2018-16570. Medicare Program; Prospective Payment System and Consolidated Billing for Skilled Nursing Facilities (SNF) Final Rule for FY 2019, SNF Value-Based Purchasing Program, and SNF Quality Reporting Program

AGENCY:

Centers for Medicare & Medicaid Services (CMS), HHS.

ACTION:

Final rule.

SUMMARY:

This final rule updates the payment rates used under the prospective payment system (PPS) for skilled nursing facilities (SNFs) for fiscal year (FY) 2019. This final rule also replaces the existing case-mix classification methodology, the Resource Utilization Groups, Version IV (RUG-IV) model, with a revised case-mix methodology called the Patient-Driven Payment Model (PDPM) beginning on October 1, 2019. The rule finalizes revisions to the regulation text that describes a beneficiary's SNF “resident” status under the consolidated billing provision and the required content of the SNF level of care certification. The rule also finalizes updates to the SNF Quality Reporting Program (QRP) and the Skilled Nursing Facility Value-Based Purchasing (VBP) Program.

DATES:

 

Effective Date: This final rule is effective October 1, 2018.

Implementation Date: The implementation date for revised case-mix methodology, PDPM, and associated policies discussed in section V. is October 1, 2019.

FOR FURTHER INFORMATION CONTACT:

 

Penny Gershman, [phone], for information related to SNF PPS clinical issues.

John Kane, [phone], for information related to the development of the payment rates and case-mix indexes, and general information.

Kia Sidbury, [phone], for information related to the wage index.

Bill Ullman, [phone], for information related to level of care determinations, and consolidated billing.

Mary Pratt, [phone], for information related to the skilled nursing facility quality reporting program.

Celeste Bostic, [phone], for information related to the skilled nursing facility value-based purchasing program.

SUPPLEMENTARY INFORMATION:

Availability of Certain Tables Exclusively Through the Internet on the CMS Website

As discussed in the FY 2014 SNF PPS final rule (78 FR 47936), tables setting forth the Wage Index for Urban Areas Based on CBSA Labor Market Areas and the Wage Index Based on CBSA Labor Market Areas for Rural Areas are no longer published in the Federal Register. Instead, these tables are available exclusively through the internet on the CMS website. The wage index tables for this final rule can be accessed on the SNF PPS Wage Index home page, at http://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​WageIndex.html.

Readers who experience any problems accessing any of these online SNF PPS wage index tables should contact Kia Sidbury at [phone].

To assist readers in referencing sections contained in this document, we are providing the following Table of Contents.

Table of Contents

I. Executive Summary

A. Purpose

B. Summary of Major Provisions

C. Summary of Cost and Benefits

D. Improving Patient Outcomes and Reducing Burden Through Meaningful Measures

E. Advancing Health Information Exchange

II. Background on SNF PPS

A. Statutory Basis and Scope

B. Initial Transition for the SNF PPS

C. Required Annual Rate Updates

III. Analysis and Responses to Public Comments on the FY 2019 SNF PPS Proposed Rule

A. General Comments on the FY 2019 SNF PPS Proposed Rule

B. SNF PPS Rate Setting Methodology and FY 2019 Update

1. Federal Base Rates

2. SNF Market Basket Update

3. Case-Mix Adjustment

4. Wage Index Adjustment

5. SNF Value-Based Purchasing Program

6. Adjusted Rate Computation Example

IV. Additional Aspects of the SNF PPS

A. SNF Level of Care—Administrative Presumption

B. Consolidated Billing

C. Payment for SNF-Level Swing-Bed Services

V. Revisions to SNF PPS Case-Mix Classification Methodology

A. Background and General Comments

B. Revisions to SNF PPS Federal Base Payment Rate Components

C. Design and Methodology for Case-Mix Adjustment of Federal Rates

D. Use of the Resident Assessment Instrument—Minimum Data Set, Version 3

E. Revisions to Therapy Provision Policies Under the SNF PPS

F. Interrupted Stay Policy

G. Relationship of PDPM to Existing Skilled Nursing Facility Level of Care Criteria

H. Effect of PDPM on Temporary AIDS Add-On Payment

I. Potential Impacts of Implementing the PDPM and Parity Adjustment

VI. Other Issues

A. Other Revisions to the Regulation Text

B. Skilled Nursing Facility (SNF) Quality Reporting Program (QRP)

C. Skilled Nursing Facility Value-Based Purchasing Program (SNF VBP)

D. Request for Information on Promoting Interoperability and Electronic Healthcare Information Exchange Through Possible Revisions to the CMS Patient Health and Safety Requirements for Hospitals and Other Medicare- and Medicaid-Participating Providers and Suppliers

VII. Collection of Information Requirements

VIII. Economic Analyses

A. Regulatory Impact Analysis

B. Regulatory Flexibility Act Analysis

C. Unfunded Mandates Reform Act Analysis

D. Federalism Analysis

E. Congressional Review Act

F. Regulatory Review Costs

I. Executive Summary

A. Purpose

This final rule updates the SNF prospective payment rates for FY 2019 as required under section 1888(e)(4)(E) of the Social Security Act (the Act). It will also respond to section 1888(e)(4)(H) of the Act, which requires the Secretary to provide for publication in the Federal Register, before the August 1 that precedes the start of each fiscal year (FY), certain specified information relating to the payment update (see section II.C. of this final rule). This final rule also replaces the existing case-mix classification methodology, the Resource Utilization Groups, Version IV (RUG-IV) model, with a revised case-mix methodology called the Patient-Driven Payment Model (PDPM) effective October 1, 2019. This rule also finalizes updates to the Skilled Nursing Facility Quality Reporting Program (SNF QRP) and Skilled Nursing Facility Value-Based Purchasing Program (SNF VBP).

B. Summary of Major Provisions

In accordance with sections 1888(e)(4)(E)(ii)(IV) and 1888(e)(5) of the Act, the federal rates in this final rule will reflect an update to the rates that we published in the SNF PPS final rule for FY 2018 (82 FR 36530), as corrected in the FY 2018 SNF PPS correction notice (82 FR 46163), which reflects the SNF market basket update for FY 2019, as required by section 1888(e)(5)(B)(iv) of the Act (as added by section 53111 of the Bipartisan Budget Act of 2018). This final rule also replaces the existing case-mix classification methodology, the Resource Utilization Groups, Version IV (RUG-IV) model, with a revised case-mix methodology called the Patient-Driven Payment Model (PDPM). It also finalizes revisions at 42 CFR 411.15(p)(3)(iv), which describes a beneficiary's SNF “resident” status under the consolidated billing provision, and 42 CFR 424.20(a)(1)(i), which describes the required content of the SNF level of care certification. Furthermore, in accordance with section 1888(h) of the Act, this final rule, beginning October 1, 2018, will reduce the adjusted federal per diem rate determined under section 1888(e)(4)(G) of the Act by 2 percent, and adjust the resulting rate by the value-based incentive payment amount earned by the SNF for that fiscal year under the SNF VBP Program. Additionally, this final rule updates policies for the SNF VBP, including requirements that apply beginning with the FY 2021 SNF VBP program year, changes to the SNF VBP scoring methodology, and the adoption of an Extraordinary Circumstances Exception policy. Finally, this rule updates requirements for the SNF QRP, including adopting a new quality measure removal factor and codifying in our regulations a number of requirements.

C. Summary of Cost and Benefits

Table 1—Cost and Benefits

Provision description	Total transfers
FY 2019 SNF PPS payment rate update\	The overall economic impact of this final rule is an estimated increase of $820 million in aggregate payments to SNFs during FY 2019.
FY 2019 SNF VBP changes	The overall economic impact of the SNF VBP Program is an estimated reduction of $211 million in aggregate payments to SNFs during FY 2019.

D. Improving Patient Outcomes and Reducing Burden Through Meaningful Measures

Regulatory reform and reducing regulatory burden are high priorities for us. To reduce the regulatory burden on the healthcare industry, lower health care costs, and enhance patient care, in October 2017, we launched the Meaningful Measures Initiative.^[1] This initiative is one component of our agency-wide Patients Over Paperwork Initiative,^[2] which is aimed at evaluating and streamlining regulations with a goal to reduce unnecessary cost and burden, increase efficiencies, and improve beneficiary experience. The Meaningful Measures Initiative is aimed at identifying the highest priority areas for quality measurement and quality improvement in order to assess the core quality of care issues that are most vital to advancing our work to improve patient outcomes. The Meaningful Measures Initiative represents a new approach to quality measures that fosters operational efficiencies, and will reduce costs including the collection and reporting burden while producing quality measurement that is more focused on meaningful outcomes.

The Meaningful Measures Framework has the following objectives:

• Address high-impact measure areas that safeguard public health;
• Patient-centered and meaningful to patients;
• Outcome-based where possible;
• Fulfill each program's statutory requirements;
• Minimize the level of burden for health care providers (for example, through a preference for EHR-based measures where possible, such as electronic clinical quality measures);
• Significant opportunity for improvement;
• Address measure needs for population based payment through alternative payment models; and
• Align across programs and/or with other payers.

In order to achieve these objectives, we have identified 19 Meaningful Measures areas and mapped them to six overarching quality priorities as shown in Table 2.

Table 2—Meaningful Measures Framework Domains and Measure Areas

Quality priority	Meaningful measure area
Making Care Safer by Reducing Harm Caused in the Delivery of Care	Healthcare-Associated Infections. Preventable Healthcare Harm.
Strengthen Person and Family Engagement as Partners in Their Care	Care is Personalized and Aligned with Patient's Goals.
	End of Life Care according to Preferences.
	Patient's Experience of Care
	Patient Reported Functional Outcomes.
	Admissions and Readmissions to Hospitals.
	Transfer of Health Information and Interoperability.
Promote Effective Prevention and Treatment of Chronic Disease	Preventive Care.
	Management of Chronic Conditions.
	Prevention, Treatment, and Management of Mental Health.
	Prevention and Treatment of Opioid and Substance Use Disorders.
	Risk Adjusted Mortality.
Work with Communities to Promote Best Practices of Healthy Living	Equity of Care.
	Community Engagement.
Make Care Affordable	Appropriate Use of Healthcare.
	Patient-focused Episode of Care.
	Risk Adjusted Total Cost of Care.

By including Meaningful Measures in our programs, we believe that we can also address the following cross-cutting measure criteria:

• Eliminating disparities;
• Tracking measurable outcomes and impact;
• Safeguarding public health;
• Achieving cost savings;
• Improving access for rural communities; and
• Reducing burden.

We believe that the Meaningful Measures Initiative will improve outcomes for patients, their families, and health care providers while reducing burden and costs for clinicians and providers and promoting operational efficiencies.

Comment: We received numerous comments from stakeholders regarding the Meaningful Measures Initiative and the impact of its implementation in CMS' quality programs. Many of these comments pertained to specific program proposals, and are discussed in the appropriate program-specific sections of this final rule. However, commenters also provided insights and recommendations for the ongoing development of the Meaningful Measures Initiative generally, including: ensuring transparency in public reporting and usability of publicly reported data; evaluating the benefit of individual measures to patients via use in quality programs weighed against the burden to providers of collecting and reporting that measure data; and identifying additional opportunities for alignment across CMS quality programs.

Response: We will continue to work with stakeholders to refine and further implement the Meaningful Measures Initiative, and will take commenters' insights and recommendations into account moving forward.

E. Advancing Health Information Exchange

The Department of Health and Human Services (HHS) has a number of initiatives designed to encourage and support the adoption of interoperable health information technology and to promote nationwide health information exchange to improve health care. The Office of the National Coordinator for Health Information Technology (ONC) and CMS work collaboratively to advance interoperability across settings of care, including post-acute care.

The Improving Medicare Post-Acute Care Transformation Act of 2015 (IMPACT Act, Pub. L. 113-185) requires assessment data to be standardized and interoperable to allow for exchange of the data among post-acute providers and other providers. To further interoperability in post-acute care, CMS has developed a Data Element Library to serve as a publicly available centralized, authoritative resource for standardized data elements and their associated mappings to health IT standards. These interoperable data elements can reduce provider burden by allowing the use and reuse of healthcare data, support provider exchange of electronic health information for care coordination, person-centered care, and support real-time, data driven, clinical decision making. Standards in the Data Element Library (https://del.cms.gov/​) can be referenced on the CMS website and in the ONC Interoperability Standards Advisory (ISA). The 2018 Interoperability Standards Advisory (ISA) is available at https://www.healthit.gov/​isa .

Most recently, the 21st Century Cures Act (Pub. L. 114-255), enacted in late 2016, requires HHS to take new steps to enable the electronic sharing of health information ensuring interoperability for providers and settings across the care continuum. Specifically, Congress directed ONC to “develop or support a trusted exchange framework, including a common agreement among health information networks nationally.” This framework (https://beta.healthit.gov/​topic/​interoperability/​trusted-exchange-framework-and-common-agreement) outlines a common set of principles for trusted exchange and minimum terms and conditions for trusted exchange in order to enable interoperability across disparate health information networks. In another important provision, Congress defined “information blocking” as practices likely to interfere with, prevent, or materially discourage access, exchange, or use of electronic health information, and established new authority for HHS to discourage these practices.

We invite providers to learn more about these important developments and how they are likely to affect SNFs.

II. Background on SNF PPS

A. Statutory Basis and Scope

As amended by section 4432 of the Balanced Budget Act of 1997 (BBA 1997, Pub. L. 105-33, enacted on August 5, 1997), section 1888(e) of the Act provides for the implementation of a PPS for SNFs. This methodology uses prospective, case-mix adjusted per diem payment rates applicable to all covered SNF services defined in section 1888(e)(2)(A) of the Act. The SNF PPS is effective for cost reporting periods beginning on or after July 1, 1998, and covers all costs of furnishing covered SNF services (routine, ancillary, and capital-related costs) other than costs associated with approved educational activities and bad debts. Under section 1888(e)(2)(A)(i) of the Act, covered SNF services include post-hospital extended care services for which benefits are provided under Part A, as well as those items and services (other than a small number of excluded services, such as physicians' services) for which payment may otherwise be made under Part B and which are furnished to Medicare beneficiaries who are residents in a SNF during a covered Part A stay. A comprehensive discussion of these provisions appears in the May 12, 1998 interim final rule (63 FR 26252). In addition, a detailed discussion of the legislative history of the SNF PPS is available online at http://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​Downloads/​Legislative_​History_​04152015.pdf.

Section 215(a) of the Protecting Access to Medicare Act of 2014 (PAMA) (Pub. L. 113-93, enacted on April 1, 2014) added section 1888(g) to the Act requiring the Secretary to specify an all-cause all-condition hospital readmission measure and an all-condition risk-adjusted potentially preventable hospital readmission measure for the SNF setting. Additionally, section 215(b) of PAMA added section 1888(h) to the Act requiring the Secretary to implement a VBP program for SNFs. Finally, section 2(c)(4) of the IMPACT Act added section 1888(e)(6) to the Act, which requires the Secretary to implement a quality reporting program for SNFs under which SNFs report data on measures and resident assessment data.

B. Initial Transition for the SNF PPS

Under sections 1888(e)(1)(A) and 1888(e)(11) of the Act, the SNF PPS included an initial, three-phase transition that blended a facility-specific rate (reflecting the individual facility's historical cost experience) with the federal case-mix adjusted rate. The transition extended through the facility's first 3 cost reporting periods under the PPS, up to and including the one that began in FY 2001. Thus, the SNF PPS is no longer operating under the transition, as all facilities have been paid at the full federal rate effective with cost reporting periods beginning in FY 2002. As we now base payments for SNFs entirely on the adjusted federal per diem rates, we no longer include adjustment factors under the transition related to facility-specific rates for the upcoming FY.

C. Required Annual Rate Updates

Section 1888(e)(4)(E) of the Act requires the SNF PPS payment rates to be updated annually. The most recent annual update occurred in a final rule that set forth updates to the SNF PPS payment rates for FY 2018 (82 FR 36530), as corrected in the FY 2018 SNF PPS correction notice (82 FR 46163).

Section 1888(e)(4)(H) of the Act specifies that we provide for publication annually in the Federal Register of the following:

• The unadjusted federal per diem rates to be applied to days of covered SNF services furnished during the upcoming FY.
• The case-mix classification system to be applied for these services during the upcoming FY.
• The factors to be applied in making the area wage adjustment for these services.

Along with other revisions discussed later in this preamble, this final rule will provide the required annual updates to the per diem payment rates for SNFs for FY 2019.

III. Analysis and Responses to Public Comments on the FY 2019 SNF PPS Proposed Rule

In response to the publication of the FY 2019 SNF PPS proposed rule, we received 290 public comments from individuals, providers, corporations, government agencies, private citizens, trade associations, and major organizations. The following are brief summaries of each proposed provision, a summary of the public comments that we received related to that proposal, and our responses to the comments.

A. General Comments on the FY 2019 SNF PPS Proposed Rule

In addition to the comments we received on specific proposals contained within the proposed rule (which we address later in this final rule), commenters also submitted the following, more general, observations on the SNF PPS and SNF care generally. A discussion of these comments, along with our responses, appears below.

Comment: A few commenters requested clarification of how a SNF may comply with the coverage requirement to provide skilled services on a daily basis and communicate intended compliance with such policy when skilled rehabilitative services are halted temporarily due to a holiday or patient illness, and the only skilled service required is rehabilitation services.

Response: As stated in the FY 2000 SNF PPS final rule (64 FR 41670), the requirement for daily skilled services should not be applied so strictly that it would not be met merely because there is a brief, isolated absence from the facility in a situation where discharge from the facility would not be practical. With regard to the “daily basis” requirement, the Medicare program does not specify in regulations or guidelines an official list of holidays of other specific occasions that a facility may observe as breaks in rehabilitation services, but recognizes that the resident's own condition dictates the amount of service that is appropriate. Accordingly, the facility itself must judge whether a brief, temporary pause in the delivery of therapy services would adversely affect the resident's condition.

This policy is also discussed at § 409.34(b), where the paragraph states that a break of 1 or 2 days in the furnishing of rehabilitation services will not preclude coverage if discharge would not be practical for the 1 or 2 days during which, for instance, the physician has suspended the therapy sessions because the patient exhibited extreme fatigue.

Comment: One commenter requested that CMS allow the addition of advanced registered nurse practitioners (ARNPs) to the rehabilitation team to meet regulatory requirements and deal with a shortage of rehabilitation physicians.

Response: We appreciate the comment. While ARNPs are eligible to enroll and participate in Medicare, it is unclear what federal regulatory requirements the commenter is concerned about that would prevent ARNPs from participating in rehabilitation team activities.

B. SNF PPS Rate Setting Methodology and FY 2019 Update

1. Federal Base Rates

Under section 1888(e)(4) of the Act, the SNF PPS uses per diem federal payment rates based on mean SNF costs in a base year (FY 1995) updated for inflation to the first effective period of the PPS. We developed the federal payment rates using allowable costs from hospital-based and freestanding SNF cost reports for reporting periods beginning in FY 1995. The data used in developing the federal rates also incorporated a Part B add-on, which is an estimate of the amounts that, prior to the SNF PPS, would be payable under Part B for covered SNF services furnished to individuals during the course of a covered Part A stay in a SNF.

In developing the rates for the initial period, we updated costs to the first effective year of the PPS (the 15-month period beginning July 1, 1998) using a SNF market basket index, and then standardized for geographic variations in wages and for the costs of facility differences in case mix. In compiling the database used to compute the federal payment rates, we excluded those providers that received new provider exemptions from the routine cost limits, as well as costs related to payments for exceptions to the routine cost limits. Using the formula that the BBA 1997 prescribed, we set the federal rates at a level equal to the weighted mean of freestanding costs plus 50 percent of the difference between the freestanding mean and weighted mean of all SNF costs (hospital-based and freestanding) combined. We computed and applied separately the payment rates for facilities located in urban and rural areas, and adjusted the portion of the federal rate attributable to wage-related costs by a wage index to reflect geographic variations in wages.

2. SNF Market Basket Update

a. SNF Market Basket Index

Section 1888(e)(5)(A) of the Act requires us to establish a SNF market basket index that reflects changes over time in the prices of an appropriate mix of goods and services included in covered SNF services. Accordingly, we have developed a SNF market basket index that encompasses the most commonly used cost categories for SNF routine services, ancillary services, and capital-related expenses. In the SNF PPS final rule for FY 2018 (82 FR 36548 through 36566), we revised and rebased the market basket index, which included updating the base year from FY 2010 to 2014.

The SNF market basket index is used to compute the market basket percentage change that is used to update the SNF federal rates on an annual basis, as required by section 1888(e)(4)(E)(ii)(IV) of the Act. This market basket percentage update is adjusted by a forecast error correction, if applicable, and then further adjusted by the application of a productivity adjustment as required by section 1888(e)(5)(B)(ii) of the Act and described in section III.B.2.d of this final rule. For FY 2019, the growth rate of the 2014-based SNF market basket in the proposed rule was estimated to be 2.7 percent, based on the IHS Global Insight, Inc. (IGI) first quarter 2018 forecast with historical data through fourth quarter 2017, before the multifactor productivity adjustment is applied. Using IGIs most recent forecast, the second quarter 2018 forecast with historical data through first quarter 2018, we calculate a growth rate of the 2014-based SNF market basket of 2.8 percent.

However, we note that section 53111 of the Bipartisan Budget Act of 2018 (BBA 2018) (Pub. L. 115-123, enacted on February 9, 2018) amended section 1888(e) of the Act to add section 1888(e)(5)(B)(iv) of the Act. Section 1888(e)(5)(B)(iv) of the Act establishes a special rule for FY 2019 that requires the market basket percentage, after the application of the productivity adjustment, to be 2.4 percent. In accordance with section 1888(e)(5)(B)(iv) of the Act, we will use a market basket percentage of 2.4 percent to update the federal rates set forth in this final rule. We proposed to revise § 413.337(d) to reflect this statutorily required 2.4 percent market basket percentage for FY 2019. In addition, to conform with section 1888(e)(5)(B)(iii) of the Act, we proposed to update the regulations to reflect the 1 percent market basket percentage required for FY 2018 (as discussed in the FY 2018 SNF PPS final rule, 82 FR 36533). Accordingly, we proposed to revise paragraph (d)(1) of § 413.337, which sets forth the market basket update formula, by revising paragraph (d)(1)(v), and by adding paragraphs (d)(1)(vi) and (d)(1)(vii). The revision to add paragraph (d)(1)(vi) reflects section 1888(e)(5)(B)(iii) of the Act (as added by section 411(a) of the Medicare Access and CHIP Reauthorization Act of 2015 (MACRA) (Pub. L. 114-10)), which, as discussed above, establishes a special rule for FY 2018 that requires the market basket percentage, after the application of the productivity adjustment, to be 1.0 percent. The revision to add paragraph (d)(1)(vii) reflects section 1888(e)(5)(B)(iv) of the Act (as added by section 53111 of BBA 2018), which establishes a special rule for FY 2019 that requires the market basket percentage, after the application of the productivity adjustment, to be 2.4 percent. These statutory provisions are self-implementing and do not require the exercise of discretion by the Secretary. In section III.B.2.e. of this final rule, we discuss the specific application of the BBA 2018-specified market basket adjustment to the forthcoming annual update of the SNF PPS payment rates. In addition, we also discuss in that section the 2 percent reduction applied to the market basket update for those SNFs that fail to submit measures data as required by section 1888(e)(6)(A) of the Act.

b. Use of the SNF Market Basket Percentage

Section 1888(e)(5)(B) of the Act defines the SNF market basket percentage as the percentage change in the SNF market basket index from the midpoint of the previous FY to the midpoint of the current FY. Absent the addition of section 1888(e)(5)(B)(iv) of the Act, added by section 53111 of BBA 2018, we would have used the percentage change in the SNF market basket index to compute the update factor for FY 2019. This factor is based on the FY 2019 percentage increase in the 2014-based SNF market basket index reflecting routine, ancillary, and capital-related expenses. In the proposed rule, the SNF market basket percentage was estimated to be 2.7 percent for FY 2019 based on IGI's first quarter 2018 forecast (with historical data through fourth quarter 2017). In this final rule, we are using IGI's more recent second quarter 2018 forecast (with historical data through first quarter 2018) and we calculate a SNF market basket percentage increase of 2.8 percent. As discussed in sections III.B.2.c and III.B.2.d of this final rule, this market basket percentage change would have been reduced by the applicable forecast error correction (as described in § 413.337(d)(2)) and by the MFP adjustment as required by section 1888(e)(5)(B)(ii) of the Act. As noted previously, section 1888(e)(5)(B)(iv) of the Act, added by section 53111 of the BBA 2018, requires us to update the SNF PPS rates for FY 2019 using a 2.4 percent SNF market basket percentage change, instead of the estimated 2.8 percent market basket percentage change adjusted by the multifactor productivity adjustment as described below. Additionally, as discussed in section II.B. of this final rule, we no longer compute update factors to adjust a facility-specific portion of the SNF PPS rates, because the initial three-phase transition period from facility-specific to full federal rates that started with cost reporting periods beginning in July 1998 has expired.

c. Forecast Error Adjustment

As discussed in the June 10, 2003 supplemental proposed rule (68 FR 34768) and finalized in the August 4, 2003 final rule (68 FR 46057 through 46059), § 413.337(d)(2) provides for an adjustment to account for market basket forecast error. The initial adjustment for market basket forecast error applied to the update of the FY 2003 rate for FY 2004, and took into account the cumulative forecast error for the period from FY 2000 through FY 2002, resulting in an increase of 3.26 percent to the FY 2004 update. Subsequent adjustments in succeeding FYs take into account the forecast error from the most recently available FY for which there is final data, and apply the difference between the forecasted and actual change in the market basket when the difference exceeds a specified threshold. We originally used a 0.25 percentage point threshold for this purpose; however, for the reasons specified in the FY 2008 SNF PPS final rule (72 FR 43425, August 3, 2007), we adopted a 0.5 percentage point threshold effective for FY 2008 and subsequent FYs. As we stated in the final rule for FY 2004 that first issued the market basket forecast error adjustment (68 FR 46058, August 4, 2003), the adjustment will reflect both upward and downward adjustments, as appropriate.

For FY 2017 (the most recently available FY for which there is final data), the estimated increase in the market basket index was 2.7 percentage points, while the actual increase for FY 2017 was 2.7 percentage points, resulting in the actual increase being the same as the estimated increase. Accordingly, as the difference between the estimated and actual amount of change in the market basket index does not exceed the 0.5 percentage point threshold, the FY 2019 market basket percentage change of 2.7 percent would not have been adjusted to account for the forecast error correction. Table 3 shows the forecasted and actual market basket amounts for FY 2017.

Table 3—Difference Between the Forecasted and Actual Market Basket Increases for FY 2017

Index	Forecasted FY 2017 increase *	Actual FY 2017 increase **	FY 2017 difference
SNF	2.7	2.7	0.0
* Published in Federal Register; based on second quarter 2016 IGI forecast (2010-based index).
** Based on the second quarter 2018 IGI forecast, with historical data through the first quarter 2018 (2010-based index).

d. Multifactor Productivity Adjustment

Section 1888(e)(5)(B)(ii) of the Act, as added by section 3401(b) of the Patient Protection and Affordable Care Act (Affordable Care Act) (Pub. L. 111-148, enacted on March 23, 2010) requires that, in FY 2012 and in subsequent FYs, the market basket percentage under the SNF payment system (as described in section 1888(e)(5)(B)(i) of the Act) is to be reduced annually by the multifactor productivity (MFP) adjustment described in section 1886(b)(3)(B)(xi)(II) of the Act. Section 1886(b)(3)(B)(xi)(II) of the Act, in turn, defines the MFP adjustment to be equal to the 10-year moving average of changes in annual economy-wide private nonfarm business multi-factor productivity (as projected by the Secretary for the 10-year period ending with the applicable FY, year, cost-reporting period, or other annual period). The Bureau of Labor Statistics (BLS) is the agency that publishes the official measure of private nonfarm business MFP. We refer readers to the BLS website at http://www.bls.gov/​mfp for the BLS historical published MFP data.

MFP is derived by subtracting the contribution of labor and capital inputs growth from output growth. The projections of the components of MFP are currently produced by IGI, a nationally recognized economic forecasting firm with which CMS contracts to forecast the components of the market baskets and MFP. To generate a forecast of MFP, IGI replicates the MFP measure calculated by the BLS, using a series of proxy variables derived from IGI's U.S. macroeconomic models. For a discussion of the MFP projection methodology, we refer readers to the FY 2012 SNF PPS final rule (76 FR 48527 through 48529) and the FY 2016 SNF PPS final rule (80 FR 46395). A complete description of the MFP projection methodology is available on our website at http://www.cms.gov/​Research-Statistics-Data-and-Systems/​Statistics-Trends-and-Reports/​MedicareProgramRatesStats/​MarketBasketResearch.html.

1. Incorporating the MFP Adjustment Into the Market Basket Update

Per section 1888(e)(5)(A) of the Act, the Secretary shall establish a SNF market basket index that reflects changes over time in the prices of an appropriate mix of goods and services included in covered SNF services. Section 1888(e)(5)(B)(ii) of the Act, added by section 3401(b) of the Affordable Care Act, requires that for FY 2012 and each subsequent FY, after determining the market basket percentage described in section 1888(e)(5)(B)(i) of the Act, the Secretary shall reduce such percentage by the productivity adjustment described in section 1886(b)(3)(B)(xi)(II) of the Act (which we refer to as the MFP adjustment). Section 1888(e)(5)(B)(ii) of the Act further states that the reduction of the market basket percentage by the MFP adjustment may result in the market basket percentage being less than zero for a FY, and may result in payment rates under section 1888(e) of the Act being less than such payment rates for the preceding fiscal year.

The MFP adjustment, calculated as the 10-year moving average of changes in MFP for the period ending September 30, 2019, is estimated to be 0.8 percent based on IGI's second quarter 2018 forecast. Also, consistent with section 1888(e)(5)(B)(i) of the Act and § 413.337(d)(2), the market basket percentage for FY 2019 for the SNF PPS is based on IGI's second quarter 2018 forecast of the SNF market basket percentage, which is estimated to be 2.8 percent. The proposed rule reflected a market basket percentage for FY 2019 of 2.7 percent and an MFP adjustment of 0.8 percent based on IGI's first quarter 2018 forecast.

If not for the enactment of section 53111 of the BBA 2018, the FY 2019 update would have been calculated in accordance with section 1888(e)(5)(B)(i) and (ii) of the Act, pursuant to which the market basket percentage determined under section 1888(e)(5)(B)(i) of the Act (that is, 2.8 percent) would have been reduced by the MFP adjustment (the 10-year moving average of changes in MFP for the period ending September 30, 2019) of 0.8 percent, which would have been calculated as described above and based on IGI's second quarter 2018 forecast. Absent the enactment of section 53111 of the BBA 2018, the resulting MFP-adjusted SNF market basket update would have been equal to 2.0 percent, or 2.8 percent less 0.8 percentage point. However, as discussed above, section 1888(e)(5)(B)(iv) of the Act, added by section 53111 of the BBA 2018, requires us to apply a 2.4 percent market basket percentage increase in determining the FY 2019 SNF payment rates set forth in this final rule (without regard to the MFP adjustment described above).

e. Market Basket Update Factor for FY 2019

Sections 1888(e)(4)(E)(ii)(IV) and 1888(e)(5)(i) of the Act require that the update factor used to establish the FY 2019 unadjusted federal rates be at a level equal to the market basket index percentage change. Accordingly, we determined the total growth from the average market basket level for the period of October 1, 2017, through September 30, 2018 to the average market basket level for the period of October 1, 2018, through September 30, 2019. This process yields a percentage change in the 2014-based SNF market basket of 2.8 percent.

As further explained in section III.B.2.c. of this final rule, as applicable, we adjust the market basket percentage change by the forecast error from the most recently available FY for which there is final data and apply this adjustment whenever the difference between the forecasted and actual percentage change in the market basket exceeds a 0.5 percentage point threshold. Since the difference between the forecasted FY 2017 SNF market basket percentage change and the actual FY 2017 SNF market basket percentage change (FY 2017 is the most recently available FY for which there is historical data) did not exceed the 0.5 percentage point threshold, the FY 2019 market basket percentage change of 2.8 percent would not be adjusted by the forecast error correction.

If not for the enactment of section 53111 of the BBA 2018, the SNF market basket for FY 2019 would have been determined in accordance with section 1888(e)(5)(B)(ii) of the Act, which requires us to reduce the market basket percentage change by the MFP adjustment (the 10-year moving average of changes in MFP for the period ending September 30, 2019) of 0.8 percent, as described in section III.B.2.d.1. of this final rule. Thus, absent the enactment of the BBA 2018, the resulting net SNF market basket update would have been equal to 2.0 percent, or 2.8 percent less the 0.8 percentage point MFP adjustment. We note that our policy has been that, if more recent data become available (for example, a more recent estimate of the SNF market basket and/or MFP adjustment), we would use such data, if appropriate, to determine the SNF market basket percentage change, labor-related share relative importance, forecast error adjustment, and MFP adjustment in the SNF PPS final rule.

Historically, we have used the SNF market basket, adjusted as described above, to adjust each per diem component of the federal rates forward to reflect the change in the average prices from one year to the next. However, section 1888(e)(5)(B)(iv) of the Act, as added by section 53111 of the BBA 2018, requires us to use a market basket percentage of 2.4 percent, after application of the MFP to adjust the federal rates for FY 2019. Under section 1888(e)(5)(B)(iv) of the Act, the market basket percentage increase used to determine the federal rates set forth in Table 4 and 5 in this final rule will be 2.4 percent for FY 2019.

In addition, we note that section 1888(e)(6)(A)(i) of the Act provides that, beginning with FY 2018, SNFs that fail to submit data, as applicable, in accordance with sections 1888(e)(6)(B)(i)(II) and (III) of the Act for a fiscal year will receive a 2.0 percentage point reduction to their market basket update for the fiscal year involved, after application of section 1888(e)(5)(B)(ii) of the Act (the MFP adjustment) and section 1888(e)(5)(B)(iii) of the Act (the 1 percent market basket increase for FY 2018). In addition, section 1888(e)(6)(A)(ii) of the Act states that application of the 2.0 percentage point reduction (after application of section 1888(e)(5)(B)(ii) and (iii) of the Act) may result in the market basket index percentage change being less than 0.0 for a fiscal year, and may result in payment rates for a fiscal year being less than such payment rates for the preceding fiscal year. Section 1888(e)(6)(A)(iii) of the Act further specifies that the 2.0 percentage point reduction is applied in a noncumulative manner, so that any reduction made under section 1888(e)(6)(A)(i) of the Act applies only with respect to the fiscal year involved, and that the reduction cannot be taken into account in computing the payment amount for a subsequent fiscal year.

Accordingly, we proposed that for SNFs that do not satisfy the reporting requirements for the FY 2019 SNF QRP, we would apply a 2.0 percentage point reduction to the SNF market basket percentage change for that fiscal year, after application of any applicable forecast error adjustment as specified in § 413.337(d)(2) and the MFP adjustment as specified in § 413.337(d)(3). In the FY 2019 SNF PPS proposed rule (83 FR 21024), we proposed that, for FY 2019, the application of this reduction to SNFs that have not met the requirements for the FY 2019 SNF QRP would result in a market basket index percentage change for FY 2019 that is less than zero (specifically, a net update of negative 0.1 percentage point, derived by subtracting 2 percent from the MFP-adjusted market basket update of 1.9 percent), and would also result in FY 2019 payment rates that are less than such payment rates for the preceding FY. However, we inadvertently applied the 2.0 percent reduction to the market basket adjustment factor that would have applied were it not for the application of the BBA 2018 stipulated market basket update factor rather than to the BBA 2018 stipulated market basket update factor of 2.4 percent. Therefore, when properly applied, the net update for providers that fail to meet the requirements for the FY 2019 SNF QRP will be 0.4 percent, rather than the negative 0.1 percent discussed in the proposed rule. We invited comments on these proposals.

Commenters submitted the following comments related to the proposed rule's discussion of the Market Basket Update Factor for FY 2019. A discussion of these comments, along with our responses, appears below.

Comment: We received a number of comments in relation to applying the FY 2019 market basket update factor in the determination of the FY 2019 unadjusted federal per diem rates, with some commenters supporting its application in determining the FY 2019 unadjusted per diem rates, while others opposed its application. In their March 2018 report (available at http://www.medpac.gov/​docs/​default-source/​reports/​mar18_​medpac_​ch8_​sec.pdf) and in their comment on the FY 2019 SNF PPS proposed rule, MedPAC recommended that we eliminate the market basket update for SNFs altogether for FY 2019 and FY 2020 and implement revisions to the SNF PPS.

Response: We appreciate all of the comments received on the proposed market basket update for FY 2019. In response to those comments opposing the application of the FY 2019 market basket update factor in determining the FY 2019 unadjusted federal per diem rates, specifically MedPAC's proposal to eliminate the market basket update for SNFs, we are required to update the unadjusted Federal per diem rates for FY 2019 by 2.4 percent under section 1888(e)(4)(E)(ii)(IV) and (e)(5)(B) of the Act. as amended by section 53111 of the BBA 2018.

Comment: A few commenters expressed concern with regard to CMS applying the 2.0 percentage point reduction to the market basket increase to the standard market basket adjustment of 1.9 percent, rather than to the market basket required as a result of the BBA 2018. These commenters requested that CMS reconsider this decision and to apply the QRP-related market basket reduction to the BBA 2018-stipulated market basket.

Response: We appreciate the comments on this issue. Further, we agree with commenters that the QRP-related reduction to the market basket should be applied to the BBA 2018-stipulated market basket. Therefore, the market basket update factor that would be applied in cases where a provider has not met the requirements of the FY 2019 SNF QRP would be a positive 0.4 percent, rather than the negative 0.1 percent discussed in the FY 2019 SNF PPS proposed rule.

Accordingly, for the reasons specified in this final rule and in the FY 2019 SNF PPS proposed rule (83 FR 21021 through 21024), we are applying the FY 2019 SNF market basket increase factor of 2.4 percent, as stipulated by the BBA 2018, in our determination of the FY 2019 SNF PPS unadjusted federal per diem rates. As described in this section, we are adjusting each per diem component of the federal rates forward to reflect the BBA 2018 stipulated update factor for FY 2019.

Tables 4 and 5 reflect the updated components of the unadjusted federal rates for FY 2019, prior to adjustment for case-mix.

Table 4—FY 2019 Unadjusted Federal Rate per Diem—Urban

Rate component	Nursing— case-mix	Therapy— case-mix	Therapy— non-case-mix	Non-case-mix
Per Diem Amount	$181.44	$136.67	$18.00	$92.60

Table 5—FY 2019 Unadjusted Federal Rate per Diem—Rural

Rate component	Nursing— case-mix	Therapy— case-mix	Therapy— Non-case-mix	Non-case-mix
Per Diem Amount	$173.34	$157.60	$19.23	$94.31

3. Case-Mix Adjustment

Under section 1888(e)(4)(G)(i) of the Act, the federal rate also incorporates an adjustment to account for facility case-mix, using a classification system that accounts for the relative resource utilization of different patient types. The statute specifies that the adjustment is to reflect both a resident classification system that the Secretary establishes to account for the relative resource use of different patient types, as well as resident assessment data and other data that the Secretary considers appropriate. In the interim final rule with comment period that initially implemented the SNF PPS (63 FR 26252, May 12, 1998), we developed the RUG-III case-mix classification system, which tied the amount of payment to resident resource use in combination with resident characteristic information. Staff time measurement (STM) studies conducted in 1990, 1995, and 1997 provided information on resource use (time spent by staff members on residents) and resident characteristics that enabled us not only to establish RUG-III, but also to create case-mix indexes (CMIs). The original RUG-III grouper logic was based on clinical data collected in 1990, 1995, and 1997. As discussed in the SNF PPS proposed rule for FY 2010 (74 FR 22208), we subsequently conducted a multi-year data collection and analysis under the Staff Time and Resource Intensity Verification (STRIVE) project to update the case-mix classification system for FY 2011. The resulting Resource Utilization Groups, Version 4 (RUG-IV) case-mix classification system reflected the data collected in 2006 through 2007 during the STRIVE project, and was finalized in the FY 2010 SNF PPS final rule (74 FR 40288) to take effect in FY 2011 concurrently with an updated new resident assessment instrument, version 3.0 of the Minimum Data Set (MDS 3.0), which collects the clinical data used for case-mix classification under RUG-IV.

We note that case-mix classification is based, in part, on the beneficiary's need for skilled nursing care and therapy services. The case-mix classification system uses clinical data from the MDS to assign a case-mix group to each patient that is then used to calculate a per diem payment under the SNF PPS. As discussed in section IV.A. of this final rule, the clinical orientation of the case-mix classification system supports the SNF PPS's use of an administrative presumption that considers a beneficiary's initial case-mix classification to assist in making certain SNF level of care determinations. Further, because the MDS is used as a basis for payment, as well as a clinical assessment, we have provided extensive training on proper coding and the timeframes for MDS completion in our Resident Assessment Instrument (RAI) Manual. For an MDS to be considered valid for use in determining payment, the MDS assessment must be completed in compliance with the instructions in the RAI Manual in effect at the time the assessment is completed. For payment and quality monitoring purposes, the RAI Manual consists of both the Manual instructions and the interpretive guidance and policy clarifications posted on the appropriate MDS website at http://www.cms.gov/​Medicare/​Quality-Initiatives-Patient-Assessment-Instruments/​NursingHomeQualityInits/​MDS30RAIManual.html.

In addition, we note that section 511 of the Medicare Prescription Drug, Improvement, and Modernization Act of 2003 (MMA, Pub. L. 108-173, enacted December 8, 2003) amended section 1888(e)(12) of the Act to provide for a temporary increase of 128 percent in the PPS per diem payment for any SNF residents with Acquired Immune Deficiency Syndrome (AIDS), effective with services furnished on or after October 1, 2004. This special add-on for SNF residents with AIDS was to remain in effect only until the Secretary certifies that there is an appropriate adjustment in the case mix to compensate for the increased costs associated with such residents. The MMA add-on for SNF residents with AIDS is also discussed in Program Transmittal #160 (Change Request #3291), issued on April 30, 2004, which is available online at www.cms.gov/​transmittals/​downloads/​r160cp.pdf. In the SNF PPS final rule for FY 2010 (74 FR 40288), we did not address this certification in that final rule's implementation of the case-mix refinements for RUG-IV, thus allowing the add-on payment required by section 511 of the MMA to remain in effect for the time being. (We discuss in section V.H. of this final rule the specific payment adjustments that we proposed under the proposed PDPM to provide for an appropriate adjustment in the case mix to compensate for the increased costs associated with such residents.)

For the limited number of SNF residents that qualify for the MMA add-on, there is a significant increase in payments. As explained in the FY 2016 SNF PPS final rule (80 FR 46397 through 46398), on October 1, 2015 (consistent with section 212 of PAMA), we converted from using ICD-9-CM code 042 to ICD-10-CM code B20 for identifying those residents for whom it is appropriate to apply the AIDS add-on established by section 511 of the MMA. For FY 2019, an urban facility with a resident with AIDS in RUG-IV group “HC2” would have a case-mix adjusted per diem payment of $453.52 (see Table 6) before the application of the MMA adjustment. After an increase of 128 percent, this urban facility would receive a case-mix adjusted per diem payment of approximately $1,034.03.

Under section 1888(e)(4)(H), each update of the payment rates must include the case-mix classification methodology applicable for the upcoming FY. The FY 2019 payment rates set forth in this final rule reflect the use of the RUG-IV case-mix classification system from October 1, 2018, through September 30, 2019. We list the final case-mix adjusted RUG-IV payment rates for FY 2019, provided separately for urban and rural SNFs, in Tables 6 and 7 with corresponding case-mix values. We use the revised OMB delineations adopted in the FY 2015 SNF PPS final rule (79 FR 45632, 45634) to identify a facility's urban or rural status for the purpose of determining which set of rate tables would apply to the facility. Tables 6 and 7 do not reflect the add-on for SNF residents with AIDS enacted by section 511 of the MMA, which we apply only after making all other adjustments (such as wage index and case-mix). Additionally, Tables 6 and 7 do not reflect adjustments which may be made to the SNF PPS rates as a result of either the SNF QRP, discussed in section VI.B of this final rule, or the SNF VBP program, discussed in sections III.B.5 and VI.C of this final rule.

Table 6—RUG-IV Case-Mix Adjusted Federal Rates and Associated Indexes—Urban

RUG-IV category	Nursing index	Therapy index	Nursing component	Therapy component	Non-case mix therapy comp	Non-case mix component	Total rate
RUX	2.67	1.87	$484.44	$255.57		$92.60	$832.89
RUL	2.57	1.87	466.30	255.57		92.60	814.74
RVX	2.61	1.28	473.56	174.94		92.60	741.34
RVL	2.19	1.28	397.35	174.94		92.60	665.11
RHX	2.55	0.85	462.67	116.17		92.60	671.66
RHL	2.15	0.85	390.10	116.17		92.60	599.06
RMX	2.47	0.55	448.16	75.17		92.60	616.13
RML	2.19	0.55	397.35	75.17		92.60	565.31
RLX	2.26	0.28	410.05	38.27		92.60	541.10
RUC	1.56	1.87	283.05	255.57		92.60	631.42
RUB	1.56	1.87	283.05	255.57		92.60	631.42
RUA	0.99	1.87	179.63	255.57		92.60	527.97
RVC	1.51	1.28	273.97	174.94		92.60	541.69
RVB	1.11	1.28	201.40	174.94		92.60	469.09
RVA	1.10	1.28	199.58	174.94		92.60	467.27
RHC	1.45	0.85	263.09	116.17		92.60	472.01
RHB	1.19	0.85	215.91	116.17		92.60	424.82
RHA	0.91	0.85	165.11	116.17		92.60	374.00
RMC	1.36	0.55	246.76	75.17		92.60	414.66
RMB	1.22	0.55	221.36	75.17		92.60	389.25
RMA	0.84	0.55	152.41	75.17		92.60	320.28
RLB	1.50	0.28	272.16	38.27		92.60	403.16
RLA	0.71	0.28	128.82	38.27		92.60	259.78
ES3	3.58		649.56		18.00	92.60	760.41
ES2	2.67		484.44		18.00	92.60	595.25
ES1	2.32		420.94		18.00	92.60	531.72
HE2	2.22		402.80		18.00	92.60	513.57
HE1	1.74		315.71		18.00	92.60	426.45
HD2	2.04		370.14		18.00	92.60	480.90
HD1	1.60		290.30		18.00	92.60	401.04
HC2	1.89		342.92		18.00	92.60	453.68
HC1	1.48		268.53		18.00	92.60	379.26
HB2	1.86		337.48		18.00	92.60	448.23
HB1	1.46		264.90		18.00	92.60	375.63
LE2	1.96		355.62		18.00	92.60	466.38
LE1	1.54		279.42		18.00	92.60	390.15
LD2	1.86		337.48		18.00	92.60	448.23
LD1	1.46		264.90		18.00	92.60	375.63
LC2	1.56		283.05		18.00	92.60	393.78
LC1	1.22		221.36		18.00	92.60	332.07
LB2	1.45		263.09		18.00	92.60	373.82
LB1	1.14		206.84		18.00	92.60	317.55
CE2	1.68		304.82		18.00	92.60	415.56
CE1	1.50		272.16		18.00	92.60	382.89
CD2	1.56		283.05		18.00	92.60	393.78
CD1	1.38		250.39		18.00	92.60	361.11
CC2	1.29		234.06		18.00	92.60	344.78
CC1	1.15		208.66		18.00	92.60	319.37
CB2	1.15		208.66		18.00	92.60	319.37
CB1	1.02		185.07		18.00	92.60	295.77
CA2	0.88		159.67		18.00	92.60	270.36
CA1	0.78		141.52		18.00	92.60	252.21
BB2	0.97		176.00		18.00	92.60	286.70
BB1	0.90		163.30		18.00	92.60	273.99
BA2	0.70		127.01		18.00	92.60	237.69
BA1	0.64		116.12		18.00	92.60	226.80
PE2	1.50		272.16		18.00	92.60	382.89
PE1	1.40		254.02		18.00	92.60	364.74
PD2	1.38		250.39		18.00	92.60	361.11
PD1	1.28		232.24		18.00	92.60	342.96
PC2	1.10		199.58		18.00	92.60	310.29
PC1	1.02		185.07		18.00	92.60	295.77
PB2	0.84		152.41		18.00	92.60	263.10
PB1	0.78		141.52		18.00	92.60	252.21
PA2	0.59		107.05		18.00	92.60	217.73
PA1	0.54		97.98		18.00	92.60	208.65

Table 7—RUG-IV Case-Mix Adjusted Federal Rates and Associated Indexes—Rural

RUG-IV category	Nursing index	Therapy index	Nursing component	Therapy component	Non-case mix therapy comp	Non-case mix component	Total rate
RUX	2.67	1.87	$462.82	$294.71		$94.31	$852.10
RUL	2.57	1.87	445.48	294.71		94.31	834.76
RVX	2.61	1.28	452.42	201.73		94.31	748.68
RVL	2.19	1.28	379.61	201.73		94.31	675.85
RHX	2.55	0.85	442.02	133.96		94.31	670.48
RHL	2.15	0.85	372.68	133.96		94.31	601.13
RMX	2.47	0.55	428.15	86.68		94.31	609.32
RML	2.19	0.55	379.61	86.68		94.31	560.77
RLX	2.26	0.28	391.75	44.13		94.31	530.34
RUC	1.56	1.87	270.41	294.71		94.31	659.64
RUB	1.56	1.87	270.41	294.71		94.31	659.64
RUA	0.99	1.87	171.61	294.71		94.31	560.81
RVC	1.51	1.28	261.74	201.73		94.31	557.95
RVB	1.11	1.28	192.41	201.73		94.31	488.59
RVA	1.10	1.28	190.67	201.73		94.31	486.86
RHC	1.45	0.85	251.34	133.96		94.31	479.76
RHB	1.19	0.85	206.27	133.96		94.31	434.67
RHA	0.91	0.85	157.74	133.96		94.31	386.12
RMC	1.36	0.55	235.74	86.68		94.31	416.86
RMB	1.22	0.55	211.47	86.68		94.31	392.59
RMA	0.84	0.55	145.61	86.68		94.31	326.70
RLB	1.50	0.28	260.01	44.13		94.31	398.57
RLA	0.71	0.28	123.07	44.13		94.31	261.59
ES3	3.58		620.56		19.23	94.31	734.31
ES2	2.67		462.82		19.23	94.31	576.52
ES1	2.32		402.15		19.23	94.31	515.83
HE2	2.22		384.81		19.23	94.31	498.50
HE1	1.74		301.61		19.23	94.31	415.27
HD2	2.04		353.61		19.23	94.31	467.29
HD1	1.60		277.34		19.23	94.31	390.99
HC2	1.89		327.61		19.23	94.31	441.28
HC1	1.48		256.54		19.23	94.31	370.19
HB2	1.86		322.41		19.23	94.31	436.08
HB1	1.46		253.08		19.23	94.31	366.72
LE2	1.96		339.75		19.23	94.31	453.41
LE1	1.54		266.94		19.23	94.31	380.59
LD2	1.86		322.41		19.23	94.31	436.08
LD1	1.46		253.08		19.23	94.31	366.72
LC2	1.56		270.41		19.23	94.31	384.06
LC1	1.22		211.47		19.23	94.31	325.11
LB2	1.45		251.34		19.23	94.31	364.99
LB1	1.14		197.61		19.23	94.31	311.23
CE2	1.68		291.21		19.23	94.31	404.87
CE1	1.50		260.01		19.23	94.31	373.66
CD2	1.56		270.41		19.23	94.31	384.06
CD1	1.38		239.21		19.23	94.31	352.85
CC2	1.29		223.61		19.23	94.31	337.24
CC1	1.15		199.34		19.23	94.31	312.97
CB2	1.15		199.34		19.23	94.31	312.97
CB1	1.02		176.81		19.23	94.31	290.43
CA2	0.88		152.54		19.23	94.31	266.15
CA1	0.78		135.21		19.23	94.31	248.81
BB2	0.97		168.14		19.23	94.31	281.76
BB1	0.90		156.01		19.23	94.31	269.62
BA2	0.70		121.34		19.23	94.31	234.94
BA1	0.64		110.94		19.23	94.31	224.54
PE2	1.50		260.01		19.23	94.31	373.66
PE1	1.40		242.68		19.23	94.31	356.32
PD2	1.38		239.21		19.23	94.31	352.85
PD1	1.28		221.88		19.23	94.31	335.51
PC2	1.10		190.67		19.23	94.31	304.30
PC1	1.02		176.81		19.23	94.31	290.43
PB2	0.84		145.61		19.23	94.31	259.22
PB1	0.78		135.21		19.23	94.31	248.81
PA2	0.59		102.27		19.23	94.31	215.87
PA1	0.54		93.60		19.23	94.31	207.20

4. Wage Index Adjustment

Section 1888(e)(4)(G)(ii) of the Act requires that we adjust the federal rates to account for differences in area wage levels, using a wage index that the Secretary determines appropriate. Since the inception of the SNF PPS, we have used hospital inpatient wage data in developing a wage index to be applied to SNFs. We proposed to continue this practice for FY 2019, as we continue to believe that in the absence of SNF-specific wage data, using the hospital inpatient wage index data is appropriate and reasonable for the SNF PPS. As explained in the update notice for FY 2005 (69 FR 45786), the SNF PPS does not use the hospital area wage index's occupational mix adjustment, as this adjustment serves specifically to define the occupational categories more clearly in a hospital setting; moreover, the collection of the occupational wage data also excludes any wage data related to SNFs. Therefore, we believe that using the updated wage data exclusive of the occupational mix adjustment continues to be appropriate for SNF payments. For FY 2019, the updated wage data are for hospital cost reporting periods beginning on or after October 1, 2014 and before October 1, 2015 (FY 2015 cost report data).

We note that section 315 of the Medicare, Medicaid, and SCHIP Benefits Improvement and Protection Act of 2000 (BIPA, Pub. L. 106-554, enacted on December 21, 2000) authorized us to establish a geographic reclassification procedure that is specific to SNFs, but only after collecting the data necessary to establish a SNF wage index that is based on wage data from nursing homes. However, to date, this has proven to be unfeasible due to the volatility of existing SNF wage data and the significant amount of resources that would be required to improve the quality of that data. More specifically, auditing all SNF cost reports, similar to the process used to audit inpatient hospital cost reports for purposes of the Inpatient Prospective Payment System (IPPS) wage index, would place a burden on providers in terms of recordkeeping and completion of the cost report worksheet. As discussed in greater detail later in this section, adopting such an approach would require a significant commitment of resources by CMS and the Medicare Administrative Contractors, potentially far in excess of those required under the IPPS given that there are nearly five times as many SNFs as there are inpatient hospitals. Therefore, while we continue to believe that the development of such an audit process could improve SNF cost reports in such a manner as to permit us to establish a SNF-specific wage index, we do not regard an undertaking of this magnitude as being feasible within the current level of programmatic resources.

In addition, we proposed to continue to use the same methodology discussed in the SNF PPS final rule for FY 2008 (72 FR 43423) to address those geographic areas in which there are no hospitals, and thus, no hospital wage index data on which to base the calculation of the FY 2019 SNF PPS wage index. For rural geographic areas that do not have hospitals, and therefore, lack hospital wage data on which to base an area wage adjustment, we would use the average wage index from all contiguous Core-Based Statistical Areas (CBSAs) as a reasonable proxy. For FY 2019, there are no rural geographic areas that do not have hospitals, and thus, this methodology would not be applied. For rural Puerto Rico, we would not apply this methodology due to the distinct economic circumstances that exist there (for example, due to the close proximity to one another of almost all of Puerto Rico's various urban and non-urban areas, this methodology would produce a wage index for rural Puerto Rico that is higher than that in half of its urban areas); instead, we would continue to use the most recent wage index previously available for that area. For urban areas without specific hospital wage index data, we would use the average wage indexes of all of the urban areas within the state to serve as a reasonable proxy for the wage index of that urban CBSA. For FY 2019, the only urban area without wage index data available is CBSA 25980, Hinesville-Fort Stewart, GA. The final wage index applicable to FY 2019 is set forth in Tables A and B available on the CMS website at http://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​WageIndex.html.

In the SNF PPS final rule for FY 2006 (70 FR 45026, August 4, 2005), we adopted the changes discussed in OMB Bulletin No. 03-04 (June 6, 2003), which announced revised definitions for MSAs and the creation of micropolitan statistical areas and combined statistical areas. In adopting the CBSA geographic designations, we provided for a 1-year transition in FY 2006 with a blended wage index for all providers. For FY 2006, the wage index for each provider consisted of a blend of 50 percent of the FY 2006 MSA-based wage index and 50 percent of the FY 2006 CBSA-based wage index (both using FY 2002 hospital data). We referred to the blended wage index as the FY 2006 SNF PPS transition wage index. As discussed in the SNF PPS final rule for FY 2006 (70 FR 45041), since the expiration of this 1-year transition on September 30, 2006, we have used the full CBSA-based wage index values.

In the FY 2015 SNF PPS final rule (79 FR 45644 through 45646), we finalized changes to the SNF PPS wage index based on the newest OMB delineations, as described in OMB Bulletin No. 13-01, beginning in FY 2015, including a 1-year transition with a blended wage index for FY 2015. OMB Bulletin No. 13-01 established revised delineations for Metropolitan Statistical Areas, Micropolitan Statistical Areas, and Combined Statistical Areas in the United States and Puerto Rico based on the 2010 Census, and provided guidance on the use of the delineations of these statistical areas using standards published on June 28, 2010 in the Federal Register (75 FR 37246 through 37252). Subsequently, on July 15, 2015, OMB issued OMB Bulletin No. 15-01, which provides minor updates to and supersedes OMB Bulletin No. 13-01 that was issued on February 28, 2013. The attachment to OMB Bulletin No. 15-01 provides detailed information on the update to statistical areas since February 28, 2013. The updates provided in OMB Bulletin No. 15-01 are based on the application of the 2010 Standards for Delineating Metropolitan and Micropolitan Statistical Areas to Census Bureau population estimates for July 1, 2012 and July 1, 2013. As we previously stated in the FY 2008 SNF PPS proposed and final rules (72 FR 25538 through 25539, and 72 FR 43423), we wish to note that this and all subsequent SNF PPS rules and notices are considered to incorporate any updates and revisions set forth in the most recent OMB bulletin that applies to the hospital wage data used to determine the current SNF PPS wage index.

On August 15, 2017, OMB announced that one Micropolitan Statistical Area now qualifies as a Metropolitan Statistical Area (OMB Bulletin No. 17-01). The new urban CBSA is as follows:

• Twin Falls, Idaho (CBSA 46300). This CBSA is comprised of the principal city of Twin Falls, Idaho in Jerome County, Idaho and Twin Falls County, Idaho.

The OMB bulletin is available on the OMB website at https://www.whitehouse.gov/​sites/​whitehouse.gov/​files/​omb/​bulletins/​2017/​b-17-01.pdf. As we stated in the proposed rule (83 FR 21028), we did not have sufficient time to include this change in the computation of the proposed FY 2019 wage index, rate setting, and tables. We also stated that this new CBSA may affect the budget neutrality factor and wage indexes, depending on the impact of the overall payments of the hospital located in this new CBSA. In the proposed rule, we provided an estimate of this new area's wage index based on the average hourly wage, unadjusted for occupational mix, for new CBSA 46300 and the national average hourly wages from the wage data for the proposed FY 2019 wage index. Currently, provider 130002 is the only hospital located in Twin Falls County, Idaho, and there are no hospitals located in Jerome County, Idaho. Thus, the wage index for CBSA 46300 is calculated using the average hourly wage data for one provider (provider 130002). In this final rule, we are providing below this new area's wage index based on the updated average hourly wage, unadjusted for occupational mix, for new CBSA 46300 and the updated national average hourly wages from the wage data for the FY 2019 wage index, and we are incorporating this change into the final FY 2019 wage index, rate setting and tables. Taking the unadjusted average hourly wage of $35.8336 of new CBSA 46300 and dividing by the national average hourly wage of $42.955567020 results in the FY 2019 wage index of 0.8334 for CBSA 46300.

In the proposed rule, we stated that in the final rule, we would incorporate this change into the final FY 2019 wage index, rate setting and tables. We did not receive any comments on this issue. Thus, in this final rule, we have incorporated this change into the final FY 2019 wage index, rate setting and tables. As we proposed, for FY 2019, we will use the OMB delineations that were adopted beginning with FY 2015 to calculate the area wage indexes, with updates as reflected in OMB Bulletin Nos. 15-01 and 17-01. As noted above, the wage index applicable to FY 2019 (with the CBSA update from OMB Bulletin No. 17-01 specified above) is set forth in Tables A and B available on the CMS website at http://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​WageIndex.html.

Once calculated, we stated in the proposed rule that we would apply the wage index adjustment to the labor-related portion of the federal rate. Each year, we calculate a revised labor-related share, based on the relative importance of labor-related cost categories (that is, those cost categories that are labor-intensive and vary with the local labor market) in the input price index. In the SNF PPS final rule for FY 2018 (82 FR 36548 through 36566), we finalized a proposal to revise the labor-related share to reflect the relative importance of the 2014-based SNF market basket cost weights for the following cost categories: Wages and Salaries; Employee Benefits; Professional Fees: Labor-Related; Administrative and Facilities Support Services; Installation, Maintenance, and Repair Services; All Other: Labor-Related Services; and a proportion of Capital-Related expenses.

We calculate the labor-related relative importance from the SNF market basket, and it approximates the labor-related portion of the total costs after taking into account historical and projected price changes between the base year and FY 2019. The price proxies that move the different cost categories in the market basket do not necessarily change at the same rate, and the relative importance captures these changes. Accordingly, the relative importance figure more closely reflects the cost share weights for FY 2019 than the base year weights from the SNF market basket.

We calculate the labor-related relative importance for FY 2019 in four steps. First, we compute the FY 2019 price index level for the total market basket and each cost category of the market basket. Second, we calculate a ratio for each cost category by dividing the FY 2019 price index level for that cost category by the total market basket price index level. Third, we determine the FY 2019 relative importance for each cost category by multiplying this ratio by the base year (2014) weight. Finally, we add the FY 2019 relative importance for each of the labor-related cost categories (Wages and Salaries, Employee Benefits, Professional Fees: Labor-Related, Administrative and Facilities Support Services, Installation, Maintenance, and Repair Services, All Other: Labor-related services, and a portion of Capital-Related expenses) to produce the FY 2019 labor-related relative importance. Table 8 summarizes the updated labor-related share for FY 2019, based on IGI's second quarter 2018 forecast with historical data through first quarter 2018, compared to the labor-related share that was used for the FY 2018 SNF PPS final rule. In the FY 2019 proposed rule, we presented the FY 2019 labor-related share based on IGI's first quarter 2018 forecast and further stated that if more recent data became available (for example, a more recent estimate of the SNF market basket and/or MFP adjustment), we would use such data, if appropriate, to determine the SNF market basket percentage change, labor-related share relative importance, forecast error adjustment, and MFP adjustment in the SNF PPS final rule.

Table 8—Labor-Related Relative Importance, FY 2018 and FY 2019

	Relative importance, labor-related, FY 2018 17:2 forecast 1	Relative importance, labor-related, FY 2019 18:2 forecast 2
Wages and salaries	50.3	50.2
Employee benefits	10.2	10.1
Professional Fees: Labor-Related	3.7	3.7
Administrative and facilities support services	0.5	0.5
Installation, Maintenance and Repair Services	0.6	0.6
All Other: Labor Related Services	2.5	2.5
Capital-related (.391)	3.0	2.9
Total	70.8	70.5
1 Published in the Federal Register; based on second quarter 2017 IGI forecast.
2 Based on second quarter 2018 IGI forecast, with historical data through first quarter 2018.

Tables 9 and 10 show the RUG-IV case-mix adjusted federal rates for FY 2019 by labor-related and non-labor-related components. Tables 9 and 10 do not reflect the add-on for SNF residents with AIDS enacted by section 511 of the MMA, which we apply only after making all other adjustments (such as wage index and case-mix). Additionally, Tables 9 and 10 do not reflect adjustments which may be made to the SNF PPS rates as a result of either the SNF QRP, discussed in section VI.B. of this final rule, or the SNF VBP program, discussed in sections III.B.5. and VI.C. of this final rule.

Section 1888(e)(4)(G)(ii) of the Act also requires that we apply this wage index in a manner that does not result in aggregate payments under the SNF PPS that are greater or less than would otherwise be made if the wage adjustment had not been made. For FY 2019 (federal rates effective October 1, 2018), we stated in the proposed rule that we would apply an adjustment to fulfill the budget neutrality requirement. We stated we would meet this requirement by multiplying each of the components of the unadjusted federal rates by a budget neutrality factor equal to the ratio of the weighted average wage adjustment factor for FY 2018 to the weighted average wage adjustment factor for FY 2019. For this calculation, we stated we would use the same FY 2017 claims utilization data for both the numerator and denominator of this ratio. We define the wage adjustment factor used in this calculation as the labor share of the rate component multiplied by the wage index plus the non-labor share of the rate component. We did not receive any comments regarding our proposed budget neutrality factor calculation. Thus, we are finalizing the budget neutrality methodology as proposed. The final budget neutrality factor for FY 2019 is 0.9999. We note that this is different from the budget neutrality factor (1.0002) provided in the FY 2018 SNF PPS proposed rule (83 FR 21031) due to an updated wage index file and updated claims file used to calculate the budget neutrality factor.

As discussed above, we have historically used, and propose to continue using, pre-reclassified IPPS hospital wage data, unadjusted for occupational mix and the rural and imputed floors, as the basis for the SNF wage index. That being said, in the proposed rule, we noted that we have received recurring comments in prior rulemaking (most recently in the FY 2018 SNF PPS final rule (82 FR 36539 through 36541)) regarding the development of a SNF-specific wage index. It has been suggested that we develop a SNF-specific wage index utilizing SNF cost report wage data instead of hospital wage data. We have noted, in response that developing such a wage index would require a resource-intensive audit process similar to that used for IPPS hospital data, to improve the quality of the SNF cost report data in order for it to be used as part of this analysis. This audit process is quite extensive in the case of approximately 3,300 hospitals, and it would be significantly more so in the case of approximately 15,000 SNFs. As discussed previously in this rule, we believe auditing all SNF cost reports, similar to the process used to audit inpatient hospital cost reports for purposes of the IPPS wage index, would place a burden on providers in terms of recordkeeping and completion of the cost report worksheet. We also believe that adopting such an approach would require a significant commitment of resources by CMS and the Medicare Administrative Contractors, potentially far in excess of those required under the IPPS given that there are nearly five times as many SNFs as there are hospitals. Therefore, while we continue to review all available data and contemplate the potential methodological approaches for a SNF-specific wage index in the future, we continue to believe that in the absence of the appropriate SNF-specific wage data, using the pre-reclassified, pre-rural and imputed floor hospital inpatient wage data (without the occupational mix adjustment) is appropriate and reasonable for the SNF PPS.

As an alternative to a SNF-specific wage index, it has also been suggested that we consider adopting certain wage index policies in use under the IPPS, such as geographic reclassification or rural floor. Although we have the authority under section 315 of BIPA to establish a geographic reclassification procedure specific to SNFs under certain conditions, as discussed previously, under BIPA, we cannot adopt a reclassification policy until we have collected the data necessary to establish a SNF-specific wage index. Thus, we cannot adopt a reclassification procedure at this time. With regard to adopting a rural floor policy, as we stated in the FY 2017 SNF PPS final rule (82 FR 36540), MedPAC has recommended eliminating the rural floor policy (which actually sets a floor for urban hospitals) from the calculation of the IPPS wage index (see, for example, Chapter 3 of MedPAC's March 2013 Report to Congress on Medicare Payment Policy, available at http://medpac.gov/​docs/​default-source/​reports/​mar13_​ch03.pdf,, which notes on page 65 that in 2007, MedPAC had “. . . recommended eliminating these special wage index adjustments and adopting a new wage index system to avoid geographic inequities that can occur due to current wage index policies (Medicare Payment Advisory Commission 2007b.”). As we stated in the FY 2017 SNF PPS final rule, if we were to adopt the rural floor under the SNF PPS, we believe that the SNF PPS wage index could become vulnerable to problems similar to those that MedPAC identified in its March 2013 Report to Congress.

Given the perennial nature of these comments and responses on the SNF PPS wage index policy, in the proposed rule (89 FR 21032) we invited further comments on the issues discussed above. Specifically, we requested comment on how a SNF-specific wage index may be developed without creating significant administrative burdens for providers, CMS, or its contractors. Further, we requested comments on specific alternatives we may consider in future rulemaking which could be implemented in advance of, or in lieu of, a SNF-specific wage index. A discussion of the comments we received, along with our responses, appear below.

Comment: One commenter encouraged CMS to continue using hospital wage data when determining the SNF wage index, since it did not have a proposal for how to obtain a SNF-specific wage index in a manner that does not cause burden on providers.

Response: We appreciate the commenter's encouragement to continue using hospital wage data as a proxy for a SNF wage index.

Comment: Several commenters recommend that CMS pursue the establishment of a SNF-specific wage index. These commenters proposed phased-in recommendations to trim hospital wage data (as an interim step), to reflect positions staffed in nursing homes, allow for a reclassification system, account for occupational mix differences between hospitals and each post-acute care (PAC) setting using published BLS data, and apply a rural floor. Further, if determining a SNF wage index using SNF cost report data is too administratively complex, it was recommended that Payroll-based Journal (PBJ) data be used. Finally, the commenters recommended communicating with hospitals through Medicare Learning Network (MLN) transmittals for education and technical support.

Response: We appreciate the commenter's recommendation for collecting SNF cost report wage data to establish a SNF-specific wage index. We note that, consistent with the preceding discussion in this final rule as well as our previous responses to these recurring comments (most recently published in the FY 2018 SNF PPS final rule (82 FR 36540 through 36541)), developing such a wage index would require a resource-intensive audit process similar to that used for IPPS hospital data, to improve the quality of the SNF cost report data in order for it to be used as part of this analysis.

Further, we appreciate these commenters' suggestion that we modify the current hospital wage data used to construct the SNF PPS wage index to reflect the SNF environment more accurately by trimming hospital wage data to reflect positions staffed in nursing homes, weighing it by occupational mix data published by the BLS, and using PBJ data. While we consider whether or not such an approach may constitute an interim step in the process of developing a SNF-specific wage index, we would note that other provider types also use the hospital wage index as the basis for their associated wage index. As such, we believe that such a recommendation should be part of a broader discussion on wage index reform across Medicare payment systems. With regard to the PBJ recommendation, we will pass this comment to our colleagues managing that initiative for further consideration.

With regard to reclassification and rural floor, as discussed above, section 315 of BIPA authorized us to establish a geographic reclassification procedure that is specific to SNFs, only after collecting the data necessary to establish a SNF-specific wage index that is based on data from nursing homes. However, to date this has been infeasible due to the volatility of existing SNF wage data and the significant amount of resources that would be required to improve the quality of that data. Furthermore, we do not believe that using hospital reclassification data would be appropriate as this data is specific to the requesting hospitals and it may or may not apply to a given SNF in a given instance. With regard to implementing a rural floor, we do not believe it would be prudent at this time to adopt such a policy, because MedPAC has recommended eliminating the rural floor policy from the calculation of the IPPS wage index (see, for example, Chapter 3 of MedPAC's March 2013 Report to Congress on Medicare Payment Policy, available at http://medpac.gov/​documents/​reports/​mar13_​entirereport.pdf, which notes on page 65 that in 2007, MedPAC had “. . . recommended eliminating these special wage index adjustments and adopting a new wage index system to avoid geographic inequities that can occur due to current wage index policies (Medicare Payment Advisory Commission 2007b.”) If we adopted the rural floor at this time under the SNF PPS, we believe that, the SNF PPS wage index could become vulnerable to problems similar to those that MedPAC identified in its March 2013 Report to Congress.

While we continue to review all available data and contemplate the potential methodological approaches for a SNF-specific wage index in the future, we continue to believe that in the absence of the appropriate SNF-specific wage data, using the pre-reclassified, pre-rural and imputed floor hospital inpatient wage data (without the occupational mix adjustment) is appropriate and reasonable for the SNF PPS. We believe the commenters' recommendations should be part of a broader discussion of wage index reform across Medicare payment systems. In the event that a SNF-specific wage index is implemented in the future, CMS will provide education and support in a manner it deems appropriate, which may include MLN transmittals. We will continue to monitor closely research efforts surrounding the development of an alternative hospital wage index for the IPPS and the potential impact or influence of that research on the SNF PPS.

5. SNF Value-Based Purchasing Program

Beginning with payment for services furnished on October 1, 2018, section 1888(h) of the Act requires the Secretary to reduce the adjusted Federal per diem rate determined under section 1888(e)(4)(G) of the Act otherwise applicable to a SNF for services furnished during a fiscal year by 2 percent, and to adjust the resulting rate for a SNF by the value-based incentive payment amount earned by the SNF based on the SNF's performance score for that fiscal year under the SNF VBP Program. To implement these requirements, we proposed to add a new paragraph (f) to § 413.337. We did not receive any public comments regarding this proposal. Therefore, we are finalizing the addition of paragraph (f) to § 413.337 as proposed, without modification.

Please see section VI.C. of this final rule for further information regarding the SNF VBP Program, including a discussion of the methodology we will use to make the payment adjustments.

6. Adjusted Rate Computation Example

Using the hypothetical SNF XYZ, Table 11 shows the adjustments made to the federal per diem rates (prior to application of any adjustments under the SNF QRP and SNF VBP programs as discussed above) to compute the provider's actual per diem PPS payment for FY 2019. We derive the Labor and Non-labor columns from Table 9. The wage index used in this example is based on the FY 2019 SNF PPS wage index that appears in Table A available on the CMS website at http://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​WageIndex.html. As illustrated in Table 11, SNF XYZ's total PPS payment for FY 2019 would equal $48,779.14.

Table 11—Adjusted Rate Computation Example SNF XYZ: Located in Frederick, MD (Urban CBSA 43524), Wage Index: 0.9880

[See wage index in Table A] ¹

RUG-IV group	Labor	Wage index	Adjusted labor	Non-labor	Adjusted rate	Percent adjustment	Medicare days	Payment
RVX	$522.48	0.9880	$516.21	$218.62	$734.83	$734.83	14	$10,287.62
ES2	419.50	0.9880	414.47	175.54	590.01	590.01	30	17,700.30
RHA	263.59	0.9880	260.43	110.29	370.72	370.72	16	5,931.52
CC2 2	242.99	0.9880	240.07	101.67	341.74	779.17	10	7,791.70
BA2	167.52	0.9880	165.51	70.09	235.60	235.60	30	7,068.00
							100	48,779.14
1 Available on the CMS website at http://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​WageIndex.html.
2 Reflects a 128 percent adjustment from section 511 of the MMA.

IV. Additional Aspects of the SNF PPS

A. SNF Level of Care—Administrative Presumption

The establishment of the SNF PPS did not change Medicare's fundamental requirements for SNF coverage. However, because the case-mix classification is based, in part, on the beneficiary's need for skilled nursing care and therapy, we have attempted, where possible, to coordinate claims review procedures with the existing resident assessment process and case-mix classification system discussed in section III.B.3. of this final rule. This approach includes an administrative presumption that utilizes a beneficiary's initial classification in one of the upper 52 RUGs of the current 66-group RUG-IV case-mix classification system to assist in making certain SNF level of care determinations.

In accordance with the regulations at § 413.345, we include in each update of the federal payment rates in the Federal Register a discussion of the resident classification system that provides the basis for case-mix adjustment. Under that discussion, we designate those specific classifiers under the case-mix classification system that represent the required SNF level of care, as provided in § 409.30. As set forth in the FY 2011 SNF PPS update notice (75 FR 42910), this designation reflects an administrative presumption under the 66-group RUG-IV system that beneficiaries who are correctly assigned to one of the upper 52 RUG-IV groups on the initial 5-day, Medicare-required assessment are automatically classified as meeting the SNF level of care definition up to and including the assessment reference date (ARD) on the 5-day Medicare-required assessment.

A beneficiary assigned to any of the lower 14 RUG-IV groups is not automatically classified as either meeting or not meeting the definition, but instead receives an individual level of care determination using the existing administrative criteria. This presumption recognizes the strong likelihood that beneficiaries assigned to one of the upper 52 RUG-IV groups during the immediate post-hospital period require a covered level of care, which would be less likely for those beneficiaries assigned to one of the lower 14 RUG-IV groups.

In the July 30, 1999 final rule (64 FR 41670), we indicated that we would announce any changes to the guidelines for Medicare level of care determinations related to modifications in the case-mix classification structure. The FY 2018 final rule (82 FR 36544) further specified that we would henceforth disseminate the standard description of the administrative presumption's designated groups via the SNF PPS website at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​index.html (where such designations appear in the paragraph entitled “Case Mix Adjustment”), and would publish such designations in rulemaking only to the extent that we actually intend to make changes in them. (We discuss in section V.G. of this final rule the modifications to the administrative level of care presumption that we are finalizing in order to accommodate the PDPM case-mix classification system.)

However, we note that this administrative presumption policy does not supersede the SNF's responsibility to ensure that its decisions relating to level of care are appropriate and timely, including a review to confirm that the services prompting the assignment of one of the designated case-mix classifiers (which, in turn, serves to trigger the administrative presumption) are themselves medically necessary. As we explained in the FY 2000 SNF PPS final rule (64 FR 41667), the administrative presumption is itself rebuttable in those individual cases in which the services actually received by the resident do not meet the basic statutory criterion of being reasonable and necessary to diagnose or treat a beneficiary's condition (according to section 1862(a)(1) of the Act). Accordingly, the presumption would not apply, for example, in those situations in which a resident's assignment to one of the upper groups is itself based on the receipt of services that are subsequently determined to be not reasonable and necessary. Moreover, we want to stress the importance of careful monitoring for changes in each patient's condition to determine the continuing need for Part A SNF benefits after the ARD of the 5-day assessment.

B. Consolidated Billing

Sections 1842(b)(6)(E) and 1862(a)(18) of the Act (as added by section 4432(b) of the BBA 1997) require a SNF to submit consolidated Medicare bills to its Medicare Administrative Contractor (MAC) for almost all of the services that its residents receive during the course of a covered Part A stay. In addition, section 1862(a)(18) of the Act places the responsibility with the SNF for billing Medicare for physical therapy, occupational therapy, and speech-language pathology services that the resident receives during a noncovered stay. (Please refer to section VI.A. of this final rule for a discussion of a revision to the regulation text that describes a beneficiary's status as a SNF “resident” for consolidated billing purposes.) Section 1888(e)(2)(A) of the Act excludes a small list of services from the consolidated billing provision (primarily those services furnished by physicians and certain other types of practitioners), which remain separately billable under Part B when furnished to a SNF's Part A resident. These excluded service categories are discussed in greater detail in section V.B.2. of the May 12, 1998 interim final rule (63 FR 26295 through 26297).

A detailed discussion of the legislative history of the consolidated billing provision is available on the SNF PPS website at http://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​Downloads/​Legislative_​History_​04152015.pdf. In particular, section 103 of the Medicare, Medicaid, and SCHIP Balanced Budget Refinement Act of 1999 (BBRA, Pub. L. 106-113, enacted on November 29, 1999) amended section 1888(e)(2)(A) of the Act by further excluding a number of individual high-cost, low probability services, identified by Healthcare Common Procedure Coding System (HCPCS) codes, within several broader categories (chemotherapy items, chemotherapy administration services, radioisotope services, and customized prosthetic devices) that otherwise remained subject to the provision. We discuss this BBRA amendment in greater detail in the SNF PPS proposed and final rules for FY 2001 (65 FR 19231 through 19232, April 10, 2000, and 65 FR 46790 through 46795, July 31, 2000), as well as in Program Memorandum AB-00-18 (Change Request #1070), issued March 2000, which is available online at www.cms.gov/​transmittals/​downloads/​ab001860.pdf.

As explained in the FY 2001 proposed rule (65 FR 19232), the amendments enacted in section 103 of the BBRA not only identified for exclusion from this provision a number of particular service codes within four specified categories (that is, chemotherapy items, chemotherapy administration services, radioisotope services, and customized prosthetic devices), but also gave the Secretary the authority to designate additional, individual services for exclusion within each of the specified service categories. In the proposed rule for FY 2001, we also noted that the BBRA Conference report (H.R. Rep. No. 106-479 at 854 (1999) (Conf. Rep.)) characterizes the individual services that this legislation targets for exclusion as high-cost, low probability events that could have devastating financial impacts because their costs far exceed the payment SNFs receive under the PPS. According to the conferees, section 103(a) of the BBRA is an attempt to exclude from the PPS certain services and costly items that are provided infrequently in SNFs. By contrast, the amendments enacted in section 103 of the BBRA do not designate for exclusion any of the remaining services within those four categories (thus, leaving all of those services subject to SNF consolidated billing), because they are relatively inexpensive and are furnished routinely in SNFs.

As we further explained in the final rule for FY 2001 (65 FR 46790), and as is consistent with our longstanding policy, any additional service codes that we might designate for exclusion under our discretionary authority must meet the same statutory criteria used in identifying the original codes excluded from consolidated billing under section 103(a) of the BBRA: they must fall within one of the four service categories specified in the BBRA; and they also must meet the same standards of high cost and low probability in the SNF setting, as discussed in the BBRA Conference report. Accordingly, we characterized this statutory authority to identify additional service codes for exclusion as essentially affording the flexibility to revise the list of excluded codes in response to changes of major significance that may occur over time (for example, the development of new medical technologies or other advances in the state of medical practice) (65 FR 46791). In the proposed rule (83 FR 21033), we specifically invited public comments identifying HCPCS codes in any of these four service categories (chemotherapy items, chemotherapy administration services, radioisotope services, and customized prosthetic devices) representing recent medical advances that might meet our criteria for exclusion from SNF consolidated billing. We stated that we may consider excluding a particular service if it met our criteria for exclusion as specified above. We further stated that commenters should identify in their comments the specific HCPCS code that is associated with the service in question, as well as their rationale for requesting that the identified HCPCS code(s) be excluded.

We note that the original BBRA amendment (as well as the implementing regulations) identified a set of excluded services by means of specifying HCPCS codes that were in effect as of a particular date (in that case, as of July 1, 1999). Identifying the excluded services in this manner made it possible for us to utilize program issuances as the vehicle for accomplishing routine updates of the excluded codes, to reflect any minor revisions that might subsequently occur in the coding system itself (for example, the assignment of a different code number to the same service). Accordingly, we stated in the proposed rule that, in the event that we identify through the current rulemaking cycle any new services that would actually represent a substantive change in the scope of the exclusions from SNF consolidated billing, we would identify these additional excluded services by means of the HCPCS codes that are in effect as of a specific date (in this case, as of October 1, 2018). By making any new exclusions in this manner, we could similarly accomplish routine future updates of these additional codes through the issuance of program instructions.

Commenters submitted the following comments related to the proposed rule's discussion of the consolidated billing aspects of the SNF PPS. A discussion of these comments, along with our responses, appears below.

Comment: Some commenters reiterated previous recommendations regarding the exclusion of certain drugs from consolidated billing that had been submitted and addressed repeatedly in a number of prior rulemaking cycles. One such recommendation involved excluding the commonly used prostate cancer drug Lupron® (leuprolide acetate). Other commenters once again raised the issue of nursing home residents bringing their own medications, as a means of minimizing the nursing home's cost of caring for the resident. Still another reiterated previous recommendations to exclude a broader range of expensive drugs beyond the category of chemotherapy alone, citing anecdotal evidence that leaving such drugs within the SNF PPS bundle may, among other things, create a disincentive for admitting those patients who require them.

Response: For the reasons discussed previously in prior rulemaking, the particular drugs cited in these comments remain subject to consolidated billing. In the case of leuprolide acetate, the most recent discussion appears in the SNF PPS final rule for FY 2015 (79 FR 45642, August 5, 2014), which explained that this drug is unlikely to meet the criterion of “low probability” specified in the BBRA. Regarding the issue of nursing homes having residents supply their own medications, the SNF PPS final rule for FY 2018 (82 FR 36548, August 4, 2017) explained that the applicable terms of the SNF's provider agreement would preclude this practice, in that they require the SNF to accept Medicare's payment for covered services as payment in full. Finally, the issue of establishing a broader exclusion that would encompass expensive non-chemotherapy drugs was addressed in the SNF PPS final rule for FY 2017 (81 FR 51985, August 5, 2016), which explained that existing law does not provide for such an expansion. In addition, it is worth noting in this context that in accounting more accurately for the costs of NTA services such as drugs, the PDPM model has the potential to ameliorate some of the concerns cited in these comments.

Comment: One commenter urged us to expand the scope of the chemotherapy exclusion, advancing an interpretation of the Secretary's authority under section 1888(e)(2)(A)(iii)(II) of the Act to designate “additional” chemotherapy items for exclusion as not actually being restricted to the types of “high-cost, low probability” chemotherapy items and services described elsewhere in that provision, and further suggesting that identifying a given item or service as either “high-cost” or “low probability” alone should be sufficient grounds for its exclusion. The commenter also submitted well over 100 codes that it suggested should be added to the chemotherapy portion of the exclusion list. The commenter reiterated previous recommendations to expand the existing chemotherapy exclusion to encompass related drugs such as anti-emetics (anti-nausea drugs)—which, while they do not in themselves fight cancer, are commonly administered along with the chemotherapy drug to ameliorate its side effects. While we have, in fact, already addressed such recommendations repeatedly in previous rulemaking (most recently, in the FY 2015 SNF PPS final rule (79 FR 45642, August 5, 2014)), the commenter cited in further support of its position the similarity between the recommended approach and the existing policy under the dialysis exclusion from consolidated billing, in which the exclusion encompasses related services along with the dialysis itself. In addition, the commenter reiterated previous concerns about the complexity of the existing set of consolidated billing exclusions, suggesting that it should be streamlined and simplified.

Response: Approximately two-thirds of the codes that the commenter submitted already appear on the chemotherapy exclusion list. Of the remaining codes, several were already in existence in 1999 when the BBRA enacted the statutory ranges of excluded codes, but were skipped over by those ranges; as discussed repeatedly in previous rulemaking—most recently, in the FY 2018 SNF PPS final rule (82 FR 36547, August 4, 2017)—this action indicates that such drugs were intended to remain within the SNF PPS bundle, subject to the BBRA Conference Report's provision for a GAO review of the code set that was conducted the following year. Still others were codes such as those for anti-emetic (anti-nausea) drugs, which serve to address the chemotherapy drug's side effects rather than actually fighting the cancer itself; as we have noted repeatedly in prior rulemaking (most recently, in the FY 2015 SNF PPS final rule, 79 FR 45642, August 5, 2014), such drugs do not, in fact, represent “chemotherapy” (that is, cancer-fighting) drugs within the meaning of this exclusion. Further, the commenter's proposed interpretation suggesting that the exclusion is not restricted to “high-cost, low probability” chemotherapy services, or that a given chemotherapy service need only be either “high-cost” or “low probability” alone in order to qualify for exclusion would not be consistent with Congress' stated intent with respect to this provision. In fact, in the above-cited BBRA Conference Report (H.R. Rep. 106-479 at 854 (1999) (Conf. Rep.)), the Congress clearly specified the overall purpose of this provision: “This provision is an attempt to exclude from the PPS certain services and costly items that are provided infrequently in SNFs” (emphasis added); thus, any “additional” chemotherapy services that the Secretary might designate for exclusion under this authority, like those already excluded, would remain subject to both the “high-cost” and “low probability” thresholds. Regarding the commenter's further suggestion that the dialysis exclusion might serve as a possible precedent for broadening the chemotherapy exclusion to include related services, we note that as one of the BBA 1997's original set of consolidated billing exclusions enacted in clause (ii) of section 1888(e)(2)(A) of the Act, the dialysis exclusion fundamentally differs from the BBRA's subsequent, more targeted set of exclusions in clause (iii) of that section (such as the one for chemotherapy) in that the BBA 1997 excluded entire service categories from consolidated billing, whereas the BBRA focused more narrowly on excluding certain designated “high-cost, low probability” services, identified by HCPCS code, within several broader categories that otherwise remained subject to the provision. In the FY 2015 SNF PPS final rule (79 FR 45644, August 5, 2014), we specifically contrasted the relatively broad exclusions enacted in the BBA 1997 with the more narrowly targeted BBRA exclusions, in the context of another one of the latter exclusions that involves radioisotope services. In that context, we noted that the statutory exclusion for “radioisotope services” at section 1888(e)(2)(A)(iii)(IV) of the Act stands in marked contrast, for example, to the ones for dialysis and erythropoietin (EPO) at section 1888(e)(2)(A)(ii) of the Act, which consist of—and, in fact, are defined by—explicit cross-references to the corresponding Part B benefit categories appearing in sections 1861(s)(2)(F) and 1861(s)(2)(O) of the Act, respectively. Under this framework, the scope of the consolidated billing provision's dialysis exclusion is effectively defined by the scope of coverage under the separate Part B dialysis benefit at section 1861(s)(2)(F) of the Act, which would encompass dialysis-related services along with the dialysis itself. By contrast, the more targeted BBRA exclusions in areas such as chemotherapy and radioisotope services focus specifically on those particular services that are directly designated as in themselves meeting the applicable criteria for exclusion.

Finally, regarding the comment about the complexity of this provision, in the FY 2010 SNF PPS final rule (74 FR 40355, August 11, 2009), we noted in response to a previous, similar comment that while the commenter's interest in promoting improved ease of administration is understandable, current law contains no authority to effect a comprehensive overhaul of the existing requirements administratively. However, we would also note in this context that we continue to conduct an active educational and training initiative on the consolidated billing provision that includes the following:

• A recently updated and expanded set of consolidated billing instructions in Chapter 6 of the Medicare Claims Processing Manual (available online at https://www.cms.gov/​Regulations-and-Guidance/​Guidance/​Manuals/​Downloads/​clm104c06.pdf), at §§ 10-20.6;
• Addressing questions that arise on this topic during CMS's recurring nationwide SNF/Long-Term Care Open Door Forums (https://www.cms.gov/​Outreach-and-Education/​Outreach/​OpenDoorForums/​ODF_​SNFLTC.html);
• Development of sample model agreements between SNFs and their suppliers, which are posted online for review at our “Best Practices” website (at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​BestPractices.html); and
• Creation of a web-based training (WBT) module accessible from the Medicare Learning Network (MLN) website at https://www.cms.gov/​Outreach-and-Education/​Medicare-Learning-Network-MLN/​MLNProducts/​WebBasedTraining.html, which offers interactive online training on consolidated billing.

Comment: One commenter recommended for exclusion a pair of oral chemotherapy drugs, ZYTIGA® (abiraterone acetate) and ERLEADA® (apalutamide), which are used in treating certain uncommon and otherwise resistant forms of prostate cancer. The commenter acknowledged our previous discussion of oral drugs in this context in the FY 2017 SNF PPS final rule (81 FR 51985, August 5, 2016), which described them as not reasonably characterized as meeting the BBRA's chemotherapy exclusion criterion of “requiring special staff expertise to administer.” However, the commenter then went on to point out that the accompanying Conference Report language (H. Conf. Rep. No. 106-479 at 854), in discussing the statutory exclusion of “high-cost, low probability” chemotherapy items, lists as examples those drugs that “. . . are not typically administered in a SNF, or are exceptionally expensive, or are given as infusions, thus requiring special staff expertise to administer” (emphasis added). Thus, the commenter suggested that while the Conference Report language itself specifies “high-cost, low probability” as the applicable standard for the chemotherapy exclusion, its use of the word “or” in the specific context of “requiring special staff expertise to administer” identifies this particular criterion as merely an illustrative example that is not an absolute prerequisite for meeting the standard in all cases. The commenter also acknowledged our explanation in the FY 2015 SNF PPS final rule (79 FR 45642, August 5, 2014) in connection with a previous comment regarding ZYTIGA® and another oral chemotherapy drug, REVLIMID® (lenalidomide), that it would not be operationally feasible to utilize a miscellaneous “not otherwise specified” (NOS) code such as J8999 to effect such an exclusion, and then urged us to consider other options, such as establishing a separate code or modifier for the particular drugs in question, or utilizing the already-existing National Drug Codes (NDCs) that are specific to those drugs. Other commenters similarly recommended the oral chemotherapy drugs REVLIMID® and GLEEVEC® (imatinib mesylate) for exclusion.

Response: We believe that the commenter's point that an oral chemotherapy drug which does not require “special staff expertise to administer” can nonetheless qualify for exclusion as long as it can otherwise meet the “high-cost, low probability” standard merits further consideration. However, we note that the four oral chemotherapy drugs at issue here differ from previously-excluded, non-oral chemotherapy drugs in that they are not covered under Part B. We note that while Part B would authorize coverage for drugs (including those chemotherapy drugs that are excluded from consolidated billing under section 1888(e)(2)(A)(iii)(II) of the Act) as either an incident to a physician's professional services (under section 1861(s)(2)(A) of the Act) or as an outpatient hospital service (under section 1861(s)(2)(B) of the Act), this authority is specifically limited in both cases to those drugs “that are not usually self-administered by the patient,” thus effectively excluding oral drugs as a class. Further, while Part B does, in fact, include a specific benefit category for oral chemotherapy drugs (at section 1861(s)(2)(Q) of the Act), that benefit is restricted to those with the same indication and active ingredient(s) as a covered non-oral anti-cancer drug, which is not the case for the specific four drugs in question.

Because the drugs at issue here would not be covered under Part B, we believe that the applicable provisions at section 1888(e)(2)(A) may not provide a basis for excluding them from consolidated billing. Accordingly, because of the need for further consideration of this issue, we are unable to adopt the commenters' recommendations at this time.

Comment: A few commenters reiterated previous recommendations to expand the existing exclusion for certain high-intensity outpatient hospital services to encompass non-hospital settings as well.

Response: Similar concerns have been raised and addressed repeatedly in prior rulemaking (most recently, in the FY 2018 SNF PPS final rule (82 FR 36547, August 4, 2017)), as follows:

• As noted in numerous previous rules, as well as in Medicare Learning Network (MLN) Matters article SE0432 (available online at https://www.cms.gov/​Outreach-and-Education/​Medicare-Learning-Network-MLN/​MLNMattersArticles/​Downloads/​SE0432.pdf), the rationale for establishing this exclusion was to address those types of services that are so far beyond the normal scope of SNF care that they require the intensity of the hospital setting in order to be furnished safely and effectively. Moreover, when the Congress enacted the consolidated billing exclusion for certain RHC and FQHC services in section 410 of the MMA, the accompanying legislative history's description of present law directly acknowledged the hospital-specific nature of this exclusion.
• Ever since its inception, this exclusion was intended to be hospital-specific; as explained in the original SNF PPS interim final rule (63 FR 26298, May 12, 1998), this exclusion was created within the context of the concurrent development of a new PPS specifically for outpatient hospital services, reflecting the need to delineate the respective areas of responsibility for the SNF under the consolidated billing provision, and for the hospital under the outpatient bundling provision, with regard to these services. This point was further reinforced in the subsequent final rule for FY 2000 (64 FR 41676, July 30, 1999), in which we explained that a key concern underlying the development of the consolidated billing exclusion of certain outpatient hospital services specifically involved the need to distinguish those services that comprise the SNF bundle from those that will become part of the outpatient hospital bundle that is currently being developed in connection with the outpatient hospital PPS. Accordingly, we noted at that time that we would not be extending the outpatient hospital exclusion from consolidated billing to encompass any other, freestanding settings.
• Finally, the FY 2010 final rule (74 FR 40355, August 11, 2009), while acknowledging that advances in medical technology over time may make it feasible to perform such high-intensity outpatient services more widely in nonhospital settings, then went on to cite the FY 2006 final rule (70 FR 45049, August 4, 2005) in noting that such a development would not argue in favor of excluding the nonhospital performance of the service from consolidated billing, but rather, would call into question whether the service should continue to be excluded from consolidated billing at all, even when performed in the hospital setting.

Comment: Several commenters reiterated comments submitted previously during the FY 2016 rulemaking cycle in the context of the SNF VBP provision, in which they had sought to portray a portable x-ray service's transportation and setup as a separately billable “physician” service by suggesting that such activities should appropriately be regarded as part of the diagnostic test's professional component (PC) for interpreting the test results rather than the technical component (TC) for performing the test itself. They now reiterated those same comments in the context of the PDPM, and additionally indicated that allowing these services to be paid separately outside of the Part A bundle would be consistent with the proration policy that applies under Part B when a single portable x-ray visit serves multiple patients, under which the trip itself is allocated among all of the patients served (regardless of payment source) in order to calculate the prorated payment amount that applies specifically to each of the Part B patients. Some of the commenters also cited certain HCPCS codes, such as R0076 (“transportation of portable EKG to facility or location, per patient”), and suggested that all of the “medical and other health services” enumerated in section 1861(s) of the Act—including the diagnostic test benefit at section 1861(s)(3) of the Act—should be regarded as excluded “physician” services.

Response: As we explained previously in the FY 2016 SNF PPS final rule (80 FR 46408, August 4, 2015), we do not share the view of those commenters who would categorize a portable x-ray service's transportation and setup as part of the separately billable PC. In that discussion, we cited § 90.5 of the Medicare Claims Processing Manual, Chapter 13 (available online at https://www.cms.gov/​Regulations-and-Guidance/​Guidance/​Manuals/​Downloads/​clm104c13.pdf), which states that the bundled TC (to which consolidated billing applies) specifically includes “any associated transportation and setup costs.” As indicated in the FY 2016 SNF PPS final rule (80 FR 46408, August 4, 2015), to be considered a separately billable “physician” service in this context, a given service must not only be furnished personally by a physician, but must actually be a type of service that ordinarily requires such performance; we further noted that a portable x-ray service's transportation and setup would never meet these criteria, as the service's excluded PC relates solely to reading the x-ray rather than taking it, and the physician's personal performance clearly would not be required for activities such as driving the supplier's vehicle to the SNF, or setting up the equipment once it arrives there.

Further, we believe the comments that cited the proration policy in this context (which involves a single portable x-ray trip that serves multiple patients) may reflect a certain amount of misunderstanding about the proration policy's actual nature and purpose. As explained in the Medicare Physician Fee Schedule (MPFS) final rule for calendar year (CY) 2016 (80 FR 70886, November 16, 2015), the reason for allocating such a trip among all of the patients served is to ensure that Medicare Part B should not pay for more than its share of the transportation costs for portable x-ray services (80 FR 71068 through 71069). However, while all of the patients served (both the Part B and non-Part B patients) would be included in calculating the proration itself, the resulting prorated amount would be payable only for the Part B patients. By contrast, for any Part A SNF residents served by the same trip, the transportation cost associated with the portable x-ray service would be subsumed in the SNF's payment to the supplier for the bundled TC, as discussed above. In terms of Part A payment, that bundled TC, in turn, would be included (along with all other bundled services) within the global PPS per diem that the SNF receives for the covered Part A stay itself. Moreover, the SNF's actual payment amount to its supplier in this scenario would not be tied to the prorated payment amount made for the Part B patients served on the same trip; as explained in § 70.4 of the Medicare Benefit Policy Manual, Chapter 8 (available online at https://www.cms.gov/​Regulations-and-Guidance/​Guidance/​Manuals/​Downloads/​bp102c08.pdf), for a bundled service, the specific details of the ensuing payment arrangement between the SNF and the outside supplier (such as the actual payment amount and timeframe) represent a private, “marketplace” transaction that is negotiated between the parties themselves.

Regarding the suggestion that all of the Part B “medical and other health services” specified in section 1861(s) of the Act (including diagnostic tests such as portable x-ray services) should be considered physician services, we note that the physician services benefit at section 1861(s)(1) of the Act actually represents only a small subset of the overall “medical and other health services” enumerated throughout section 1861(s) of the Act, and that the diagnostic test benefit at section 1861(s)(3) of the Act (which would encompass the TC for a portable x-ray service) is, in fact, a separate and distinct benefit category from the one at section 1861(s)(1) of the Act for physician services. Finally, regarding the comments on certain HCPCS codes, we acknowledge that among the various consolidated billing exclusions listed in section 1888(e)(2)(A)(ii) of the Act are “transportation costs of electrocardiogram equipment for electrocardiogram test services (HCPCS code R0076).” However, that portion of the law additionally specifies that this particular exclusion is in effect “only with respect to services furnished during 1998;” accordingly, the statutory exclusion for these particular services has long since expired.

C. Payment for SNF-Level Swing-Bed Services

Section 1883 of the Act permits certain small, rural hospitals to enter into a Medicare swing-bed agreement, under which the hospital can use its beds to provide either acute- or SNF-level care, as needed. For critical access hospitals (CAHs), Part A pays on a reasonable cost basis for SNF-level services furnished under a swing-bed agreement. However, in accordance with section 1888(e)(7) of the Act, SNF-level services furnished by non-CAH rural hospitals are paid under the SNF PPS, effective with cost reporting periods beginning on or after July 1, 2002. As explained in the FY 2002 final rule (66 FR 39562), this effective date is consistent with the statutory provision to integrate swing-bed rural hospitals into the SNF PPS by the end of the transition period, June 30, 2002.

Accordingly, all non-CAH swing-bed rural hospitals have now come under the SNF PPS. Therefore, all rates and wage indexes outlined in earlier sections of this final rule for the SNF PPS also apply to all non-CAH swing-bed rural hospitals. A complete discussion of assessment schedules, the MDS, and the transmission software (RAVEN-SB for Swing Beds) appears in the FY 2002 final rule (66 FR 39562) and in the FY 2010 final rule (74 FR 40288). As finalized in the FY 2010 SNF PPS final rule (74 FR 40356 through 40357), effective October 1, 2010, non-CAH swing-bed rural hospitals are required to complete an MDS 3.0 swing-bed assessment which is limited to the required demographic, payment, and quality items. The latest changes in the MDS for swing-bed rural hospitals appear on the SNF PPS website at http://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​index.html. We refer readers to section V.D.2. of this final rule for a discussion of the revisions we are finalizing to the MDS 3.0 swing-bed assessment effective October 1, 2019.

V. Revisions to SNF PPS Case-Mix Classification Methodology

A. Background and General Comments

In the FY 2019 SNF PPS proposed rule, we discussed our proposed changes to the SNF PPS, specifically the proposed comprehensive revisions to the SNF PPS case-mix classification system whereby we proposed to replace the current RUG-IV system with the Patient Driven Payment Model (PDPM) effective October 1, 2019. In section V.A of the FY 2019 SNF PPS proposed rule (83 FR 21034-21036), we discuss the basis for the proposed PDPM and our reasons for proposing to replace the existing case-mix classification system with the PDPM, effective October 1, 2019.

Section 1888(e)(4)(G)(i) of the Act requires the Secretary to make an adjustment to the per diem rates to account for case-mix. The statute specifies that the adjustment is to be based on both a resident classification system that the Secretary establishes that accounts for the relative resource use of different resident types, as well as resident assessment and other data that the Secretary considers appropriate.

In general, the case-mix classification system currently used under the SNF PPS classifies residents into payment classification groups, called RUGs, based on various resident characteristics and the type and intensity of therapy services provided to the resident. Under the existing SNF PPS methodology, there are two case-mix-adjusted components of payment: Nursing and therapy. Each RUG is assigned a CMI for each payment component to reflect relative differences in cost and resource intensity. The higher the CMI, the higher the expected resource utilization and cost associated with residents assigned to that RUG. The case-mix-adjusted nursing component of payment reflects relative differences in a resident's associated nursing and non-therapy ancillary (NTA) costs, based on various resident characteristics, such as resident comorbidities, and treatments. The case-mix-adjusted therapy component of payment reflects relative differences in a resident's associated therapy costs, which is based on a combination of PT, OT, and SLP services. Resident classification under the existing therapy component is based primarily on the amount of therapy the SNF chooses to provide to a SNF resident. Under the RUG-IV model, residents are classified into rehabilitation groups, where payment is determined primarily based on the intensity of therapy services received by the resident, and into nursing groups, based on the intensity of nursing services received by the resident and other aspects of the resident's care and condition. However, only the higher paying of these groups is used for payment purposes. For example, if a resident is classified into a both the RUA (Rehabilitation) and PA1 (Nursing) RUG-IV groups, where RUA has a higher per-diem payment rate than PA1, the RUA group is used for payment purposes. It should be noted that the vast majority of Part A covered SNF days (over 90 percent) are paid using a rehabilitation RUG. A variety of concerns have been raised with the current SNF PPS, specifically the RUG-IV model, which we discuss below.

When the SNF PPS was first implemented in 1998 (63 FR 26252), we developed the RUG-III case-mix classification model, which tied the amount of payment to resident resource use in combination with resident characteristic information. Staff time measurement (STM) studies conducted in 1990, 1995, and 1997 provided information on resource use (time spent by staff members on residents) and resident characteristics that enabled us not only to establish RUG-III but also to create CMIs. This initial RUG-III model was refined by changes finalized in the FY 2006 SNF PPS final rule (70 FR 45032), which included adding nine case-mix groups to the top of the original 44-group RUG-III hierarchy, which created the RUG-53 case-mix model.

In the FY 2010 SNF PPS proposed rule (74 FR 22208), we proposed the RUG-IV model based on, among other reasons, concerns that incentives in the SNF PPS had changed the relative amount of nursing resources required to treat SNF residents (74 FR 22220). These concerns led us to conduct a new Staff Time Measurement (STM) study, the Staff Time and Resource Intensity Verification (STRIVE) project, which served as the basis for developing the current SNF PPS case-mix classification model, RUG-IV, which became effective in FY 2011. At that time, we considered alternative case mix models, including predictive models of therapy payment based on resident characteristics; however, we had a great deal of concern that by separating payment from the actual provision of services, the system, and more importantly, the beneficiaries would be vulnerable to underutilization (74 FR 22220). Other options considered at the time included a non-therapy ancillary (NTA) payment model based on resident characteristics (74 FR 22238) and a DRG-based payment model that relied on information from the prior inpatient stay (74 FR 22220); these and other options are discussed in detail in a CMS Report to Congress issued in December 2006 (available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​Downloads/​RC_​2006_​PC-PPSSNF.pdf).

As we explained in the proposed rule (83 FR 21034), in the years since we implemented the SNF PPS, finalized RUG-IV, and made statements regarding our concerns about underutilization of services in previously considered models, we have witnessed a significant trend that has caused us to reconsider these concerns. More specifically, as discussed in section V.E. of the FY 2015 SNF PPS proposed rule (79 FR 25767), we documented and discussed trends observed in therapy utilization in a memo entitled “Observations on Therapy Utilization Trends” (which may be accessed at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​Downloads/​Therapy_​Trends_​Memo_​04212014.pdf). The two most notable trends discussed in that memo were that the percentage of residents classifying into the Ultra-High therapy category has increased steadily and, of greater concern, that the percentage of residents receiving just enough therapy to surpass the Ultra-High and Very-High therapy thresholds has also increased. In that memo, we state “the percentage of claims-matched MDS assessments in the range of 720 minutes to 739 minutes, which is just enough to surpass the 720 minute threshold for RU groups, has increased from 5 percent in FY 2005 to 33 percent in FY 2013” and this trend has continued since that time. We stated in the proposed rule (83 FR 21035) that while it might be possible to attribute the increasing share of residents in the Ultra-High therapy category to increasing acuity within the SNF population, we believe the increase in “thresholding” (that is, of providing just enough therapy for residents to surpass the relevant therapy thresholds) is a strong indication of service provision predicated on financial considerations rather than resident need. We discussed this issue in response to comments in the FY 2015 SNF PPS final rule. In that rule, in response to comments regarding the lack of “current medical evidence related to how much therapy a given resident should receive,” we stated that with regard to the comments which highlight the lack of existing medical evidence for how much therapy a given resident should receive, we would note that the number of therapy minutes provided to SNF residents within certain therapy RUG categories is, in fact, clustered around the minimum thresholds for a given therapy RUG category. We further stated that given the comments highlighting the lack of medical evidence related to the appropriate amount of therapy in a given situation, it is all the more concerning that practice patterns would appear to be as homogenized as the data would suggest. (79 FR 45651).

In response to comments which highlighted potential explanatory factors for the observed trends, such as internal pressure within SNFs that would override clinical judgment, we stated that we found these potential explanatory factors troubling and entirely inconsistent with the intended use of the SNF benefit. Specifically, the minimum therapy minute thresholds for each therapy RUG category are certainly not intended as ceilings or targets for therapy provision. As discussed in Chapter 8, Section 30 of the Medicare Benefit Policy Manual (Pub. 100-02), to be covered, the services provided to a SNF resident must be “reasonable and necessary for the treatment of a patient's illness or injury, that is, are consistent with the nature and severity of the individual's illness or injury, the individual's particular medical needs, and accepted standards of medical practice.” Therefore, we stated that services which are not specifically tailored to meet the individualized needs and goals of the resident, based on the resident's condition and the evaluation and judgment of the resident's clinicians, may not meet this aspect of the definition for covered SNF care, and we stated we believe that internal provider rules should not seek to circumvent the Medicare statute, regulations and policies, or the professional judgment of clinicians. (79 FR 45651 through 45652).

In addition to this discussion of observed trends, we noted in the proposed rule (83 FR 21035) that others have also identified potential areas of concern within the current SNF PPS. The two most notable sources are the Office of the Inspector General (OIG) and the Medicare Payment Advisory Commission (MedPAC).

For the OIG, three recent OIG reports describe the OIG's concerns with the current SNF PPS. In December 2010, the OIG released a report entitled “Questionable Billing by Skilled Nursing Facilities” (which may be accessed at https://oig.hhs.gov/​oei/​reports/​oei-02-09-00202.pdf). In this report, among its findings, the OIG found that “from 2006 to 2008, SNFs increasingly billed for higher paying RUGs, even though beneficiary characteristics remained largely unchanged” (OEI-02-09-00202, ii), and among other things, recommended that we should “consider several options to ensure that the amount of therapy paid for by Medicare accurately reflects beneficiaries' needs” (OEI-02-09-00202, iii). Further, in November 2012, the OIG released a report entitled “Inappropriate Payments to Skilled Nursing Facilities Cost Medicare More Than a Billion Dollars in 2009” (which may be accessed at https://oig.hhs.gov/​oei/​reports/​oei-02-09-00200.pdf). In this report, the OIG found that “SNFs billed one-quarter of all claims in error in 2009” and that the “majority of the claims in error were upcoded; many of these claims were for ultrahigh therapy.” (OEI-02-09-00200, Executive Summary). Among its recommendations, the OIG stated that “the findings of this report provide further evidence that CMS needs to change how it pays for therapy” (OEI-02-09-00200, 15). Finally, in September 2015, the OIG released a report entitled “The Medicare Payment System for Skilled Nursing Facilities Needs to be Reevaluated” (which may be accessed at https://oig.hhs.gov/​oei/​reports/​oei-02-13-00610.pdf). Among its findings, the OIG found that “Medicare payments for therapy greatly exceed SNFs' costs for therapy,” further noting that “the difference between Medicare payments and SNFs' costs for therapy, combined with the current payment method, creates an incentive for SNFs to bill for higher levels of therapy than necessary” (OEI-02-13-00610, 7). Among its recommendations, the OIG stated that CMS should “change the method of paying for therapy”, further stating that “CMS should accelerate its efforts to develop and implement a new method of paying for therapy that relies on beneficiary characteristics or care needs.” (OEI-02-13-00610, 12).

For MedPAC's recommendations in this area, Chapter 8 of MedPAC's March 2017 Report to Congress (available at http://www.medpac.gov/​docs/​default-source/​reports/​mar17_​medpac_​ch8.pdf) includes the following recommendation: “The Congress should . . . direct the Secretary to revise the prospective payment system (PPS) for skilled nursing facilities” and “. . . make any additional adjustments to payments needed to more closely align payment with costs.” (March 2017 MedPAC Report to Congress, 220). This recommendation is seemingly predicated on MedPAC's own analysis of the current SNF PPS, where they state that “almost since its inception the SNF PPS has been criticized for encouraging the provision of excessive rehabilitation therapy services and not accurately targeting payments for nontherapy ancillaries” (March 2017 MedPAC Report to Congress, 202). Finally, with regard to the possibility of changing the existing SNF payment system, MedPAC stated that “since 2015, [CMS] has gathered four expert panels to receive input on aspects of possible design features before it proposes a revised PPS” and further that “the designs under consideration are consistent with those recommended by the Commission” (March 2017 MedPAC Report to Congress, 203).

As we discussed in the proposed rule (83 FR 21035), the combination of the observed trends in the current SNF PPS discussed above (which strongly suggest that providers may be basing service provision on financial reasons rather than resident need), the issues raised in the OIG reports discussed above, and the issues raised by MedPAC, has caused us to consider significant revisions to the existing SNF PPS, in keeping with our overall responsibility to ensure that payments under the SNF PPS accurately reflect both resident needs and resource utilization.

We explained in the proposed rule (83 FR 20135 through 21036) that under the RUG-IV system, therapy service provision determines not only therapy payments but also nursing payments. This is because, as noted above, payment is based on the highest RUG category that the resident could be assigned to, so only one of a resident's assigned RUG groups, rehabilitation or nursing, is used for payment purposes. Each rehabilitation group is assigned a nursing CMI to reflect relative differences in nursing costs for residents in those rehabilitation groups, which is less specifically tailored to the individual nursing costs for a given resident than the nursing CMIs assigned for the nursing RUGs. We explained that, as mentioned above, because most resident days are paid using a rehabilitation RUG, and since assignment into a rehabilitation RUG is based on therapy service provision, this means that therapy service provision effectively determines nursing payments for those residents who are assigned to a rehabilitation RUG. Thus, we stated that we believe any attempts to revise the SNF PPS payment methodology to better account for therapy service provision under the SNF PPS would need to be comprehensive and affect both the therapy and nursing case-mix components. Moreover, we noted that in the FY 2015 SNF PPS final rule, in response to comments regarding access for certain “specialty” populations (such as those with complex nursing needs), that we agreed with the commenter that access must be preserved for all categories of SNF residents, particularly those with complex medical and nursing needs. We stated that, as appropriate, we would examine our current monitoring efforts to identify any revisions which may be necessary to account appropriately for these populations. (79 FR 45651).

In addition, MedPAC, in its March 2017 Report to Congress, stated that it has previously recommended that we revise the current SNF PPS to “base therapy payments on patient characteristics (not service provision), remove payments for NTA services from the nursing component, [and] establish a separate component within the PPS that adjusts payments for NTA services” (March 2017 MedPAC Report to Congress, 202). Accordingly, included among the proposed revisions we discussed in the proposed rule were revisions to the SNF PPS to address longstanding concerns regarding the ability of the RUG-IV system to account for variation in nursing and NTA services.

In May 2017, CMS released an Advance Notice of Proposed Rulemaking with comment (82 FR 20980) (the ANPRM), in which we discussed the history of and analyses conducted during the SNF Payment Models Research (PMR) project, which sought to address these concerns with the RUG-IV model, and sought comments on a possible replacement to the current RUG-IV model, which we called the Resident Classification System, Version I (RCS-I). As we stated in the proposed rule (83 FR 21036), this model was intended as an improvement over the RUG-IV model because it would better account for resident characteristics and care needs, thus better aligning SNF PPS payments with resource use and eliminating therapy provision-related financial incentives inherent in the current payment model used in the SNF PPS. We received many comments from stakeholders on a wide variety of aspects of the RCS-I model. After considering these comments, we made significant revisions to the RCS-I model to account for the concerns or questions raised by stakeholders, resulting in a revised case-mix classification model which we proposed in the FY 2019 SNF PPS proposed rule (83 FR 21018). To make clear the purpose and intent of replacing the existing RUG-IV system, the model we proposed is called the Patient-Driven Payment Model (PDPM). We refer readers to the FY 2019 SNF PPS proposed rule (83 FR 21036) for a discussion of the SNF PMR project, and the resulting SNF PMR technical report which contains supporting language and documentation related to the RCS-I model (available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​Downloads/​SNF_​Payment_​Models_​Research_​Technical_​Report201704.pdf), and the SNF PDPM technical report which presents analyses and results that were used to develop the proposed PDPM (available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​Downloads/​PDPM_​Technical_​Report_​508.pdf). We invited comments on any and all aspects of the proposed PDPM, including the research analyses described in the proposed rule, the SNF PDPM technical report and the SNF PMR technical report.

As further detailed below, and as we stated in the proposed rule (83 FR 21036), we believe that the PDPM represents an improvement over the RUG-IV model and the RCS-I model because it would better account for resident characteristics and care needs while reducing both systemic and administrative complexity. To better ensure that resident care decisions appropriately reflect each resident's actual care needs, we believe it is important to remove, to the extent possible, service-based metrics from the SNF PPS and derive payment from verifiable resident characteristics. In the sections that follow, we describe the comprehensive revisions we are implementing to the SNF PPS through the PDPM. Additionally, we discuss the comments we received on each of the proposed policies, our responses to these comments and the PDPM-related policies we are finalizing in this rule.

Before moving into the specific policy areas, we first discuss general comments we received on the PDPM, along with our responses.

Comment: Many commenters expressed support for the goals of the proposed PDPM, acknowledging that changes must be made to the current payment system. Many commenters also expressed concerns regarding the potential impacts on patient care which could result from implementing PDPM, specifically that PDPM will introduce new incentives into SNF payment that will have a negative impact on patient care. Some commenters believe that SNF providers could stint on care, most notably therapy services, and that such providers will be overcompensated for care that is not being delivered. Some commenters urged CMS to monitor the impacts on patient care of implementing PDPM and take action upon evidence of adverse trends. One commenter noted that PDPM does not correct the problems in the existing reimbursement model, assigning too few resources to nursing and NTAs.

Response: We appreciate the support we have received for PDPM and its goals. With respect to the concerns raised by commenters with regard to the potential impact of PDPM on patient care, specifically the possibility that some providers may stint on care or provide fewer services to patients, we plan to monitor closely service utilization, payment, and quality trends which may change as a result of implementing PDPM. If changes in practice and/or coding patterns arise, then we may take further action, which may include administrative action against providers as appropriate and/or proposing changes in policy (for example, system recalibration, rebasing case-mix weights, case mix creep adjustment) to address any concerns. We will also continue to work with the HHS Office of Inspector General, should any specific provider behavior be identified which may justify a referral for additional action.

With regard to the comment that PDPM does not correct the issues with the current reimbursement model and assigns too few resources to nursing and NTAs, we would refer the commenter to the impact analysis presented in Table 37, which indicates that the broadest shifts in payment are to those patients with high nursing and NTA needs.

Comment: Several commenters raised concerns regarding the use of historical data as the basis for developing PDPM. One commenter stated that PDPM is overly complex and that the majority of patient days are captured in a small number of case-mix groups. One commenter stated that because PDPM is based on historical utilization, it does not sufficiently reflect current best practices or high quality care.

Response: Historical data are the only form of data that can be used for any data analysis, so it is not clear what other data, that are not historical, CMS could have used to develop PDPM. Further, as these data are reported by SNFs, we believe that these data should be best reflective of SNF costs and patient needs. With regard to the comment that the majority of patient days are captured in a small number of case-mix groups, we agree with this comment and believe that this is precisely part of the motivation for implementing a new case-mix classification model. The current case-mix model has caused a homogenization in patient classification such that the current payment model does not adequately reflect differences among SNF patients. We believe that PDPM is a significant improvement in this regard, better reflecting the myriad differences between SNF patients in terms of their characteristics, care needs, and goals.

With regard to the comment that the historical data do not sufficiently reflect current best practices or high quality care, while we are concerned about this assertion from a patient care perspective, we do not believe that this would affect the accuracy of the reported data in terms of reflecting relative differences in costs, which is all that is necessary for developing accurate case-mix groups.

Comment: Several commenters requested clarification on the effect of implementing PDPM on the development of a unified Post-Acute Care (PAC) PPS and how PDPM would interact with a PAC PPS. One commenter requested that CMS establish a panel to advise on payment system changes across the PAC continuum.

Response: As a PAC PPS has not been established, we cannot provide guidance as to how the PDPM would interact with such a system, once developed. However, given that PDPM shifts away from the current case-mix model that utilizes service-based metrics as the primary determinant of payment for most days paid under the SNF Part A benefit to a model that utilizes patient characteristics as the basis for payment, and that most other PAC payment systems already rely more heavily on patient characteristics within their payment model, we believe that PDPM will better align the SNF PPS for this eventual transition to a PAC PPS as it brings the SNF PPS closer to those other PAC payment systems. We will consider the commenter's recommendation to establish a panel on payment system changes across the PAC continuum, particularly as we work to develop a PAC PPS.

Comment: Some commenters suggested that CMS consider including quality measures of effective rehabilitation services when evaluating the impact of PDPM.

Response: We appreciate these commenters' suggestion. In monitoring the impact of the PDPM, we will consider including measures for a variety of service areas as a component of our planned monitoring efforts.

Comment: Several commenters suggested that CMS should establish a plan to recalibrate the system to address any unanticipated impacts. More specifically, these commenters requested that CMS provide more details on plans to recalibrate the system in case of unanticipated service and performance changes, as well as plans to recalibrate the payment weights associated with the revised payment model.

Response: We appreciate the suggestions made by these commenters with regard to CMS providing plans for recalibrating the payment system after implementing PDPM. However, such recalibrations will depend largely on the results of our monitoring efforts and could take various forms. For example, in the FY 2012 SNF PPS final rule (76 FR 48486), we recalibrated the parity adjustment that was intended to ensure that SNF payments under RUG-IV matched those that would have been made under RUG-III, similar to how the parity adjustment discussed below for PDPM is intended to ensure that SNF payments under PDPM mirror those that would have been made under RUG-IV. As discussed in that rule, our assumptions regarding case-mix distribution that were used to calculate the RUG-IV parity adjustment subsequently proved to be inaccurate, which caused us to recalculate the RUG-IV parity adjustment in the following year. We anticipate similarly monitoring PDPM implementation closely and may propose adjustments as appropriate if we discover evidence that payments are either higher or lower than anticipated, or if provider costs change in such a manner that the current relationship between provider costs and provider payments changes from that currently observed.

Comment: One commenter raised the concern that the PDPM model has low explanatory power and lacks an objective threshold for inclusion of various components in the model. This commenter suggested that if CMS intends to update this model with new data over time to reflect changes in clinical practice and resource utilization, there is a need for a systematic determination of the minimum acceptable R-squared values for the model features. Model components currently excluded may increase in predictive power over time and merit inclusion in future versions of PDPM. In addition, current model components may decrease in predictive power such that they should be removed from the model.

Response: Setting an absolute minimum threshold would not only be arbitrary but also deviates from the practical use of the R-squared metric, which is to evaluate the proportion of variance explained and compare models with the same dependent variable vector. Additionally, R-squared is not the only measure we use to evaluate PDPM. In fact, because the current system is heavily based on service provision and most residents are classified into the Ultra-High therapy category, we are dealing with a dataset with little explainable variance. Each of the PDPM case-mix groups meets clinical expectations, which is a convincing validation of the model given the data available. We note that with the change to a patient driven model, we expect more variation will appear in therapy costs. This will allow for future development of models with higher explanatory power.

Comment: Several commenters requested clarification on how PDPM would interact with other CMS initiatives, such as the SNF Quality Reporting Program (QRP), Value-Based Purchasing (VBP) program, revised conditions of participation and other such initiatives. A few commenters also requested clarification on how PDPM accounts for or would interact with the Jimmo v Sebelius settlement surrounding the provision of maintenance therapy. These commenters requested clarification on how CMS would track maintenance therapy services, as compared to other forms of therapy. Several commenters requested clarification on how Comprehensive Person-Centered Plan maintenance services, new Requirements of Participation and other CMS initiatives will be factored into CMS burden estimates and that CMS should revise existing burden estimates to incorporate these changes.

Response: We anticipate that PDPM will only serve to strengthen the various quality and payment reform initiatives throughout CMS, by shifting payment away from the current service-driven model that has produced nearly homogenized care for SNF beneficiaries, to a more resident-centered model that focuses more on the individual patient's needs and characteristics. We also believe that through the use of standardized assessment items (as discussed in section V.D. of this final rule) and changes to the assessment schedule to mirror that of other PAC settings that use a similar admission/discharge assessment model (as discussed in section V.E. of this final rule), the PDPM would better align with the current direction of PAC reform and standardization efforts supported by the IMPACT Act.

With regard to the comment about tracking maintenance services, we do not believe it is necessary at this time to track maintenance services separately. Such tracking would be burdensome and it would be difficult to do so accurately, as it is possible that many patients have both maintenance and restorative goals, and allocating therapy minutes among these varied goals would be particularly complicated for providers.

With regard to the burden of the Comprehensive Person-Centered Plan, new requirements of participation, and other CMS initiatives, the burdens estimated in relation to PDPM are only those in relation to implementation of the PDPM and its related policies. As the Comprehensive Person-Centered Plan and other issues mentioned are outside of these PDPM related policies, we do not address the potential burden of such issues in this section.

Comment: Several commenters expressed concerns regarding the potential impact of implementing PDPM on Medicaid programs. A few commenters raised concerns regarding the impact of PDPM on calculating the Upper Payment Limit (UPL), which is utilized as part of calculating Medicaid payment rates. One commenter questioned if states would be permitted to still use RUG-IV as the basis for estimating the UPL. One commenter requested clarification on if any changes would be necessary for Medicaid claims systems. One commenter stated that Medicaid providers will have less incentive to provide therapy and Medicaid beneficiaries will have lower nursing case-mix scores under PDPM, thereby incentivizing states to transition to PDPM in order to reduce Medicaid spending. Commenters suggested that CMS work closely with states, who may wish to transition to PDPM, to ensure a smooth transition. Some commenters also stated that, should certain states not transition to PDPM, this would mean operating two different payment systems. A few commenters requested clarification on if CMS would continue to support previous payment systems for states that do not make the transition to PDPM or have access to MDS data for Medicaid rate-setting purposes. These commenters also requested if CMS could provide a further breakdown of certain cost categories, such as NTA costs, in a manner that would be more helpful to states in conducting UPL calculations.

Response: We appreciate the commenters' concerns with the potential impact of PDPM on Medicaid programs. We agree with the commenters that this is an area that deserves significant attention in terms of education and training, and we plan to work with states to ensure a smooth transition between the current RUG-IV model and PDPM. With regard to questions on how PDPM may relate to UPL calculations, these calculations are based on how Medicare pays for services under Part A and not based on a prior payment system. Therefore, UPL calculations, after PDPM has been implemented, would need to be based on the payments made under PDPM. That being said, we expect that, because PDPM bases payment on patient characteristics and not service utilization, payments made under PDPM will more accurately reflect patient needs and goals, which should also improve the basis for Medicaid payments which may be related to Medicare payments. With regard to having the data necessary for such UPL calculations, whether in regard to specific rate components (for example, NTA costs) or more generally, we will work with states to help ensure that they have the necessary information so PDPM implementation does not negatively impact on their ability to manage their Medicaid programs.

With regard to the comment that states may have more of an incentive to transition to PDPM in order to reduce Medicaid spending, we believe that the primary reason that Medicaid programs may adopt PDPM is due to its focus on patient characteristics and goals, rather than on service utilization. Given the improvements in Medicare payment that this transition represents, we would expect a similar improvement in Medicaid payments in states that make this transition.

With regard to the comment that Medicaid providers will be incentivized to provide less therapy or that Medicaid beneficiaries will have lower nursing case-mix scores, we would encourage states that decide to transition to PDPM to ensure they are monitoring the impacts of such a change on their beneficiaries and the care they receive.

In terms of those states that opt not to transition to PDPM and instead use some form of legacy payment system, we would note that a number of states use systems quite distinct from the existing RUG-IV model and we are not aware of any difficulties or complexities for providers or states in managing these systems concurrently. These states still have access to MDS data for ratesetting purposes and nothing associated with PDPM implementation, in and of itself, would affect state access to MDS data. That being said, we would likely need to evaluate the costs and benefits of continued support for certain legacy payment systems, most notably any RUG-III based payment models.

Comment: One commenter requested that CMS consider the possibility that some Medicare Advantage plans could reform their payment models to mirror PDPM, while others may maintain their existing payment models, which could include models that resemble RUG-IV. The commenter requested that CMS consider working with those plans that opt to modify their payment models to resemble PDPM and consider the impact of having multiple payment models that providers must operate under simultaneously.

Response: We acknowledge that some Medicare Advantage plans could change their payment models to mirror PDPM, while others may not change their payment models in relation to the changes finalized in this rule. We would note, however, that, as private plans, Medicare Advantage plans currently take a wide variety of forms, with some already approximating the structure of PDPM, using patient characteristics rather than service utilization as the basis for payment. We will work generally with stakeholders, including these private plans, to help ensure that adequate education and resources are available for all parties.

Comment: One commenter requested clarification on how CMS will track and reconcile patient diagnosis and classification information reported at admission with such information at discharge, expressing concern regarding what might occur in the case that the information from these two points is different, as well as diagnosis or procedural information from the preceding hospital stay, noting that some information for SNF payment comes from the hospital, and how these issues could affect provider risk of alleged improper billing and recovery efforts.

Response: We plan to develop a robust monitoring program that utilizes data from many sources, such as assessments, claims, cost reports and other data that would prove valuable in assessing both the impact of implementing PDPM, as well as identify any provider level issues related to PDPM payments. While the vast majority of information related to PDPM classification and payment is derived from the SNF, there is one area (surgical procedural information) which may come from the preceding hospital stay. However, nothing in PDPM should change the relationship or need for information between the hospital and SNF, given that the information that PDPM requires is no more information than the SNF would need simply for basic care planning purposes. As such, there should be no impact on improper billing or recovery efforts that derive from the implementation of PDPM.

Comment: One commenter requested clarification on how PDPM will address the number of face-to-face hours the registered therapist spends treating the patients. This commenter states they have observed nursing staff instructed to complete certain activities with patients who are receiving therapy.

Response: PDPM does not address the specific number of face-to-face hours that therapists spend with their patients. The expectations for what is considered skilled therapy and reasonable and necessary care found in Chapter 8 of the Medicare Benefit Policy Manual (https://www.cms.gov/​Regulations-and-Guidance/​Guidance/​Manuals/​downloads/​bp102c08.pdf) and the MDS 3.0 RAI manual (https://downloads.cms.gov/​files/​MDS-30-RAI-Manual-v115-October-2017.pdf) will not change under PDPM. We continue to expect that patients will receive high quality skilled rehabilitation services based on their individual needs and we do not believe that patients should have any nursing care that they require reduced because they happen to be receiving therapy. If a patient requires nursing care (including restorative nursing), the SNF should provide that nursing care as medically necessary. Similarly, if a patient requires therapy, the SNF should provide the therapy as medically necessary. One should not impact the other and PDPM does not affect this either.

Comment: Several commenters requested clarification about how SNFs are expected to comply with Medicare and Medicaid Conditions of Participation and whether SNFs will continue to be required to complete the discharge assessments required by the Omnibus Budget Reconciliation Act (OBRA), as well as the end of therapy-related assessments.

Response: PDPM is not intended to affect any of the Medicare and Medicaid Conditions of Participation for SNFs. Facilities should continue to follow these regulations as they always have. Additionally, even though under PDPM, the majority of PPS assessments will now be removed (as discussed later in this final rule), all OBRA assessments will still be required. PDPM will not affect the OBRA requirements. With regard to existing therapy-related assessments (the Start of Therapy, End of Therapy, and Change of Therapy assessments), these assessments would no longer exist under PDPM.

Comment: Several commenters expressed concern that PDPM may not fully account for mild cognitive impairment and encouraged CMS to collect more sensitive data, in line with the IMPACT Act, to ensure necessary attention to cognition.

Response: We appreciate these commenters' concerns and also believe that attention should be paid to cognition as an area for potential future system refinements. However, as the only change in the proposed use of cognition as a factor in payment classification is under the SLP component, and because for this component, we proposed to use even mild cognitive impairment as the basis for a payment classification, we believe that PDPM does adequately account for mild cognitive impairment. We will consider the commenter's concerns as we continue to evaluate potential refinements to our assessment tools.

Comment: One commenter expressed concern that PDPM does not incorporate incentives for quality improvement.

Response: PDPM, as a case-mix classification system, is intended to classify SNF patients for purposes of reimbursement based on the resource utilization associated with treating those patients. However, there do exist programs, such as the SNF VBP program, that is a part of the SNF PPS which does incentivize quality improvement. Therefore, while we agree that PDPM, in and of itself, does not include incentives for quality improvement, other aspects of the SNF PPS do include such incentives.

Comment: Some commenters requested clarification about the appeals process that will be available to help patients in case of shortcomings in their care and coverage, including any inaccurate assignments to payment classifications.

Response: We appreciate this comment, but would note that nothing associated with PDPM implementation would affect existing patient appeal rights or processes.

Comment: One commenter requested clarification on how items Z0100A and Z0150A on the MDS would be populated and how the classifications would translate to a billable claim code.

Response: We will provide detail on how these MDS items, which relate to patient billing codes, will be populated as part of our updates to the MDS manual.

Comment: One commenter requested clarification on how a patient's voice would be heard in a care design driven by medical information.

Response: While patient case-mix classification, for purposes of payment, would be driven by medical information, as occurs under the current payment system, care design should be driven by patient goals and needs, as well as discussions with the patient and his or her family. Further, while under the current payment model over 90 percent of patient days are paid for using a therapy RUG, which utilizes only therapy minutes and ADLs as the basis for payment, PDPM provides a more holistic approach to payment classifications. More specifically, by separately adjusting for the nursing component, which utilizes patient interviews as a major component of patient classification, we believe that this achieves the commenter's goal of elevating the patient's voice.

Comment: Some commenters requested that CMS consider adopting an outlier policy as part of the SNF PPS to account for patients whose costs far exceed the cost of typical patients. These commenters stated that a SNF outlier policy would ensure access to clinically complex patients and align with other PAC systems.

Response: Under the current statutory provisions governing the SNF PPS, there is no specific statutory authority for an outlier payment as part of the SNF PPS.

B. Revisions to SNF PPS Federal Base Payment Rate Components

1. Background on SNF PPS Federal Base Payment Rates and Components

Section 1888(e)(4) of the Act requires that the SNF PPS per diem federal payment rates be based on FY 1995 costs, updated for inflation to the first effective period of the PPS. These base rates are then required to be adjusted to reflect differences among facilities in patient case-mix and in average wage levels by area. In keeping with this statutory requirement, the base per diem payment rates were set in 1998 and reflect average SNF costs in a base year (FY 1995), updated for inflation to the first period of the SNF PPS, which was the 15-month period beginning on July 1, 1998. The federal base payment rates were calculated separately for urban and rural facilities and based on allowable costs from the FY 1995 cost reports of hospital-based and freestanding SNFs, where allowable costs included all routine, ancillary, and capital-related costs (excluding those related to approved educational activities) associated with SNF services provided under Part A, and all services and items for which payment could be made under Part B prior to July 1, 1998.

In general, routine costs are those included by SNFs in a daily service charge and include regular room, dietary, and nursing services, medical social services and psychiatric social services, as well as the use of certain facilities and equipment for which a separate charge is not made. Ancillary costs are directly identifiable to residents and cover specialized services, including therapy, drugs, and laboratory services. Lastly, capital-related costs include the costs of land, building, and equipment and the interest incurred in financing the acquisition of such items (63 FR 26253).

There are four federal base payment rate components which may factor into SNF PPS payment. Two of these components, “nursing case-mix” and “therapy case-mix,” are case-mix adjusted components, while the remaining two components, “therapy non-case-mix” and “non-case-mix,” are not case-mix adjusted. While we discussed the details of the proposed PDPM and justifications for certain associated policies we proposed throughout section V of the FY 2019 SNF PPS proposed rule, we note that, as part of the PDPM case-mix model, we proposed to bifurcate the “nursing case-mix” component of the federal base payment rate into two case-mix adjusted components and separate the “therapy case-mix” component of the federal base payment rate into three case-mix adjusted components, thereby creating five case-mix adjusted components of the federal base per diem rate. More specifically, we proposed to separate the “therapy case-mix” rate component into a “Physical Therapy” (PT) component, an “Occupational Therapy” (OT) component, and a “Speech-Language Pathology” (SLP) component. Our rationale for separating the therapy case-mix component in this manner is presented in section V.D.3.b. of the proposed rule. Based on the results of the SNF PMR, we also proposed to separate the “nursing case-mix” rate component into a “Nursing” component and a “Non-Therapy Ancillary” (NTA) component. Our rationale for proposing to bifurcate the nursing case-mix component in this manner is presented in section V.D.3.d. of the proposed rule. Given that all SNF residents under PDPM would be assigned to a classification group for each of the three proposed therapy-related case-mix adjusted components as further discussed below, we proposed eliminating the “therapy non-case-mix” rate component under PDPM and stated that we would distribute the dollars associated with this current rate component amongst the proposed PDPM therapy components. We also stated in the proposed rule (83 FR 21038) that the existing non-case-mix component would be maintained as it is currently constituted under the existing SNF PPS. We explained that although the case-mix components of the proposed PDPM case-mix classification system would address costs associated with individual resident care based on an individual's specific needs and characteristics, the non-case-mix component addresses consistent costs that are incurred for all residents, such as room and board and various capital-related expenses. As these costs are not likely to change, regardless of what changes we might make to the SNF PPS, we proposed to maintain the non-case-mix component as it is currently used.

In the next section, we discuss the methodology used to create the proposed PDPM case-mix adjusted components, as well as the data sources used in this calculation. As we stated in the proposed rule (83 FR 21038), the proposed methodology does not calculate new federal base payment rates but simply proposes to modify the existing base rate case-mix components for therapy and nursing. The methodology and data used in this calculation are based on the data and methodology used in the calculation of the original federal payment rates in 1998, as further discussed below.

2. Data Sources Utilized for Revision of Federal Base Payment Rate Components

Section II.A.2. of the interim final rule with comment period that initially implemented the SNF PPS (63 FR 26256 through 26260) provides a detailed discussion of the data sources used to calculate the original federal base payment rates in 1998. Except as discussed below, we proposed to use the same data sources (that is, cost information from FY 1995 cost reports) to determine the portion of the therapy case-mix component base rate that would be assigned to each of the proposed therapy component base rates (PT, OT, and SLP). As we stated in the proposed rule (83 FR 21038), we believe that using the same data sources, to the extent possible, that were used to calculate the original federal base payment rates in 1998 results in base rates for the components that resemble as closely as possible what they would have been had these components initially been established in 1998. The portion of the nursing component base rate that corresponds to NTA costs was already calculated using the same data source used to calculate the federal base payment rates in 1998. As explained below and in the proposed rule (83 FR 21038), we used the previously calculated percentage of the nursing component base rate corresponding to NTA costs to set the NTA base rate and verified this calculation with the analysis described in section V.C.3. of the FY 2019 SNF PPS proposed rule. Therefore, the steps described below address the calculations performed to separate out the therapy base rates alone.

As discussed in the proposed rule (83 FR 21038), the percentage of the current therapy case-mix component of the federal base payment rates that would be assigned to the three proposed therapy components (PT, OT, and SLP) of the federal base payment rates was determined using cost information from FY 1995 cost reports, after making the following exclusions and adjustments: First, only settled and as-submitted cost reports for hospital-based and freestanding SNFs for periods beginning in FY 1995 and spanning 10 to 13 months were included. This set of restrictions replicates the restrictions used to derive the original federal base payment rates as set forth in the 1998 interim final rule with comment period (63 FR 26256). Following the methodology used to derive the SNF PPS base rates, routine and ancillary costs from as-submitted cost reports were adjusted down by 1.31 and 3.26 percent, respectively. As discussed in the 1998 interim final rule with comment period, the specific adjustment factors were chosen to reflect average adjustments resulting from cost report settlement and were based on a comparison of as-submitted and settled reports from FY 1992 to FY 1994 (63 FR 26256); these adjustments are in accordance with section 1888(e)(4)(A)(i) of the Act. We used similar data, exclusions, and adjustments as in the original base rates calculation so the resulting base rates for the components would resemble as closely as possible what they would have been had they been established in 1998. However, as we discussed in the proposed rule, there were two ways in which the PT, OT, and SLP percentage calculations deviate from the 1998 base rates calculation. First, the 1998 calculation of the base rates excluded reports for facilities exempted from cost limits in the base year. The available data do not identify which facilities were exempted from cost limits in the base year, so this restriction was not implemented. As we stated in the proposed rule, we do not believe this had a notable impact on our estimate of the PT, OT, and SLP percentages, because only a small fraction of facilities were exempted from cost limits. Consistent with the 1998 base rates calculation, we excluded facilities with per diem costs more than three standard deviations higher than the geometric mean across facilities. Therefore, facilities with unusually high costs did not influence our estimate. Second, the 1998 calculation of the base rates excluded costs related to exceptions payments and costs related to approved educational activities. The available cost report data did not identify costs related to exceptions payments nor indicate what percentage of overall therapy costs or costs by therapy discipline were related to approved educational activities, so these costs are not excluded from the PT, OT, and SLP percentage calculations. We stated in the proposed rule that because exceptions were only granted for routine costs, we believe the inability to exclude these costs should not affect our estimate of the PT, OT, and SLP percentages as exceptions would not apply to therapy costs. Additionally, the data indicate that educational costs made up less than one-hundredth of 1 percent of overall SNF costs. Therefore, we stated that we believe the inability to exclude educational costs should have a negligible impact on our estimates.

In addition to Part A costs from the cost report data, the 1998 federal base rates calculation incorporated estimates of amounts payable under Part B for covered SNF services provided to Part A SNF residents, as required by section 1888(e)(4)(A)(ii) of the Act. We stated in the proposed rule (83 FR 21038) that in calculating the PT, OT, and SLP percentages, we also estimated the amounts payable under Part B for covered SNF services provided to Part A residents. All Part B claims associated with Part A SNF claims overlapping with FY 1995 cost reports were matched to the corresponding facility's cost report. For each cost center (PT, OT, and SLP) in each cost report, a ratio was calculated to determine the amount by which Part A costs needed to be increased to account for the portion of costs payable under Part B. This ratio for each cost center was determined by dividing the total charges from the matched Part B claims by the total charges from the Part A SNF claims overlapping with the cost report. The 1998 interim final rule (63 FR 26256) states that to estimate the amounts payable under Part B for covered SNF services provided to Part A SNF residents, CMS (then known as HCFA) matched 100 percent of Part B claims associated with Part A covered SNF stays to the corresponding facility's cost report. Part B allowable charges were then incorporated at the facility level by the appropriate cost report center. Although the interim final rule does not provide further detail on how Part B allowable charges were incorporated at the facility level, we stated in the proposed rule that we believe our methodology reasonably approximates the methodology described in the interim final rule, and provides a reasonable estimate of the amounts payable under Part B for covered SNF services provided to Part A residents for purposes of calculating the PT, OT, and SLP percentages. Therefore, we stated that we believe it is reasonable to use this methodology to calculate the PT, OT, and SLP percentages of the therapy case-mix component.

Finally, the 1998 federal base rates calculation standardized the cost data for each facility to control for the effects of case-mix and geographic-related wage differences, as required by section 1888(e)(4)(C) of the Act. As we stated in the proposed rule, when calculating the PT, OT and SLP shares of the current therapy base rate, we replicated the method used in 1998 to standardize for wage differences, as described in the 1998 interim final rule with comment period (63 FR 26259 through 26260). We applied a hospital wage index to the labor-related share of costs, estimated at 75.888 percent, and used an index composed of hospital wages from FY 1994. We noted in the proposed rule that the PT, OT, and SLP percentage calculations did not include the case-mix adjustment used in the 1998 calculation because the 1998 adjustment relied on the obsolete RUG-III classification system. In the 1998 federal base rates calculation, information from SNF and inpatient claims was mapped to RUG-III clinical categories at the resident level to case-mix adjust facility per diem costs. However, the 1998 interim final rule did not document this mapping, and the data used as the basis for this adjustment are no longer available, and therefore, this step could not be replicated. We stated in the proposed rule that we believe the inability to apply the case-mix adjustment likely has a small impact on our estimate of the PT, OT, and SLP percentages. The 1998 interim final rule indicates that the case-mix adjustment was applied by dividing facility per diem costs for a given component by average facility case mix for that component; in other words, multiplying by the inverse of average facility case mix. As we discussed in the proposed rule, as long as average facility case-mix values are within a relatively narrow range, adjustment for facility case mix should not have a large impact on the estimated PT, OT, and SLP percentages. Because the RUG-III case-mix indexes shown in the 1998 interim final rule are within a relatively narrow range (for example, therapy indexes range from 0.43 to 2.25), we stated that we do not expect the inability to apply the case-mix adjustment to facility per diem costs to have a large influence on the estimated PT, OT, and SLP percentages. These data sources are described in more detail in section 3.10. of the SNF PDPM technical report, available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html.

We invited comments on the data sources used to determine the PT, OT, and SLP rate components, as discussed above.

Commenters submitted the following comments related to the proposed rule's discussion of the Data Sources Utilized for Proposed Revision of Federal Base Payment Rate Components. A discussion of these comments, along with our responses, appears below.

Comment: One commenter requested additional information on the data sources used to develop PDPM. Specifically, the commenter requested that CMS clarify which year of claims and cost report data was used to develop PDPM.

Response: As detailed in section 3.1 of the SNF PDPM technical report and FY 2019 SNF PPS proposed rule (83 FR 21041), we used data, including claims and assessments, corresponding to Medicare Part A SNF stays with admissions in FY 2017. This was the most complete year of data available when PDPM was developed and continues to be the most complete year of data available as of the FY 2019 final rule. Foundational analyses—for example, those discussed throughout the SNF PMR technical report that accompanied the 2017 ANPRM—used FY 2014 data, as that was the most recent complete year of data available when those analyses were completed. Finally, based on suggestions from commenters responding to the 2017 ANPRM, the analysis that established the list of comorbidities used for payment in the PDPM NTA component and the points associated with each comorbid condition used multiple years of data to generate more robust results. Specifically, resource utilization and assessment data from FYs 2014-2017 were used to determine the comorbid conditions associated with high NTA utilization and estimate the specific resource utilization associated with each condition for the purpose of assigning points and payment to these conditions under PDPM. This methodology is discussed in further detail in section 3.7 of the SNF PDPM technical report and in the FY 2019 SNF PPS proposed rule (83 FR 21056). In terms of cost reports, since providers have their own fiscal year and reporting schedule, we used the cost report closest to the stay window among the cost reports of that provider recorded in the database as of November 2017.

Comment: Some commenters questioned whether it is appropriate to use the same data sources and methodology from 1998 (that is, 1995 cost reports) to set base rates given updated technology and changes in SNF care practices since then. Particularly, a few commenters stated that the estimated share of the nursing base rate attributed to NTA services (43 percent) is outdated and not representative of the proportion of the nursing base rate that corresponds to NTA services. These commenters requested that we consider recalculating SNF base rates using more recent data on SNF costs.

Response: We appreciate the commenters' suggestion to use more recent data in calculating the SNF base rates. However, in accordance with section 1888(e)(4)(A) of the Act, the federal per diem rates used for SNF payment are based on the FY 1995 cost reports. Therefore, we cannot consider recalculating the SNF base rates using more recent data. Additionally, given this statutory requirement, we believed that it was appropriate to use these cost reports to set the base rates for the proposed new components to reflect as closely as possible what the base rates would have been for these components if they had been separately established in 1998. Finally, while it may be the case that, as the commenter stated, changes in SNF care practices may have occurred, such changes would more likely be reflected in differences in the relative costs of treating different types of patients and these types of changes in relative costs are reflected in the revised case-mix weights under PDPM, which does use more recent data than FY 1995. Specifically, as discussed in section 3.1 of the SNF PDPM technical report and FY 2019 SNP PPS proposed rule (83 FR 21041), we developed PDPM using data, including claims and assessments, corresponding to Medicare Part A SNF stays with admissions in FY 2017.

Comment: One commenter recommends that CMS treat respiratory therapy as “therapy” and not “nursing” for purpose of payment, and recommends CMS consider incorporating an add-on payment for respiratory therapy to ensure it is reimbursed appropriately to safeguard the continuation of these therapy services.

Response: Under Chapter 8 of the Medicare Benefit Policy manual, section 30.4, “skilled therapy services” includes physical therapy, occupational therapy, and speech-language pathology therapy (reflecting the regulations at 42 CFR 409.23). Respiratory therapy, on the other hand, is treated as a separate service category in section 50.8.2 of the same chapter (reflecting the regulations at § 409.27(b)). As such, respiratory therapy is distinct from other forms of therapy and is not included among the other therapy components. Additionally, therapy services, as defined in § 409.33 make specific reference to skilled therapy services provided by physical and occupational therapists and speech-language pathologists. Finally, while respiratory therapists have specialized training in addressing respiratory issues, much of the work conducted by respiratory therapists falls within the scope of practice for nurses, which further supports the closer relationship between respiratory therapy and nursing, rather than with the three therapy disciplines. With regard to developing an add-on payment for respiratory therapy, given that such services are currently captured through the global per diem payment, we do not believe that an add-on payment would be warranted.

3. Methodology Used for the Calculation of Federal Base Payment Rate Components

As discussed previously in this section, we proposed to separate the current therapy components into a PT component, an OT component, and an SLP component. To do this, we calculated the percentage of the current therapy component of the federal base rate that corresponds to each of the three proposed PDPM therapy components (PT, OT, and SLP) in accordance with the methodology set forth below and in the FY 2019 SNF PPS proposed rule (83 FR 21039).

The data described in section V.C.2. of the proposed rule (primarily, cost information from FY 1995 cost reports) provides cost estimates for the Medicare Part A SNF population for each cost report that met the inclusion criteria. Cost reports stratify costs by a number of cost centers that indicate different types of services. For instance, costs are reported separately for each of the three therapy disciplines (PT, OT, and SLP). Cost reports also include the number of Medicare Part A utilization days during the cost reporting period. As we stated in the proposed rule, this allows us to calculate both average total therapy costs per day and average therapy costs by discipline in the facility during the cost reporting period. Therapy costs are defined as the sum of costs for the three therapy disciplines.

As explained in the proposed rule (83 FR 21039), the goal of this methodology is to estimate the proportion of therapy costs that corresponds to each of the three therapy disciplines. We use the facility-level per-diem costs developed from 1995 cost reports to derive average per diem amounts for both total therapy costs and for PT, OT, and SLP costs separately. To do this, we followed the methodology outlined in section II.A.3. of the 1998 interim final rule with comment period (63 FR 26260), which was used by CMS (then known as HCFA) to create the federal base payment rates:

(1) For each of the four measures of cost (PT, OT, SLP, and total therapy costs per day), we computed the mean based on data from freestanding SNFs only. This mean was weighted by the total number of Medicare days of the facility.

(2) For each of the four measures of cost (PT, OT, SLP, and total therapy costs per day), we computed the mean based on data from both hospital-based and freestanding SNFs. This mean was weighted by the total number of Medicare days of the facility.

(3) For each of the four measures of cost (PT, OT, SLP, and total therapy costs per day), we calculated the arithmetic mean of the amounts determined under steps (1) and (2) above.

In section 3.10.3. of the SNF PDPM technical report (available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html), we show the results of these calculations.

The three steps outlined above produce a measure of costs per day by therapy discipline and a measure of total therapy costs per day. We divided the discipline-specific (PT, OT, SLP) cost measure by the total therapy cost measure to obtain the percentage of the therapy component that corresponds to each therapy discipline. As we discussed in the proposed rule (83 FR 21039), we believe that following a methodology to derive the discipline-specific therapy percentages that is consistent with the methodology used to determine the base rates in the 1998 interim final rule with comment period is appropriate because a consistent methodology helps to ensure that the resulting base rates for the components resemble what they would be had they been established in 1998. We found that PT, OT, and SLP costs correspond to 43.4 percent, 40.4 percent, and 16.2 percent of the therapy component of the federal per diem rate for urban SNFs, and 42.9 percent, 39.4 percent, and 17.7 percent of the therapy component of the federal per diem rate for rural SNFs. Under the proposed PDPM, we stated that the current therapy case-mix component would be separated into a Physical Therapy component, an Occupational Therapy component, and a Speech-Language Pathology component using the percentages derived above. We stated that this process would be done separately for urban and for rural facilities. In the appendix of the SNF PDPM technical report (available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html) we provided the specific cost centers used to identify PT, OT, and SLP costs.

In addition, we proposed to separate the current nursing case-mix component into a nursing case-mix component and an NTA component. Similar to the therapy component, we calculated the percentage of the current nursing component of the federal base rates that corresponds to each of the two proposed PDPM components (NTA and nursing). The 1998 reopening of the comment period for the interim final rule (63 FR 65561, November 27, 1998) states that NTA costs comprise 43.4 percent of the current nursing component of the urban federal base rate, and the remaining 56.6 percent accounts for nursing and social services salary costs. These percentages for the nursing component of the federal base rate for rural facilities are 42.7 percent and 57.3 percent, respectively (63 FR 65561). Therefore, we proposed to assign 43 percent of the current nursing component of the federal base rates to the new NTA component of the federal base rates and assign the remaining 57 percent to the new nursing component of the federal base rates to reflect what the base rates would have been for these components if they had been separately established in 1998.

As discussed in the proposed rule (83 FR 21040), we verified the 1998 calculation of the percentages of the nursing component federal base rates that correspond to NTA costs by developing a measure of NTA costs per day for urban and rural facilities. We used the same data (that is, cost information from 1995 cost reports) and followed the same methodology described above to develop measures of PT, OT, and SLP costs per day and total therapy costs per day. The measure of NTA costs per day produced by this analysis was $47.70 for urban facilities and $47.30 for rural facilities. The original 1998 federal base rates for the nursing component, which relied on a similar methodology, were $109.48 for urban facilities and $104.88 for rural facilities. Therefore, our measure of NTA costs in urban facilities was equivalent to 43.6 percent of the urban 1998 federal nursing base rate, and our measure of NTA costs in rural facilities was equivalent to 45.1 percent of the rural 1998 federal nursing base rate. These results are similar to the estimates published in the 1998 reopening of the comment period for the interim final rule (63 FR 65561, November 27, 1998), which we stated we believe supports the validity of the 43 percent figure stated above.

For illustration purposes, Tables 12 and 13 set forth what we stated the unadjusted federal per diem rates would be for each of the case-mix adjusted components if we were to apply the proposed PDPM to the FY 2019 base rates given in Tables 4 and 5. These were derived by dividing the FY 2019 SNF PPS base rates according to the percentages described above. Tables 12 and 13 also show what the unadjusted federal per diem rates for the non-case-mix component would be, which are not affected by the change in case-mix methodology from RUG-IV to PDPM. We used these unadjusted federal per diem rates in calculating the impact analysis discussed in section V.J. of the proposed rule.

Table 12—FY 2019 PDPM Unadjusted Federal Rate per Diem—Urban ¹

Rate component	Nursing	NTA	PT	OT	SLP	Non-case-mix
Per Diem Amount	$103.46	$78.05	$59.33	$55.23	$22.15	$92.63
1 The rates shown in Tables 12 and 13 illustrate what the adjusted federal per diem rates would be for each of the case-mix adjusted components if we were to apply the proposed PDPM to the proposed FY 2019 base rates given in Tables 4 and 5.

Table 13—FY 2019 PDPM Unadjusted Federal Rate per Diem—Rural

Rate component	Nursing	NTA	PT	OT	SLP	Non-case-mix
Per Diem Amount	$98.83	$74.56	$67.63	$62.11	$27.90	$94.34

We invited comments on the proposed data sources and proposed methodology for calculating the unadjusted federal per diem rates that would be used in conjunction with the proposed PDPM effective October 1, 2019.

Commenters submitted the following comments related to the proposed rule's discussion of the Methodology Used for the Calculation of Federal Base Payment Rate Components. A discussion of these comments, along with our responses, appears below.

Comment: Some commenters supported the proposed changes to the SNF PPS base rates. One commenter specifically highlighted their support for including an NTA component. Some commenters sought clarification regarding how CMS intends to distribute system resources currently associated with the “therapy non-case-mix” base rate. Specifically, they stated that the FY 2019 SNF proposed rule and the SNF PDPM technical report that accompanied the proposed rule appear to be inconsistent in describing how resources associated with this payment component will be distributed under the new payment model. Commenters note that the proposed rule stated that resources associated with the “therapy non-case-mix” base rate will be redistributed among the three PDPM case-mix therapy components, while the SNF PDPM technical report states that the “therapy non-case-mix” payment component is dropped from the payment model under PDPM.

Response: We appreciate the support for our proposed changes. As stated in the proposed rule, we believe it is appropriate to eliminate the non-case-mix therapy base rate because facilities will be compensated for residents who receive nominal amounts of therapy (for example, therapy evaluations) through the three PDPM base rates corresponding to the three disciplines of therapy provided in the SNF setting (PT, OT, and SLP) under the new payment model. In other words, whereas under the existing RUG-IV reimbursement model, facilities receive a non-case-mix therapy payment for residents who receive nominal amounts of therapy, under PDPM facilities would receive payment for these residents through the PT, OT, and SLP payment components.

Additionally, in setting component base rates under PDPM, we sought to replicate the methodology used to estimate the SNF PPS original base rates in 1998 as closely as possible. This is consistent with the requirements of section 1888(e)(4) of the Act, which requires that SNF PPS per diem federal payment rates be based on FY 1995 costs reports. Therefore, to ensure that the PDPM base rates resembled as closely as possible what they would have been had these components been established in 1998, we used FY 1995 cost reports to determine the share of therapy costs accounted for by PT, OT, and SLP. As described in the proposed rule (83 FR 21038 through 21039) and in section 3.10 of the SNF PDPM technical report, we then used the percentage of costs associated with each of these disciplines to calculate the corresponding base rates for the PT, OT, and SLP components under PDPM.

Finally, as further discussed in section 3.11 of the SNF PDPM technical report, we adjusted CMIs for each of the five case-mix-adjusted components of PDPM to ensure budget neutrality between RUG-IV and PDPM. In doing so, we applied a multiplier to CMIs for all five case-mix-adjusted PDPM payment components so that total estimated payments under PDPM are budget neutral relative to RUG-IV. This procedure effectively distributes resources that are currently associated with the “therapy non-case-mix” component of RUG-IV across all five case-mix components of PDPM. We acknowledge that the proposed rule inadvertently stated that the resources associated with the therapy non-case mix component were distributed across only the three PDPM case-mix therapy components. Thus, we are clarifying that, while we did eliminate the therapy non-case mix component from the model, we redistributed resources associated with this component across the five PDPM case-mix components as described in section 3.11 of the PDPM technical report.

Comment: Many commenters expressed concern regarding the base rate for the SLP component, specifically that it is much lower than that of the other therapy base rates. Commenters suggested that this may be taken to devalue SLP services and that low reimbursement will lead to a decrease in the utilization of SLP services. Some commenters further suggested that such low reimbursement rates could lead to layoffs among SLPs and believe that PDPM should pay equally for all three therapy disciplines.

Response: We appreciate the concerns raised by these commenters regarding the potential impact on SLP services resulting from the payment policies in relation to SLP services discussed in the proposed rule. With regard to the comment about the SLP component base rate, as described above, we utilized the proportion of the current therapy base rate corresponding to each therapy discipline as the basis for allocating the therapy base rate as the basis for allocating the therapy base rate among each of the individual components. As SLP services represented approximately 17 percent, on average, of overall therapy costs, we believed it was appropriate to allocate this percentage as the base rate for the SLP component. If we were to make all three components equal, as one commenter had suggested, then this would overinflate SLP payment in relation to SLP costs. We would note, however, that while the base rate for the SLP component is lower than the other therapy component base rates, the case-mix weights for this component, as described in section V.B.3.c. of this final rule, are far greater for the SLP component than for either of the PT or OT components. This reflects that when SLP services are predicted to be necessary, there is adequate reimbursement for these services. Therefore, we expect that utilization of and access to SLP services should not be adversely affected merely because the base rate is lower for this component.

Accordingly, after considering the comments received, for the reasons specified in the FY 2019 SNF PPS proposed rule and in this final rule, we are finalizing, effective October 1, 2019, our proposals related to the calculation of the federal base payment rate components, as described in this section, with the following clarification. As discussed above, we are clarifying that, while we did eliminate the therapy non-case mix component from the model, we redistributed resources associated with this component across the five PDPM case-mix components as described in the PDPM technical report.

4. Updates and Wage Adjustments of Revised Federal Base Payment Rate Components

In section III.B. of the proposed rule, we described the process used to update the federal per diem rates each year. Additionally, as discussed in section III.B.4 of the proposed rule, SNF PPS rates are adjusted for geographic differences in wages using the most recent hospital wage index data. Under PDPM, we proposed to continue to update the federal base payment rates and adjust for geographic differences in wages following the current methodology used for such updates and wage index adjustments under the SNF PPS (83 FR 21040). Specifically, we proposed to continue the practice of using the SNF market basket, adjusted as described in section III.B. of the proposed rule to update the federal base payment rates and to adjust for geographic differences in wages as described in section III.B.4. of the proposed rule.

We received comments on the proposed methodology for updating the federal base payment rates and adjusting the per diem rates for geographic differences in wages under the PDPM. Those comments, and our responses, appear below.

Comment: Most commenters agreed with using the standard rate update policy and the existing wage index policy as the basis for updating the payment rates and adjusting the rates for geographic variation. One commenter stated that the lack of separate labor-share adjustment for each component may lead to provision of fewer services as each component would not be appropriately wage adjusted. This commenter stated that because CMS has already calculated payment amounts for each component and because cost reports contain all the information necessary to determine the labor share for each component, it would be appropriate for CMS to make separate wage adjustment calculations for each PDPM component.

Response: We appreciate the support for this proposal. With regard to the comment that CMS should separately wage adjust each PDPM component, the labor-related share reflects the facility Medicare-allowable costs (including all of the PDPM components) that are labor-intensive and vary with the local labor market. Specifically, it is equal to the following cost categories from the 2014-based SNF market basket: Wages and Salaries; Employee Benefits; Professional Fees: Labor-Related; Administrative and Facilities Support Services; Installation, Maintenance, and Repair Services; All Other: Labor-Related Services; and a proportion of Capital-Related expenses. The majority of these labor-related costs are derived using the MCR data; however, a notable portion is based on other government data sources. A complete description of the methodology used to derive the 2014-based SNF market basket is available in the FY 2018 final rule (82 FR 36548 through 36566). Given that these categories cut across PDPM components, to wage adjust for each component separately would require a substantial increase in the specificity of reporting these MCR data items, as well as developing a methodology for accurately assigning these costs to each component. We believe that the additional reporting burden associated with implementing this suggestion would not justify the increased specificity of applying the wage index adjustment to each component under PDPM.

Accordingly, after considering the comments received, for the reasons specified in the FY 2019 SNF PPS proposed rule (83 FR 21040) and discussed in this section, we are finalizing our proposal, without modification, for updating the federal base payment rates and for adjusting the per diem rates for geographic differences in wages under the PDPM, effective October 1, 2019.

C. Design and Methodology for Case-Mix Adjustment of Federal Rates

1. Background on PDPM

Section 1888(e)(4)(G)(i) of the Act requires that the Secretary provide an appropriate adjustment to account for case mix and that such an adjustment shall be based on a resident classification system that accounts for the relative resource utilization of different patient types. The current case-mix classification system uses a combination of resident characteristics and service intensity metrics (for example, therapy minutes) to assign residents to one of 66 RUGs, each of which corresponds to a therapy CMI and a nursing CMI, which are indicative of the relative cost to a SNF of treating residents within that classification category. However, as noted in section V.A. of the proposed rule, incorporating service-based metrics into the payment system can incentivize the provision of services based on a facility's financial considerations rather than resident needs. To better ensure that resident care decisions appropriately reflect each resident's actual care needs, we stated in the proposed rule (83 FR 21040) that we believe it is important to remove, to the extent possible, service-based metrics from the SNF PPS and derive payment from verifiable resident characteristics that are patient, and not facility, centered. To that end, as we stated in the proposed rule, the proposed PDPM was developed to be a payment model which derives payment classifications almost exclusively from verifiable resident characteristics.

Additionally, the current RUG-IV case-mix classification system reduces the varied needs and characteristics of a resident into a single RUG-IV group that is used for payment. As of FY 2017, of the 66 possible RUG classifications, over 90 percent of covered SNF PPS days are billed using one of the 23 Rehabilitation RUGs, with over 60 percent of covered SNF PPS days billed using one of the three Ultra-High Rehabilitation RUGs. As we stated in the proposed rule (83 FR 21040), the implication of this pattern is that more than half of the days billed under the SNF PPS effectively utilize only a resident's therapy minutes and Activities of Daily Living (ADL) score to determine the appropriate payment for all aspects of a resident's care. Both of these metrics, more notably a resident's therapy minutes, may not derive so much from the resident's own characteristics, but rather, from the type and amount of care the SNF decides to provide to the resident. We stated that even assuming that the facility takes the resident's needs and unique characteristics into account in making these service decisions, the focus of payment remains centered, to a potentially great extent, on the facility's own decision making and not on the resident's needs.

We explained in the proposed rule (83 FR 21041) that while the RUG-IV model utilizes a host of service-based metrics (type and amount of care the SNF decides to provide) to classify the resident into a single RUG-IV group, the proposed PDPM would separately identify and adjust for the varied needs and characteristics of a resident's care and combine this information together to determine payment. We stated we believe the proposed PDPM would improve the SNF PPS by basing payments predominantly on clinical characteristics rather than service provision, thereby enhancing payment accuracy and strengthening incentives for appropriate care. For these reasons, we proposed that, effective October 1, 2019, SNF residents would be classified using the PDPM, as further discussed below. As discussed in the proposed rule and in section V.I. of this final rule, we proposed to implement the PDPM on October 1, 2019 to allow all stakeholders adequate time for systems updates and staff training needed to assure smooth implementation.

2. Data Sources Utilized for Developing PDPM

To understand, research, and analyze the costs of providing Part A services to SNF residents, we utilized a variety of data sources in the course of research. In the proposed rule (83 FR 21041) and in this section, we discuss these sources and how they were used in the SNF PMR in developing the proposed PDPM. A more thorough discussion of the data sources used during the SNF PMR is available in section 3.1. of the SNF PDPM technical report (available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html).

a. Medicare Enrollment Data

Beneficiary enrollment and demographic information was extracted from the CMS enrollment database (EDB) and Common Medicare Environment (CME). Beneficiaries' Medicare enrollment was used to apply restrictions to create a study population for analysis. For example, beneficiaries were required to have continuous Medicare Part A enrollment during a SNF stay. Demographic characteristics (for example, age) were incorporated as being predictive of resource use. Furthermore, enrollment and demographic information from these data sources were used to assess the impact of the proposed PDPM on subpopulations of interest. In particular, the EDB and CME include indicators for potentially vulnerable subpopulations, such as those dually-enrolled in Medicaid and Medicare.

b. Medicare Claims Data

Medicare Parts A and B claims from the CMS Common Working File (CWF) were used to conduct claims analyses as part of the SNF PMR. SNF claims (CMS-1450 form, OMB control number 0938-0997), including type of bill (TOB) 21x (SNF Inpatient Part A) and 18x (hospital swing bed), were used to identify Medicare Part A stays paid under the SNF PPS. Part A stays were constructed by linking claims that share the same beneficiary, facility CMS Certification Number (CCN), and admission date. Stays created from SNF claims were linked to other claims data and assessment data via beneficiary identifiers.

Acute care hospital stays that qualified the beneficiary for the SNF benefit were identified using Medicare inpatient hospital claims. The dates of the qualifying hospital stay listed in the span codes of the SNF claim were used to connect inpatient claims with those dates listed as the admission and discharge dates. Although there are exceptions, the claims from the preceding inpatient hospitalization commonly contain clinical and service information relevant to the care administered during a SNF stay. Components of this information were used in the regression models predicting therapy and NTA costs and to better understand patterns of post-acute care (PAC) referrals for patients requiring SNF services. Additionally, the most recent hospital stay was matched to the SNF stay, which often (though not always) was the same as the preceding inpatient hospitalization, and used in the regression models.

Other Medicare claims, including outpatient hospital, physician, home health, hospice, durable medical equipment, and drug prescriptions, were incorporated, as necessary, into the analysis in one of three ways: (1) To verify information found on assessments or on SNF or inpatient claims; (2) to provide additional resident characteristics to test outside of those found in assessment and SNF and inpatient claims data; and (3) to stratify modeling results to identify effects of the system on beneficiary subpopulations. These claims were linked to SNF claims using beneficiary identifiers.

c. Assessment Data

Minimum Data Set (MDS) assessments were the primary source of resident characteristic information used to explain resource utilization in the SNF setting. The data repositories include MDS assessments submitted by SNFs and swing-bed hospitals. MDS version 2.0 assessments were submitted until October 2010, at which point MDS version 3.0 assessments began. MDS data were extracted from the Quality Improvement Evaluation System (QIES). MDS assessments were then matched to SNF claims data using the beneficiary identifier, assessment indicator, assessment date, and Resource Utilization Group (RUG).

d. Facility Data

Facility characteristics, while not considered as explanatory variables when modeling service use, were used for impact analyses. By incorporating this facility-level information, we could identify any disproportionate effects of the proposed case-mix classification system on different types of facilities.

Facility-level characteristics were taken from the Certification and Survey Provider Enhanced Reports (CASPER). From CASPER, we draw facility-level characteristics such as ownership, location, facility size, and facility type. CASPER data were supplemented with information from publicly available data sources. The principal data sources that are publicly available include the Medicare Cost Reports (Form 2540-10, 2540-96, and 2540-92) extracted from the Healthcare Cost Report Information System (HCRIS) files, Provider-Specific Files (PSF), Provider of Service files (POS), and Nursing Home Compare (NHC). These data sources have information on facility costs, payment, and characteristics that directly affect PPS calculations.

We received comments from stakeholders regarding the data used to develop PDPM, though we address these comments later in this section in relation to the specific PDPM component to which the comments were addressed.

3. Resident Classification Under PDPM

a. Background

As noted above, section 1888(e)(4)(G)(i) of the Act requires that the Secretary provide for an appropriate adjustment to account for case mix and that such an adjustment shall be based on a resident classification system that accounts for the relative resource utilization of different patient types. As we stated in the proposed rule (83 FR 21040), the proposed PDPM was developed to be a payment model which derives almost exclusively from resident characteristics. We stated that the proposed PDPM would separately identify and adjust five different case-mix components for the varied needs and characteristics of a resident's care and then combine these together with the non-case-mix component to form the full SNF PPS per diem rate for that resident.

We stated in the proposed rule (83 FR 21041 through 21042) that, as with any case-mix classification system based on resident characteristics, the proposed predictors that would be part of case-mix classification under PDPM are those which our analysis identified as associated with variation in costs for the given case-mix component. We explained that the proposed federal per diem rates discussed above serve as “base rates” specifically because they set the basic average cost of treating a typical SNF resident. Based on the presence of certain needs or characteristics, caring for certain residents may cost more or less than that average cost. We explained that a case-mix system identifies certain aspects of a resident or of a resident's care which, when present, lead to average costs for that group being higher or lower than the average cost of treating a typical SNF resident. For example, if we found that therapy costs were the same for two residents regardless of having a particular condition, then that condition will not be relevant in predicting increases in therapy costs. If, however, we found that, holding all else constant, the presence of a given condition was correlated with an increase in therapy costs for residents with that condition over those without that condition, then this could mean that this condition is indicative, or predictive, of increased costs relative to the average cost of treating SNF residents generally.

In the subsections that follow, we describe each of the five case-mix adjusted components under the proposed PDPM and the basis for each of the predictors that we stated would be used within the PDPM to classify residents for payment purposes.

b. Physical and Occupational Therapy Case-Mix Classification

As we stated in the proposed rule (83 FR 21042), a fundamental aspect of the proposed PDPM is to use resident characteristics to predict the costs of furnishing similarly situated residents with SNF care. Costs derived from the charges on claims and cost-to-charge ratios (CCRs) on facility cost reports were used as the measure of resource use to develop the proposed PDPM. We explained that costs better reflect differences in the relative resource use of residents as opposed to charges, which partly reflect decisions made by providers about how much to charge payers for certain services. We further explained that costs derived from charges are reflective of therapy utilization as they are correlated to the therapy minutes recorded for each therapy discipline. Under the current RUG-IV case-mix model, therapy minutes for all three therapy disciplines (PT, OT, SLP) are added together to determine the appropriate case-mix classification for the resident. However, as shown in section 3.3.1. of the SNF PDPM technical report (available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html), and as explained in the proposed rule, when we began to investigate resident characteristics predictive of therapy costs for each therapy discipline, we found that PT and OT costs per day are only weakly correlated with SLP costs per day (correlation coefficient of 0.04). The set of resident characteristics from the MDS that predicted PT and OT utilization was different than the set of characteristics predicting SLP utilization. Additionally, many predictors of high PT and OT costs per day predicted lower SLP costs per day, and vice versa. For example, we found that residents with cognitive impairments receive less physical and occupational therapy but receive more speech-language pathology. As a result of this analysis, as we explained in the proposed rule, we found that basing case-mix classification on total therapy costs per day obscured differences in the determinants of PT, OT, and SLP utilization.

In contrast, we stated in the proposed rule (83 FR 21042) that the correlation coefficient between PT and OT costs per day was high (0.62). Additionally, regression analyses found that predictors of high PT costs per day were also predictive of high OT costs per day. For example, the analyses found that late-loss ADLs are strong predictors of both PT and OT costs per day. We then used a range of resident characteristics to predict PT and OT costs per day separately and we found that the coefficients in both models followed similar patterns. Finally, we noted that resident characteristics were found to be better predictors of the sum of PT and OT costs per day than for either PT or OT costs separately. These analyses used a variety of items from the MDS as independent variables and used PT, OT, and SLP costs per day as dependent variables. In the proposed rule, we referred readers to section 3.3.1. of the SNF PMR technical report that accompanied the ANPRM available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html for more information on these analyses.

Given the results of this analytic work, as well as feedback from multiple stakeholders, we proposed three separate case-mix adjusted components, one corresponding to each therapy discipline: PT, OT, and SLP. In the original RCS-I model presented in the ANPRM, we stated that we were considering addressing PT and OT services through a single component, given the strong correlation between PT and OT costs and our finding that very similar predictors explained variation in the utilization of both therapy disciplines. However, as we explained in the proposed rule (83 FR 21042), commenters on the ANPRM stated that having a single combined PT and OT component could encourage providers to inappropriately substitute PT for OT and vice versa. We stated that this belief comports with feedback received from professional organizations and other stakeholders during technical expert panels (TEPs). The TEP commenters stated that PT and OT services should be addressed via separate components given the different aims of the two therapy disciplines and differences in the clinical characteristics of the resident subpopulations for which PT or OT services are warranted. For example, clinicians consulted during development of PDPM advised that personal hygiene, dressing, and upper extremity motion may bear a closer clinical relationship to OT utilization, while lower extremity motion may be more closely related to PT utilization. We stated in the proposed rule that while we do not believe that RCS-I, which included two separate components for PT/OT and SLP, contained stronger incentives for substitution across therapy disciplines compared to RUG-IV, which reimburses all three therapy disciplines through a single therapy component, we concur with the TEP commenters that PT and OT have different aims and that there are clinically relevant differences between residents who could benefit from PT, residents who could benefit from OT, and residents who could benefit from both disciplines. For the foregoing reasons, we decided to separate the combined PT/OT component presented in the ANPRM into two separate case-mix adjusted components in the proposed PDPM. As we stated in the proposed rule, because of the strong correlation between the dependent variables used for both components and the similarity in predictors, we decided to maintain the same case-mix classification model for both components. We stated that in practice, this means that the same resident characteristics will determine a resident's classification for PT and OT payment. However, we stated that each resident would be assigned separate case-mix groups for PT and OT payment, which correspond to separate case-mix indexes and payment rates. We explained that we believe providing separate case-mix-adjusted payments for PT and OT may allay concerns about inappropriate substitution across disciplines and encourage provision of these services according to clinical need. We further noted that as clinical practices evolve independently of incentives created by the current RUG-IV payment model, we would re-evaluate the different sets of resident characteristics that are predictive of PT and OT utilization after the PDPM is implemented. We stated that if based on this re-evaluation we determine that different sets of characteristics are predictive of PT and OT resource utilization, we could consider revising the payment model to better reflect clinical differences between residents who receive PT services and those who receive OT services.

After delineating the three separate case-mix adjusted therapy components, we continued our analysis, as described in the proposed rule (83 FR 21043), by identifying resident characteristics that were best predictive of PT and OT costs per day. To accomplish this, we conducted cost regressions with a host of variables from the MDS assessment, the prior inpatient claims, and the SNF claims that were believed to be potentially predictive of relative increases in PT and OT costs. As we stated in the proposed rule, the variables were selected with the goal of being as inclusive as possible with respect to characteristics related to the SNF stay and the prior inpatient stay. The selection also incorporated clinical input. We explained that these initial costs regressions were exploratory and meant to identify a broad set of resident characteristics that are predictive of PT and OT resource utilization. The results were used to inform which variables should be investigated further and ultimately included in the payment system. A table of all of the variables considered as part of this analysis appears in the appendix of the SNF PMR technical report that accompanied the ANPRM available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html. As explained in the proposed rule, based on our regression analyses, we found that the three most relevant categories of predictors of PT and OT costs per day were the clinical reasons for the SNF stay, the resident's functional status, and the presence of a cognitive impairment. More information on this analysis can be found in section 3.4.1. of the SNF PDPM technical report available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html.

Under the RUG-IV case-mix model, residents are first categorized based on being a rehabilitation resident or a non-rehabilitation resident, then categorized further based on additional aspects of the resident's care. As explained in the proposed rule (83 FR 21043), under the proposed PDPM, for the purposes of determining the resident's PT and OT groups and, as will be discussed below, the resident's SLP group, the resident would first be categorized based on the clinical reasons for the resident's SNF stay. We stated that empirical analyses demonstrated that the clinical basis for the resident's stay (that is, the primary reason the resident is in the SNF) is a strong predictor of therapy costs. For example, we explained that all of the clinical categories (described below) developed to characterize the primary reason for a SNF stay (except the clinical category used as the reference group) were found to be statistically significant predictors of therapy costs per day. More detail on these analyses can be found in section 3.4.1. of the SNF PMR technical report that accompanied the ANPRM (available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html ). In consultation with stakeholders (industry representatives, beneficiary representatives, clinicians, and payment policy experts) at multiple technical expert panels (TEPs), we created a set of ten inpatient clinical categories that we believe capture the range of general resident types which may be found in a SNF. These proposed clinical categories were provided in Table 14 of the proposed rule (83 FR 21043) and are reflected in Table 14.

Major Joint Replacement or Spinal Surgery.

Non-Surgical Orthopedic/Musculoskeletal.

Orthopedic Surgery (Except Major Joint Replacement or Spinal Surgery).

Acute Infections.

Medical Management.

Cancer.

Pulmonary.

Cardiovascular and Coagulations.

Acute Neurologic.

Non-Orthopedic Surgery.

We proposed to categorize a resident into a PDPM clinical category using item I8000 on the MDS 3.0. We stated in the proposed rule (83 FR 21043) that providers would use the first line in item I8000 to report the ICD-10-CM code that represents the primary reason for the resident's Part A SNF stay. We further stated that this code would be mapped to one of the ten clinical categories provided in Table 14 of the proposed rule (set forth at Table 14 of this final rule). The mapping between ICD-10-CM codes and the ten clinical categories is available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html. As explained in the proposed rule, the mapping indicates that in some cases, a single ICD-10-CM code maps to more than one clinical category because the care plan for a resident with this diagnosis may differ depending on the inpatient procedure history. In these cases, we explained that a resident may be categorized into a surgical clinical category if the resident received a surgical procedure during the immediately preceding inpatient stay that relates to the primary reason for the Part A SNF stay and typically requires extensive post-surgical rehabilitation or nursing care. If the resident did not receive a related surgical procedure during the prior inpatient stay that typically requires extensive post-surgical rehabilitation or nursing care, the resident may be categorized into a non-surgical clinical category. For example, we explained that certain wedge compression fractures that were treated with an invasive surgical procedure such as a fusion during the prior inpatient stay would be categorized as Major Joint Replacement or Spinal Surgery, but if these cases were not treated with a surgical procedure they would be categorized as Non-Surgical Orthopedic/Musculoskeletal. For residents who received a related surgical procedure during the prior inpatient stay, we stated that a provider would need to indicate the type of surgical procedure performed for the resident to be appropriately classified under PDPM. Thus, in these cases we proposed to require providers to record the type of inpatient surgical procedure performed during the prior inpatient stay so that residents can be appropriately classified into a PDPM clinical category for purposes of PT, OT, and SLP classification. We proposed that providers record the type of surgical procedure performed during the prior inpatient stay by coding an ICD-10-PCS code that corresponds to the inpatient surgical procedure in the second line of item I8000 in cases where inpatient surgical information is required to appropriately categorize a resident under PDPM. We noted that if we were to use the second line of item I8000 to record inpatient surgical information, we would provide a list of ICD-10-PCS codes that map to the surgical clinical categories. We stated that we believe this approach would allow for patients to be appropriately classified under the PDPM because it would provide sufficient information on the primary reason for SNF care and inpatient surgical procedures to assign a resident to the appropriate surgical or non-surgical clinical category. We invited comments on this proposal. In addition, we solicited comments on alternative methods for recording the type of inpatient surgical procedure to appropriately classify a patient into a clinical category. We explained that the clinical category into which the resident is classified would be used to classify the resident into a PT and OT category as discussed below, as well as an SLP category, as explained in section V.D.3.c. of the proposed rule.

As discussed above, we proposed to categorize a resident into a PDPM clinical category for purposes of PT, OT, and SLP classification using the ICD-10-CM code in the first line of item I8000, and if applicable, the ICD-10 PCS code in the second line of item I8000. As an alternative to using item I8000 to classify a resident into a clinical category, we stated in the proposed rule (83 FR 21044) that we were considering using a resident's primary diagnosis as reflected in MDS item I0020 as the basis for assigning the resident to a clinical category, and were evaluating the categories provided in item I0020 to determine if there is sufficient overlap between the categories used in item I0020 and the proposed PDPM clinical categories provided in Table 14 that this item could serve as the basis for a resident's initial classification into a clinical category under PDPM. We stated that the MDS item I0020 would require facilities to select a primary diagnosis from a pre-populated list of primary diagnoses representing the most common types of beneficiaries treated in a SNF, while item I8000, if used to assign residents to clinical categories, would require facilities to code a specific ICD-10-CM code that corresponds to the primary reason for the resident's Part A SNF stay. As indicated above, we also proposed that providers would code a specific ICD-10-PCS code in the second line of item I8000 when surgical information from the prior inpatient stay is necessary to assign a resident to a clinical category. We explained that if we were to use item I0020 to categorize residents under PDPM, we would not require providers to record additional information on inpatient surgical procedures as we expect the primary diagnosis information provided through item I0020 to be adequate to appropriately assign a resident to a clinical category.

We invited comments on our proposal to categorize a resident into a PDPM clinical category using the ICD-10-CM code recorded in the first line of item I8000 on the MDS 3.0, and the ICD-10-PCS code recorded on the second line of item I8000 on the MDS 3.0. In addition, we solicited comments on the alternative of using item I0020 on the MDS 3.0, as discussed above, as the basis for resident classification into one of the ten clinical categories in Table 14.

Commenters submitted the following comments related to the proposed rule's discussion of the clinical category assignments under PDPM. A discussion of these comments, along with our responses, appears below.

Comment: Several commenters expressed concern about ICD-10 coding requirements under PDPM. Some commenters are concerned that these requirements, especially ICD-10-PCS coding requirements, would create compliance risks because of SNFs' limited expertise in using ICD codes. A few commenters request that CMS offer ICD-10 coding training for clinicians, billers, coders, and other SNF personnel, prior and subsequent to PDPM implementation. Another commenter requested that CMS provide case studies and other resources as part of its educational strategy with respect to ICD-10 coding in the SNF setting. This commenter also recommended that CMS develop explicit instructions for how providers should record diagnosis and procedure information in MDS item I8000 for resident classification purposes under PDPM. One commenter recommends requiring the employment of credentialed medical record staff to ensure accurate coding. One commenter seeks clarification about potential consequences of ICD-10 coding errors during RAC audits. Another commenter questioned if the proposed ICD-10 coding is considered a transaction under the HIPPA transaction coding requirement.

Response: We appreciate the commenters' concerns regarding ICD-10 coding, but do not concur with stakeholder claims that SNF providers are unfamiliar with ICD-10 coding practices. While ICD-10 codes are not, in most instances, a factor in determining payment under the current SNF Part A benefit, ICD-10 has been an aspect of Medicare since FY 2016. Moreover, ICD-10 provides the most accurate coding and diagnosis information on patients, which can only serve to improve provider understanding of their patient's condition and resultant care plan. Finally, we believe that given the type of homogeneity of care currently provided by SNF providers (as evidenced by the existing case-mix distribution that has over 90 percent of patient billed days in one of 23 RUGs that utilize therapy service utilization as the primary determinant), moving to a system that utilizes the primary patient diagnosis as the key determinant of payment will help to ensure that the patient's unique condition and goals is the primary driver of care planning and care delivery and case mix classification, rather than the patient's ability to tolerate a high volume of therapy services.

With regard to the comment that CMS offer ICD-10 coding training for clinicians and other personnel, we do not believe it is the role of CMS to offer this type of professional training, as it is the responsibility of the provider to ensure that their staff is properly trained to perform these types of more general tasks that are not specific to a given payer or requirement. With regard to the comment that CMS provide case studies and other resources as part of an educational strategy, we appreciate this comment and will take it into consideration as we develop the educational materials for PDPM. In terms of the explicit instructions for how providers record diagnosis and procedure information, we do intend to provide such information in the MDS RAI manual.

With regard to the comment that we should require that providers employ credentialed medical record staff to ensure accurate coding, we agree that the emphasis on ICD-10 could cause changes in staffing at some providers. However, we do not believe it would be appropriate for CMS, in this instance, to specifically identify the type of staff that providers must employ to ensure accurate coding, as this is a decision best left to the provider. With regard to the potential consequences of ICD-10 coding errors on RAC audits, as under the current payment system, the information reported to CMS must be accurate. Inaccuracies in the data reported to CMS, or a failure to document the basis for such data, will necessitate the same types of administrative actions as occur today.

Finally, with regard to the question of whether the reporting of ICD-10 coding information constitutes a HIPAA transaction, we note that while some HIPAA Administrative Simplification requirements at 45 CFR part 162 require the use of ICD-10 codes, reporting ICD-10 codes does not in and of itself constitute a HIPAA transaction.

Comment: One commenter stated that CMS's proposal to use the first line of I8000 to capture the primary reason for SNF stay, the second line to capture procedure code, and the remaining spaces to capture comorbidities is overly complex. The commenter expressed concern that coding a procedure code in the second line of I8000 would not follow current RAI coding instructions. Some commenters support using MDS item I0020 to record the primary diagnosis, stating this will reduce provider burden. Other commenters opposed using item I0020 for this purpose because this item is designed for the Quality Reporting Program and does not align well with the PDPM clinical categories. One commenter stated that coding primary reason for SNF care in both item I8000 and item I0020 for different purposes will be confusing and will lead to errors. Another commenter sought clarity on whether providers would still be required to code ICD-10 diagnosis or procedure codes in item I8000 if item I0020 is used for resident classification. This commenter also questioned what providers should do if a resident does not fall into one of the I0020 categories. A few commenters suggest instead adding checkboxes in section I of the MDS to indicate the ten PDPM clinical categories. One commenter recommended the use of MDS item J2000 for procedure information, because SNFs have minimal experience with ICD-10-PCS codes and it can be difficult to obtain precise information on procedures performed during the preceding inpatient stay.

Response: We appreciate these comments regarding the complexity of the proposed methodology for collecting diagnosis and procedure information and appreciate the suggestions for ways to improve coding without compromising the overall integrity of the information reported. We agree with commenters who stated that the I0020 categories are not currently aligned with the clinical categories used under PDPM, specifically that the categories used under I0020 do not match the clinical categories that we use under PDPM, which means that using I0020 at this time would not be appropriate. We will continue to work to determine if refinements may be made in that item in the future which could allow for a transition to this item. With regard to comments concerning the potential for confusion associated with coding the patient's primary diagnosis in both I8000 and I0020 for different purposes, we believe this both affords the potential to confirm the primary diagnosis coding on the MDS (to the extent that we can identify areas of alignment between the two items) and helps us to refine the categories for a potential future transition to item I0020 under PDPM. With regard to the question of what providers should do if a patient does not fall into one of the I0020 categories, we would recommend that the provider refer to the I0020 coding instructions in the MDS manual for guidance on this issue.

With regard to suggestions of using a checkbox for recording diagnosis information, we believe that the use of such checkboxes for recording diagnosis information may not provide sufficient granularity for CMS to monitor properly the effects of PDPM implementation or to accurately classify patients for payment purposes, nor provide enough information for the SNF in terms of care planning. Given the use of ICD-10 diagnosis coding in other Medicare payment systems and given efforts to align payment across multiple post-acute care payment systems, we believe that using the actual diagnosis code, rather than a checkbox for a category, will provide greater consistency between payment systems and would provide a smoother transition to the extent such payment systems are aligned further in the future.

With regard to the comment that CMS consider using item J2000 to report procedural information, we believe that while the actual ICD-10 code is important in the case of diagnosis coding, we agree with the commenter that the procedural information may be coded at a more aggregated level, as this information is only being used to augment the patient's classification rather than as the primary basis for the classification. However, we believe that item J2000 (which requires providers to report if the patient experienced a surgical procedure in the preceding 100 days) would not adequately link to the care being delivered in the SNF, potentially close to 100 days after the surgical event. To address this, consistent with this commenter's suggestion, and in response to other concerns about the complexity of the proposed methodology, we believe that it would be appropriate and sufficient to develop subitems for item J2000 that would allow providers to report, through a checkbox-style mechanism, if a surgical procedure occurred during the preceding hospital stay (as opposed to the previous 100 days, as is used for J2000), and then provide a series of procedural categories, related to the PDPM clinical categories, that providers could select using a checkbox style mechanism, that would allow the provider to report on the relevant procedural information (rather than recording the specific ICD-10-PCS code). We believe this is a substantial improvement to the procedure we proposed for recording surgical procedure information, as it reduces the burden and complexity of provider reporting on procedural information while maintaining payment accuracy and integrity. Moreover, similar to how PDPM utilizes the procedural information to augment the patient's clinical category classification, we believe that using a checkbox mechanism also augments care-planning by helping ensure that the procedural history information from the hospital is properly taken into account in determining the resident's care needs and care plan. Therefore, we are developing sub-items for item J2000, which will allow providers to report the patient's procedural information in a way that uses a checkbox mechanism, and this procedural information will be used in concert with the patient's diagnosis information, as was discussed above and in the FY 2019 SNF PPS proposed rule, to classify the patient into a clinical category. We will provide both the subitems under item J2000, and the instructions regarding their use, for this purpose in the RAI manual.

Comment: One commenter was opposed to PDPM's focus on one primary diagnosis, as SNF residents can be admitted with complex medical conditions and multiple diagnoses. The commenter recommends that SNFs should select all resident conditions and allow the software to select the highest case-mix index achieved. In a similar vein, another commenter requested that CMS clarify which inpatient procedure SNFs should select for purposes of resident classification and payment under PDPM when the patient record includes multiple procedure codes.

Response: While we agree with the commenter that a SNF patient may suffer from multiple conditions, we believe that one of these reasons prompted transfer to the SNF. This reason would function as the patient's primary diagnosis, as it represents the primary reason for the patient being in the SNF. We would also note that primary diagnosis, as a concept, is used throughout the Medicare program as the basis for payment and, in each area in which it is used, patients have the potential to present with multiple conditions and multiple diagnoses. Therefore, we do not believe it would be appropriate for providers simply to report all conditions and be paid for the highest case-mix index, but rather that providers should determine the primary reason for the patient's stay, as this should also be the primary motivation behind the patient's SNF care.

With regard to the comment related to multiple inpatient surgical procedures, we expect that the checkbox mechanism discussed above, which would include more aggregated procedural groupings, should address much of this possibility, as often times multiple procedures may be done of the same type. In the case of different types of procedures, providers should code or check-off all information supportable by the patient's medical record.

Comment: One commenter stated that ICD-10 codes do not contain adequate specificity to indicate whether a condition is active/stable or active/non-stable. This information, according to the commenter, is needed to identify relevant comorbidities. As a result, the commenter states that SNFs may inappropriately use active/stable conditions to achieve higher reimbursement although these conditions may not indicate higher resource utilization.

Response: We do not agree with the commenter that the ICD-10 codes do not contain this degree of specificity. Further, to the extent that providers would code conditions solely for purposes of achieving higher reimbursement, this type of behavior can be identified through medical record reviews, which could prompt additional administrative action.

Comment: One commenter stated that chronic conditions may not be coded consistently year over year. Specifically, a chronic condition may be coded one year but not the following year for a long term care resident moving in and out of post-acute stays or a post-acute-care patient with more than one spell of illness. For example, the commenter noted that care may have been provided to the patient but the provider did not accurately capture it in reporting. The commenter further stated that such coding inconsistencies may lead to unexpected payment changes. The commenter recommended that CMS should clarify how chronic conditions should be reported and handled by medical reviewers as PDPM is implemented.

Response: We do not believe that any of the PDPM-related policies should affect the reporting of chronic conditions. Care should be properly documented, regardless of whether it is for a chronic or acute condition. Failure to document and code such information accurately could lead not only to payment errors, but also to patient care errors. We encourage providers to ensure the accuracy and completeness of their documentation.

Comment: Several commenters expressed concern about potential logistical issues arising from the time lag in SNFs receiving clinical information on admitted patients from the prior inpatient stay. Specifically, they state that it is difficult for SNF providers to obtain diagnosis and procedure information, as well as other clinical information such as discharge summaries, from the facility where a resident was treated during their prior inpatient stay. A commenter recommended that CMS require hospitals to provide diagnostic and procedural information within 48 hours of discharge to the receiving facility. This commenter requested that CMS clarify which medical records SNFs may rely upon to determine the principal reason for a SNF stay or which inpatient procedures were performed. The commenter questioned how SNFs should assess this information if they lack adequate documentation. Additionally, commenters stated that ICD-10-CM and ICD-10-PCS coding require a high level of clinical detail that may be difficult to obtain without clinical information from the prior inpatient stay.

Response: For case-mix classification under the PDPM, SNFs will not be required to collect any information from the hospital where the prior inpatient stay took place beyond that which is required under the current RUG-IV system, except for the procedural information discussed above. The information that SNFs already collect from hospitals should already include sufficient information for the SNF to be able to properly care plan and provide care based on the patient's condition. In order to do this effectively, SNFs should already be receiving documentation and records from the hospital that substantiate the need for care and the type of care that is required for that patient. This level of information, that is essential in developing an appropriate care plan for the patient, should be sufficient for addressing the payment requirements under PDPM. For proper classification and payment under PDPM, facilities will only be required to record the primary reason for SNF care at the time of SNF admission and record the associated ICD-10-CM code and procedural information. As discussed in Chapter 8 of the Medicare Benefit Policy Manual, a beneficiary in a Medicare Part A SNF stay must require skilled nursing care for a condition that was treated during the qualifying hospital stay, or for a condition that arose while in the SNF for treatment of a condition for which the beneficiary was previously treated in the hospital. However, CMS recognizes that in many cases, the primary reason for SNF care may not be the same as the primary reason for the prior inpatient stay. For example, a beneficiary may be treated in a SNF for a secondary condition that arose during the prior inpatient stay but that is different from the condition that precipitated the acute inpatient stay in the first place. PDPM requires facilities to code the diagnosis that corresponds most closely to the primary reason for SNF care (in this case, the secondary condition that arose during the hospital stay) rather than the primary reason for the prior hospitalization. Facilities currently must assess beneficiaries' health status and reason for SNF care at admission in order to treat them appropriately and formulate a patient-centered care plan. PDPM does not require a level of data collection that exceeds the requirements of the existing admission and care planning processes. Therefore, PDPM does not require SNFs to obtain additional clinical information from the inpatient setting, beyond the surgical procedure information discussed above.

Comment: One commenter recommended that CMS allow providers to correct the diagnosis or procedure information recorded at admission any time prior to discharge and to direct Medicare Administrative Contractors, Recovery Audit Contractors, and other contractors to assign low priority to reviewing ICD-10 codes in the medical review process.

Response: We appreciate the commenter's concern and would note that there are existing processes for modifying and correcting MDS assessments, as described in Chapter 5 of the MDS RAI manual. With regard to the comment on CMS directing contractor review activities, we see no reason to assign low priority to any issues at this time.

Comment: One commenter requested additional information about codes listed as “Return to Provider” in the PDPM Clinical Category Mapping. Specifically, the commenter requested that CMS provide clarity on why these codes are not accepted as valid primary diagnoses for the purposes of resident classification. Additionally, the commenter requests clarification on what actions providers are required to take when a code is returned.

Response: As discussed above and in the proposed rule (83 FR 21043), PDPM would use ICD-10-CM diagnosis codes entered in the first line of section I8000 on the MDS assessment to assign residents to clinical categories for classification and payment purposes in three PDPM payment components (PT, OT, and SLP). Codes listed in the PDPM Clinical Category Mapping as “Return to Provider” are not deemed appropriate to enter as the primary reason for SNF care. Such codes either lack certainty and specificity required to properly categorize a resident under PDPM or the underlying condition cannot be the main reason of care in SNFs. Therefore, these codes cannot be used to assign a resident to a clinical category for payment purposes under PDPM. When a code is returned to a provider, the provider is to select an appropriate ICD-10-CM diagnosis code from the SNF PDPM Clinical Category Mapping available at CMS' website.

Comment: Another commenter stated that the PDPM Clinical Category Mapping file inappropriately includes ICD-10-CM codes that correspond to an initial encounter. The commenter states that initial encounter codes include “A” as the 7th character and can only occur in a hospital where the initial treatment is completed. According to the commenter, initial encounter codes cannot be used in the SNF setting and should be excluded from the clinical category mapping. Additionally, the commenter states that Z codes are not appropriate to assign to patients receiving aftercare for traumatic fractures. These issues, state the commenter, lead to non-traumatic major joint replacements being assigned to Major Joint Replacement while major joint replacements as a result of traumatic injury are assigned to Orthopedic Surgery (Except Major Joint Replacement or Spinal Surgery). The commenter stated that this is inappropriate because aftercare of a traumatic injury resulting in hip replacement needs higher complexity of care than a scheduled non-traumatic hip replacement.

Response: We do not agree with the commenter's assertion that initial encounter codes cannot be used in the SNF and should be excluded from the clinical category mapping. Particularly given the increased focus of some commenters on the ability of PDPM to allow alignment across different payer types, we believe it is possible that some conditions could result as an initial encounter in the SNF. Moreover, as SNF services may be covered for conditions that arise in the hospital or arise in the SNF, we believe that it is important to allow for initial encounter codes to be coded within the SNF and mapped to clinical categories in case such a condition serves as the primary diagnosis for a SNF stay.

With regard to the comment that Z codes are not appropriate for traumatic fractures, as detailed in the ICD-10-CM Official Guidelines for Coding and Reporting, the aftercare codes cover situations when the initial treatment of a disease has been performed and the patient requires continued care during the healing or recovery phase, or for the long-term consequences of the disease. The aftercare Z codes should not be used if treatment is directed at a current, acute disease. Therefore, the aftercare Z codes should not be used for aftercare for traumatic fractures. For aftercare of a traumatic fracture, providers are instructed to assign the acute fracture code with the appropriate 7th character. We agree with the commenter and will update the PDPM mapping accordingly.

Comment: Some commenters expressed concern over the use of MS-DRGs to develop the PDPM clinical categories. Commenters noted that hospital MS-DRGs are unrelated to the reason for SNF admission and are poor predictors of cost in post-acute care. These commenters stated that if SNF MDS coding produces a substantially different set of case-mix adjustments from the case-mix derived from hospital DRG assignments, then the model will produce inappropriate payment rates for the cases which deviate from the “predicted” case mix rate. They suggested that CMS should consider retroactively evaluating this case-mix adjustment as soon as it has SNF data following PDPM implementation to correct any inaccurate payments in future updates of the PDPM. A commenter states that PDPM will need significant recalibration due to payment inaccuracies based on the discrepancy between inpatient hospital and SNF reason for admission.

Response: We appreciate the commenters' concerns with the use of MS-DRGs to develop the PDPM clinical categories. We would note, however, that while the MS-DRGs were used to identify patient categories in the SNF, they were not used to determine the cost of treating these types of patients. Given this distinction, while we might expect some difference in the distribution of SNF case-mix based on the potential differences between the prior hospital MS-DRG and SNF-generated diagnosis information under PDPM, we do not believe that using the MS-DRGs compromised the integrity of the clinical categories themselves. In developing PDPM clinical categories, we used MS-DRGs from the prior inpatient stay to define the primary reason for SNF care and assign residents to clinical categories. As stated in section 3.4.1 of the SNF PDPM technical report, we selected this source of diagnosis information because of data quality concerns relating to the principal diagnosis from the SNF claim. At the time the clinical categories were developed, we found that 47 percent of SNF claims assigned generic ICD-9-CM codes, with roughly a third assigned V57.89 “care involving other specified rehabilitation procedure”, as the principal diagnosis, limiting the usefulness of diagnoses from SNF claims in classifying residents. Per the Medicare Benefit Policy Manual, the SNF reason for admission must be related to a condition treated during the qualifying inpatient stay. Therefore, we believe it is reasonable to use clinical information from the prior inpatient stay to characterize the major types of beneficiaries who receive SNF care. Additionally, the clinical categories were validated by multiple clinicians consulted by CMS and participants at technical expert panels. Therefore, we believe the proposed clinical categories are appropriate to use to classify major clinical types found in the SNF setting. With regard to the possibility that the actual case-mix distribution may be distinct from the “predicted” case-mix distribution, we intend to monitor for these types of effects and may make adjustments to the payment rates as may be appropriate. We also appreciate the commenter's suggestion to recalibrate PDPM in the future.

Accordingly, after considering the comments received, for the reasons discussed above and in the proposed rule, we are finalizing our proposals discussed above relating to PT and OT case-mix classification under the PDPM, with the modification discussed below. As discussed above, rather than requiring providers to record the type of inpatient surgical procedure performed during the prior inpatient hospital stay by coding an ICD-10-PCS code in the second line of item I8000 as we proposed, we will instead require providers to select, as necessary, a surgical procedure category in a sub-item within Item J2000 which would identify the relevant surgical procedure that occurred during the patient's preceding hospital stay and which would augment the patient's PDPM clinical category.

(i) Clinical Categories

Once we identified these clinical categories as being generally predictive of resource utilization in a SNF, we then undertook the necessary work to identify those categories predictive of PT and OT costs specifically. As we discussed in the proposed rule (83 FR 21044), we conducted additional regression analyses to determine if any of these categories predicted similar levels of PT and OT as other categories, which may provide a basis for combining categories. As a result of this analysis, for the RCS-I model presented in the ANPRM, we found that the ten inpatient clinical categories could be collapsed into five clinical categories, which predict varying degrees of PT and OT costs. However, as explained in the proposed rule, we received comments on the ANPRM regarding the number of possible case-mix group combinations under RCS-I, so we sought to try and reduce this number of possible case-mix group combinations by further simplifying the model. As part of that effort, we observed similar PT and OT resource utilization patterns in the clinical categories of Non-Orthopedic Surgery and Acute Neurologic and, therefore, proposed to collapse these categories for the purpose of PT and OT classification. Additionally, as reflected in the RCS-I model presented in the ANPRM, we proposed that under PDPM, the remaining clinical categories would be collapsed as follows: Acute infections, cancer, pulmonary, cardiovascular and coagulations, and medical management would be collapsed into one clinical category entitled “Medical Management” because their residents had similar PT and OT costs. Similarly, we proposed that orthopedic surgery (except major joint replacement or spinal surgery) and non-surgical orthopedic/musculoskeletal would be collapsed into a new “Other Orthopedic” category for equivalent reasons. Finally, the remaining category, Major Joint Replacement, showed a distinct PT and OT cost profile and, thus, we proposed to retain it as an independent category. More information on this analysis can be found in section 3.4.2. of the SNF PMR technical report that accompanied the ANPRM and in section 3.4.2. of the SNF PDPM technical report, both available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html. These proposed collapsed categories, which would be used to categorize a resident initially under the proposed PT and OT case-mix components, were presented in Table 15 of the proposed rule (and are reflected in Table 15 of this final rule).

Table 15—Collapsed Clinical Categories for PT and OT Classification

PDPM clinical category	Collapsed PT and OT clinical category
Major Joint Replacement or Spinal Surgery	Major Joint Replacement or Spinal Surgery.
Non-Orthopedic Surgery	Non-Orthopedic Surgery and Acute Neurologic.
Acute Neurologic.
Non-Surgical Orthopedic/Musculoskeletal	Other Orthopedic.
Orthopedic Surgery (Except Major Joint Replacement or Spinal Surgery).
Medical Management	Medical Management.
Acute Infections.
Cancer.
Pulmonary.
Cardiovascular and Coagulations.

We received several comments regarding the collapsed PT and OT clinical categories. These comments, along with our responses, appear below.

Comment: A commenter disagreed with the decision to collapse the initial 10 clinical categories into five clinical groupings for purposes of resident classification and payment in the PT and OT components. The commenter stated that the five clinical categories used for resident classification in the PT and OT components are too broad and not representative of the clinical needs of residents. Another commenter recommends that CMS not finalize the proposal to combine the Acute Neurologic and Non-Orthopedic Surgery residents into a single category because patients should be classified based on clinically coherent categories, not on similar cost patterns observed under the current SNF case-mix classification model, for the latter is reflective of current reimbursement incentives to provide therapy based on financial considerations. A commenter suggests that CMS consider separate clinical category for elective major joint replacement of the lower extremity because its cost profile is different from other episode types. The commenter suggests that joint replacements as a result of a fracture could possibly be combined into the Other Orthopedic category.

Response: As described in section 3.4.2 of the SNF PMR technical report that accompanied the 2017 ANPRM, in developing RCS-I (the predecessor to PDPM), we created 10 broad clinical categories to characterize the major patient types found in the SNF setting. In using the CART algorithm to develop resident groups for PT and OT payment, we included the 10 clinical categories as a categorical variable. Allowing the CART algorithm to group the 10 clinical categories into a smaller number of groups resulted in fewer resident groups but a similar R-squared value for predicting costs. In building PDPM we first retained these five collapsed clinical categories to characterize major patient types relevant to predicting PT and OT utilization. As detailed in the proposed rule, we then further collapsed the clinical categories into four categories, in response to comments on the ANPRM regarding the number of possible case-mix group combinations under RCS-I. Based on the greater simplicity achieved in using fewer clinical categories for PT and OT classification and the maintenance in predictive accuracy, we believe using the collapsed four categories is a superior option to capture variation in PT and OT utilization and to characterize the major types of clinical conditions relevant to PT and OT utilization in the SNF population. Non-Orthopedic Surgery and Acute Neurologic are combined into one category based on their similar PT and OT resource utilization pattern, as shown in section 3.4.2 and Table 16 of the SNF PDPM technical report. We recognize that the observed data are reflective of current reimbursement incentives to provide therapy based on financial considerations, which may disguise the relationship between clinical traits and patient need based on best practice assumptions. We will monitor closely the resource utilization pattern of the 10 clinical categories after the implementation of PDPM. Regarding the elective major joint replacement comment, as detailed in section 3.4.1 of the SNF PMR technical report, we observed that MS-DRG groups with a high percentage of elective surgeries correspond to two types of procedures: Major joint replacements and spinal surgeries, while MS-DRG groups with a high percentage of emergent surgeries include other types of orthopedic surgeries involving extremities, often related to falls. We discovered that average therapy costs per day were similar for resident in a given surgical orthopedic MS-DRG group regardless of whether they received elective or emergent surgery.

Accordingly, after considering the comments received, for the reasons discussed above and in the proposed rule, we are finalizing our proposals without modification relating to the collapsed clinical categories for the PT and OT components.

(ii) Functional Status

As discussed previously in this section and in the proposed rule (83 FR 21044), regression analyses demonstrated that the resident's functional status is also predictive of PT and OT costs in addition to the resident's initial clinical categorization. In the RCS-I model discussed in the ANPRM, we presented a function score similar to the existing ADL score to measure functional abilities for the purposes of PT and OT payment. In response to the ANPRM, we received comments requesting that we consider replacing the functional items used to build the RCS-I function score with newer, IMPACT Act-compliant items from section GG. Therefore, we constructed, and proposed as discussed below, a new function score for PT and OT payment based on section GG functional items.

Under the RUG-IV case-mix system, a resident's ADL or function score is calculated based on a combination of self-performance and support items coded by SNFs in section G of the MDS 3.0 for four ADL areas: Transfers, eating, toileting, and bed mobility. These four areas are referred to as late-loss ADLs because they are typically the last functional abilities to be lost as a resident's function declines. Each ADL is assigned a score of up to four points, with a potential total score as high as 16 points. Under the proposed PDPM, we proposed that section G items would be replaced with functional items from section GG of the MDS 3.0 (Functional Abilities and Goals) as the basis for calculating the function score for resident classification used under PDPM. We explained that section GG offers standardized and more comprehensive measures of functional status and therapy needs. Additionally, we stated that the use of section GG items better aligns the payment model with other quality initiatives. SNFs have been collecting section GG data since October 2016 as part of the requirements for the IMPACT Act. We stated that given the advantages of section GG and of using a more comprehensive measure of functional abilities, we received numerous comments in response to the ANPRM requesting the incorporation of section GG items and of early ADLs items into the function score.

As explained in the proposed rule (83 FR 21045), multiple stakeholders commented on the ANPRM that late-loss items do not adequately reflect functional abilities on their own. These commenters stated that early-loss ADL items also capture essential clinical information on functional status. Therefore, we stated in the proposed rule that in building a new function score based on section GG items, we also investigated the incorporation of early-loss items. To explore the incorporation of section GG items, we evaluated each item's relationship with PT and OT costs. We ran individual regressions using each of the 12 section GG items assessed at admission to separately predict PT and OT costs per day. As explained in the proposed rule, the regression results showed that early-loss items are indeed strong predictors of PT and OT costs, with the exception of two wheeling items. Both wheeling items were excluded from the functional measure due to their weak predictive relationship with PT and OT costs. We observed high predictive ability among the remaining items. In total, we selected ten items for inclusion in the functional measure for the PT and OT components based on the results of the analysis. Thus, under the proposed functional measure for the PT and OT components, a resident's function would be measured using four late-loss ADL activities (bed mobility, transfer, eating, and toileting) and two early-loss ADL activities (oral hygiene and walking). Specifically, the proposed measure includes: Two bed mobility items, three transfer items, one eating item, one toileting item, one oral hygiene item, and two walking items that were all found to be highly predictive of PT and OT costs per day. A list of proposed section GG items that would be included in the functional measure for the PT and OT components was included in Table 18 of the proposed rule (and is shown in Table 18 of this final rule). Section 3.4.1. in the SNF PDPM technical report (available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html) provides more detail on these analyses.

We explained in the proposed rule (83 FR 21045) that, similar to the RUG-IV ADL score, each of these ADL areas would be assigned a score of up to 4 points. However, in contrast to the RUG-IV ADL score, we stated that points were assigned to each response level to track functional independence rather than functional dependence. In other words, higher points are assigned to higher levels of independence. We stated that this approach is consistent with functional measures in other care settings, such as the IRF PPS. Further, under the RUG-IV model, if the SNF codes that the “activity did not occur” or “occurred only once,” these items are assigned the same point value as “independent.” However, as explained in the proposed rule, we observed that residents who were unable to complete an activity had similar PT and OT costs as dependent residents. Therefore, we stated that when the activity cannot be completed, the equivalent section GG responses (“Resident refused,” “Not applicable,” “Not attempted due to medical condition or safety concerns”) are grouped with “dependent” for the purpose of point assignment. For the two walking items, we proposed an additional response level to reflect residents who skip the walking assessment due to their inability to walk. We stated that we believe this is appropriate because this allows us to assess the functional abilities of residents who cannot walk and assign them a function score. We explained that without this modification, we could not calculate a function score for residents who cannot walk because they would not be assessed on the two walking items included in the function score. We further stated that residents who are coded as unable to walk receive the same score as dependent residents to match with clinical expectations. In Tables 16 and 17 of the proposed rule (set forth at Tables 16 and 17 in this final rule), we provided the proposed scoring algorithm for the PT and OT functional measure.

Table 16—PT and OT Function Score Construction

[Except walking items]

Response	Score
05, 06—Set-up assistance, Independent	4
04—Supervision or touching assistance	3
03—Partial/moderate assistance	2
02—Substantial/maximal assistance	1
01, 07, 09, 88—Dependent, Refused, N/A, Not Attempted	0

Table 17—PT and OT Function Score Construction for Walking Items

Response	Score
05, 06—Set-up assistance, Independent	4
04—Supervision or touching assistance	3
03—Partial/moderate assistance	2
02—Substantial/maximal assistance	1
01, 07, 09, 88—Dependent, Refused, N/A, Not Attempted, Resident Cannot Walk *	0
* Coded based on response to GG0170H1 (Does the resident walk?).

We explained in the proposed rule (83 FR 21046) that, unlike section G, section GG measures functional areas with more than one item. We noted that this results in substantial overlap between the two bed mobility items, the three transfer items, and the two walking items. Because of this overlap, we stated that a simple sum of all scores for each item may inappropriately overweight functional areas measured by multiple items. Therefore, to adjust for this overlap, we proposed to calculate an average score for these related items. That is, we would average the scores for the two bed mobility items, the three transfer items, and the two walking items. We stated that the average bed mobility, transfer, and walking scores would then be summed with the scores for eating, oral hygiene, and toileting hygiene, resulting in equal weighting of the six activities. This proposed scoring algorithm produces a function score that ranges from 0 to 24. In section 3.4.1. of the SNF PDPM technical report (available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html), we provide additional information on the analyses that led to the construction of this proposed function score.

Table 18—Section GG Items Included in PT and OT Functional Measure

Section GG item	Score
GG0130A1—Self-care: Eating	0-4.
GG0130B1—Self-care: Oral Hygiene	0-4.
GG0130C1—Self-care: Toileting Hygiene	0-4.
GG0170B1—Mobility: Sit to lying	0-4 (average of 2 items).
GG0170C1—Mobility: Lying to sitting on side of bed.
GG0170D1—Mobility: Sit to stand	0-4 (average of 3 items).
GG0170E1—Mobility: Chair/bed-to-chair transfer.
GG0170F1—Mobility: Toilet transfer.
GG0170J1—Mobility: Walk 50 feet with 2 turns	0-4 (average of 2 items).
GG0170K1—Mobility: Walk 150 feet.

We received comments on the use of section GG items as the basis for determining the patient's PDPM functional score for purposes of classifying under the PT and OT components. Those comments, along with our responses, appear below.

Comment: Some comments welcomed the use of IMPACT-Act compliant section GG data to build new function scores for the PT, OT, and nursing components of PDPM, which was a recommendation provided by many commenters on the ANPRM. However, commenters also expressed concern about using section GG data, stating that this data should first be validated and that the results of this validation should be made public. Commenters stated that the first year of section GG data likely contains inaccuracies as providers adjust to the new items. Some commenters therefore stated that it is inappropriate to base resident classification and payment on a single year of section GG data and request that CMS collect a minimum of two years of section GG data to ensure reliability and validity before using this data to determine payment. One commenter suggested that, due to the issues with section GG, CMS should continue to use section G as the basis for functional assessment under the payment system. Another commenter sought clarification regarding whether CMS compared the first 6 months of section GG data to the second 6 months of section GG data to determine whether there were any changes in assessment practices for the new assessment items.

Response: We conducted several investigations to validate the section GG data. First, we verified that the relationship between section G responses and PT and OT utilization was very similar to the relationship between corresponding section GG responses and PT and OT utilization. Second, we determined that section GG items performed similarly to section G items in predicting PT and OT utilization. Finally, we compared coding of section GG items during the first 6 months of FY 2017 to coding of these items during the second 6 months of FY 2017 and found only small changes in the frequency of responses. Based on the results of these checks, we believe the FY 2017 section GG data are valid and reliable, and therefore, appropriate to use as a basis of resident classification and payment under PDPM.

Comment: One commenter stated that the proposed PDPM function scores ignore missing values for section GG assessment items and urged CMS to map missing values to a function score. Another commenter stated that the function score should incorporate the new response “10. Not attempted due to environmental limitations”. A few commenters requested that CMS consider assigning a score of 1 to “dependent” responses instead of 0, stating that this scoring aligns better with the SNF Quality Reporting Program. These commenters also seek clarification on the rationale for grouping “dependent” responses with “resident refused,” “not applicable,” and “not attempted due to medical conditions or safety concerns.” One commenter pointed out that the MDS item GG0170H1 (Does the resident walk) will be retired on September 30, 2018, and recommended that CMS adopt MDS item GG0170I (Walk 10 feet) as a substitute for retired item GG0170H1.

Response: We appreciate the comment that missing values for section GG assessments items are not currently mapped to a point value for computing function score. CMS will follow this suggestion to map all values to a function score by assigning missing section GG responses to receive zero points for the function score calculation as other incomplete responses are also assigned zero points. This is also consistent with the current RUG-IV ADL scoring methodology, which assigns the same point value for missing responses and other incomplete responses. Similarly, we will map the new response of 10: “Not attempted due to environmental limitations,” which was highlighted by another commenter, to receive zero points for function score assignment to make sure every response has a corresponding point value. We believe these point value assignments are appropriate as they are consistent with other similar responses that receive zero points for function score assignment, including “resident refused,” “not applicable,” and “not attempted due to medical condition or safety concerns”. In response to the comment requesting us to consider assigning 1 point to “dependent” responses instead of 0, this suggested scoring would group “dependent” responses with “substantial/maximal assistance” responses. However, we found that dependent residents have different levels of PT and OT resource utilization than residents receiving substantial/maximal assistance. As described in section 3.4.1 the SNF PDPM technical report, we observed that residents who were unable to complete an activity had similar PT and OT costs as dependent residents. Therefore, we grouped the equivalent section GG responses (“resident refused,” “not applicable,” and “not attempted due to medical condition or safety concerns”) with “dependent” responses for the purpose of point assignment in constructing the function score for PT and OT classification and payment. In terms of alignment with the SNF QRP quality measures, the PDPM function score uses similar scoring logic as the QRP functional outcome measure. As with the PDPM function score, the QRP Change in Self-Care score assigns higher points to higher levels of functional independence and assigns the same point value to “dependent” and incomplete responses. The QRP functional outcome measure, however, differs in scale. Whereas the PDPM function score ranges from 0-4, the QRP Change in Self-Care score ranges from 1-6. The QRP functional outcome measure assigns 1 point to “dependent” and all “activity was not attempted” codes (“resident refused,” “not applicable,” and “not attempted due to medical condition or safety concerns”), and 2 points to “substantial/maximal assistance”. This score assignment is very similar to that of the PDPM function score. Additionally, one item currently used to compute function score, MDS GG0170H1 (Does the resident walk), which is used to determine if the resident can walk before proceeding to assess GG0170J1 (Walk 50 Feet with Two Turns) and GG0170K1 (Walk 150 Feet), is set to be retired on September 30, 2018 with the introduction of the newer, more detailed SNF QRP mobility and self-care outcome measure items. CMS concurs with the commenter's suggestion to select a replacement for PDPM implementation. Consistent with the commenter's suggestion, MDS item GG0170I1 (Walk 10 feet) will be used as the substitute for MDS GG0170H1 since the inability to walk at least 10 feet or to complete the assessment for this item suggests a significant mobility impairment that is essentially equivalent to the definition of the retired “cannot walk” MDS item. Responses 07: “resident refused,” 09: “not applicable,” 10: “not attempted due to environmental limitations,” or 88: “not attempted due to medical condition or safety concerns” from MDS item GG0170I1 will be used to identify residents who cannot walk.

Comment: Commenters also stated that the proposed function scores should be updated to reflect new section GG items for FY 2019. Specifically, they stated that toileting, dressing, and bathing are important activities of daily living that are addressed by occupational therapy, and therefore, should be considered in measuring residents' functional status under PDPM.

Response: In constructing the function score for PT and OT payment, we investigated the use of all existing section GG items. Toileting is one of the items included in the proposed function scores for the PT, OT, and nursing components of PDPM. We are aware that additional section GG items are scheduled to be implemented in FY 2019, including items that measure a resident's dressing and bathing abilities. However, because these new items have not yet been implemented, there is no data available on resource utilization associated with these items. Therefore, it is not appropriate to include these items in the calculation of the PDPM function scores at this time. We will consider adding section GG items that are demonstrated to have a meaningful relationship with utilization of SNF resources as new items are added and an appropriate amount of data (for example, one year) is available to assess this relationship. We will also consider other changes to the function score as necessary to reflect additional updates to the section GG items, for example, the addition, deletion, or modification of particular items or responses.

Comment: One commenter advised CMS to account for weight bearing restrictions among residents who are categorized into the Major Joint Replacement or Spinal Surgery or Other Orthopedic clinical categories. The commenter stated that patients who cannot bear weight have a more complicated post-surgical recovery.

Response: We appreciate the concern of the commenter regarding post-surgical residents who cannot bear weight. However, we believe the ability of a resident to bear weight is adequately captured by the mobility items in MDS item GG0170, which are included in the function score used for classification and payment in the PT and OT components. Therefore, we do not believe additional modifications are necessary at this time.

Comment: One commenter noted that in some cases, PT and OT payment is higher for case-mix groups with higher functional independence. The commenter said this is counterintuitive because it implies that some residents who are more dependent require less therapy. Another commenter sought clarification on the relationship between function score and average PT and OT costs per day.

Response: The commenter is correct that in some cases payment is higher for residents who have higher levels of functional independence. This reflects the finding that PT and OT utilization is highest for residents with moderate functional independence and lower for residents with both the highest levels of functional dependence and independence. In the first case, this likely reflects residents whose functional abilities are too impaired to receive intensive therapy, while the second case likely corresponds to residents who require less therapy because they already have a high level of functional independence. Therefore, we believe PDPM appropriately assigns payment according to the observed relationship between functional independence and PT/OT utilization.

Comment: One commenter expressed concern regarding the potential for gaming the function score and recommended that CMS remove the function score from use as a patient classifier.

Response: We appreciate this concern for gaming of the function score and plan to monitor closely for any changes in functional coding before and after implementation of PDPM. That being said, we do believe that a patient's functional score is relevant in terms of predicting payment accurately, as described elsewhere in this section. Therefore, we believe it is important to keep function as an aspect of patient classification for payment.

Accordingly, after considering the comments received, for the reasons discussed in the proposed rule and in this final rule, we are finalizing our proposals relating to the use of the section GG items as the basis for determining the patient's PDPM functional score and for classifying the patient under PDPM PT and OT components, with modifications. As discussed above, in response to comments, all missing values for section GG assessment items will receive zero points as a function score. Similarly, the function score will incorporate a new response “10. Not attempted due to environmental limitations” and we will assign it a point value of zero. Furthermore, consistent with a commenter's suggestion, we will adopt MDS item GG0170I1 (Walk 10 feet) as a substitute for retired item GG0170H1 (Does the resident walk), and we will use responses 07: “resident refused,” 09: “not applicable,” 10: “not attempted due to environmental limitations,” or 88: “not attempted due to medical condition or safety concerns” from MDS item GG0170I1 to identify residents who cannot walk.

(iii) Cognitive Status

Under the RCS-I case-mix model presented in the ANPRM, we used cognitive status to classify residents under the PT and OT components in addition to the primary reason for SNF care and functional ability. As explained in the proposed rule (83 FR 21046) and in greater detail below, after publication of the ANPRM, we removed cognitive status as a determinant of resident classification for the PT and OT components. Still, although cognitive status was not ultimately selected as a determinant of PT and OT classification, it was considered as a possible element in developing the proposed resident groups for these components via the Classification and Regression Trees (CART) algorithm described in greater detail in the proposed rule and below. Because we included cognitive status as an independent variable in the CART analysis used to develop case-mix groups for PT and OT, we stated that we believed it was appropriate to discuss construction of the proposed new cognitive measure here even though it was not ultimately selected as a determinant of payment for PT and OT. Thus, we discussed construction of the instrument used to measure cognitive status under the proposed PDPM in the section addressing case-mix classification under the PT and OT components, rather than introducing it when discussing SLP classification, in which we proposed cognitive status as a determinant of resident classification. Under the current SNF PPS, cognitive status is used to classify a small portion of residents that fall into the Behavioral Symptoms and Cognitive Performance RUG-IV category. For all other residents, cognitive status is not used in determining the appropriate payment for a resident's care. However, as we explained in the proposed rule, industry representatives and clinicians at multiple TEPs suggested that a resident's cognitive status can have a significant impact on a resident's PT and OT costs. Based on this feedback, we explored a resident's cognitive status as a predictor of PT and OT costs.

Under the RUG-IV model, cognitive status is assessed using the Brief Interview for Mental Status (BIMS) on the MDS 3.0. The BIMS is based on three items: “repetition of three words,” “temporal orientation,” and “recall.” These items are summed to produce the BIMS summary score. The BIMS score ranges from 0 to 15, with 0 assigned to residents with the worst cognitive performance and 15 assigned to residents with the highest performance. Residents with a BIMS score less than or equal to 9 classify for the Behavioral Symptoms and Cognitive Performance category. Residents with a summary score greater than 9 but not 99 (resident interview was not successful) are considered cognitively intact for the purpose of classification under RUG-IV.

As we explained in the proposed rule (83 FR 21046), in approximately 15 percent of 5-day MDS assessments, the BIMS is not completed: in 12 percent of cases the interview is not attempted, and for 3 percent of cases the interview is attempted but cannot be completed. The MDS directs assessors to skip the BIMS if the resident is rarely or never understood (this is scored as “skipped”). In these cases, the MDS requires assessors to complete the Staff Assessment for Mental Status (items C0700 through C1000). The Cognitive Performance Scale (CPS) is then used to assess cognitive function based on the Staff Assessment for Mental Status and other MDS items (“Comatose” (B0100), “Makes Self Understood” (B0700), and the self-performance items of the four late-loss ADLs). The Staff Assessment for Mental Status consists of four items: “Short-term Memory OK,” “Long-term Memory OK,” “Memory/Recall Ability,” and “Cognitive Skills for Daily Decision Making.” Only “Short-term Memory OK” and “Cognitive Skills for Daily Decision Making” are currently used for payment. In MDS 2.0, the CPS was used as the sole measure of cognitive status. A resident was assigned a CPS score from 0 to 6 based on the Staff Assessment for Mental Status and other MDS items, with 0 indicating the resident was cognitively intact and 6 indicating the highest level of cognitive impairment. In addition to the items on the Staff Assessment for Mental Status, MDS items “Comatose” (B0100), “Makes Self Understood” (B0700), and the self-performance items of the four late-loss ADLs factored into the CPS score. Any score of 3 or above was considered cognitively impaired. The CPS on the current version of the MDS (3.0) functions very similarly. Instead of assigning a score to each resident, a resident is determined to be cognitively impaired if he or she meets the criteria to receive a score of 3 or above on the CPS, based on the MDS items mentioned above. In other words, whereas the MDS 2.0 assigned a CPS score to each resident, the MDS 3.0 only determines whether a resident's score is greater than or equal to 3 and does not assign a specific score to each resident for whom the CPS is used to assess cognitive status. Residents who are determined to be cognitively impaired based on the CPS are classified in the Behavioral Symptoms and Cognitive Performance category under RUG-IV, if they do not meet the criteria for a higher-paying category.

We stated in the proposed rule (83 FR 21047) that given that the 15 percent of residents who are not assessed on the BIMS must be assessed using a different scale that relies on a different set of MDS items, there is currently no single measure of cognitive status that allows comparison across all residents. To address this issue, Thomas et al., in a 2015 paper, proposed use of a new cognitive measure, the Cognitive Function Scale (CFS), which combines scores from the BIMS and CPS into one scale that can be used to compare cognitive function across all residents (Thomas KS, Dosa D, Wysocki A, Mor V; The Minimum Data Set 3.0 Cognitive Function Scale. Med Care. https://www.ncbi.nlm.nih.gov/​pubmed/​?term=​25763665). Following a suggestion from the June 2016 TEP, we explored using the CFS as a measure of cognition and found that there is a relationship between the different levels of the cognitive scale and resident costs. Specifically, we observed that as cognitive function declines, PT and OT costs per day decrease, while SLP costs per day more than double. More information on this analysis can be found in section 3.4.1. of the SNF PMR technical report that accompanied the ANPRM available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html. Based on these initial investigations, we used the CFS as a cognitive measure in the RCS-I payment model described in the ANPRM. As we noted above, the RUG-IV system incorporates both the BIMS and CPS score separately, but the CFS blends them together into one measure of cognitive status. Details on how the BIMS score and CPS score are determined using the MDS assessment are described above. The CFS uses these scores to place residents into one of four cognitive performance categories, as shown in Table 19 of the proposed rule (set forth in Table 19 of this final rule). After publication of the ANPRM, we received stakeholder comments questioning this scoring methodology, specifically the classification of a CPS score of 0 as “mildly impaired.” Based on a subsequent analysis showing that residents with a CPS score of 0 were similar to residents classified as “cognitively intact” under the CFS methodology, as well as clinical feedback, we determined that it was appropriate to reclassify residents with a CPS score of 0 as cognitively intact, consistent with ANPRM feedback. This analysis is described in more detail in section 3.4.1. of the SNF PDPM technical report, available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html. The scoring methodology for the proposed PDPM cognitive measure was shown in Table 20 of the proposed rule (set forth in Table 20 of this final rule). We would note once again that while we discussed this scoring methodology in section V.D.3.b of the proposed rule (83 FR 21046 through 21047) and this section of the final rule because cognitive status was considered in developing the PT and OT classification, the cognitive score was not proposed as a factor in classification for the PT and OT components under PDPM, as further discussed in the proposed rule (83 FR 21047) and below.

Table 19—Cognitive Function Scale (CFS) Scoring Methodology

Cognitive level	BIMS score	CPS score
Cognitively Intact	13-15
Mildly Impaired	8-12	0-2
Moderately Impaired	0-7	3-4
Severely Impaired		5-6

Table 20—PDPM Cognitive Measure Classification Methodology

Cognitive level	BIMS score	CPS score
Cognitively Intact	13-15	0
Mildly Impaired	8-12	1-2
Moderately Impaired	0-7	3-4
Severely Impaired		5-6

(iv) PT and OT Case-Mix Groups

As explained in the proposed rule (83 FR 21047), once each of these variables—clinical reasons for the SNF stay, the resident's functional status, and the presence of a cognitive impairment—was identified, we then used a statistical regression technique called Classification and Regression Trees (CART) to explore the most appropriate splits in PT and OT case-mix groups using these three variables. In other words, CART was used to investigate how many PT and OT case-mix groups should exist under the proposed PDPM and what types of residents or score ranges should be combined to form each of those PT and OT case-mix groups. CART is a non-parametric decision tree learning technique that produces either classification or regression trees, depending on whether the dependent variable is categorical or numeric, respectively. We stated that using the CART technique to create payment groups is advantageous because it is resistant to both outliers and irrelevant parameters. The CART algorithm has been used to create payment groups in other Medicare settings. For example, it was used to determine Case Mix Groups (CMGs) splits within rehabilitation impairment groups (RICs) when the inpatient rehabilitation facility (IRF) PPS was developed. This methodology is more thoroughly explained in section 3.4.2. of the SNF PDPM technical report (available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html).

As explained in the proposed rule (83 FR 21047), we used CART to develop splits within the four collapsed clinical categories shown in Table 15 of the proposed rule (set forth in Table 15 of this final rule). Splits within each of these four collapsed clinical categories were based on the two independent variables included in the algorithm: Function score and cognitive status. The CART algorithm split residents into 18 groups for the PT component and 14 groups for the OT component. These splits are primarily based on differences in resident function. As stated in the proposed rule, in the CART-generated groups, cognitive status plays a role in categorizing less than half of the PT groups and only two of the 14 OT groups. In addition, we stated that to create the proposed resident classification for the PT and OT components, we made certain administrative decisions that further refined the PT and OT case-mix classification groups beyond those produced through use of the CART algorithm. For example, while CART may have created slightly different breakpoints for the function score in different clinical categories, we state that we believe using a consistent split in scores across clinical categories improves the simplicity of the case-mix model without compromising its accuracy. Therefore, we used the splits created by the CART algorithm as the basis for the consistent splits selected for the case-mix groups, simplifying the CART output while retaining important features of the CART-generated splits. In our proposed classification for the PT and OT components, we retained function as the sole determinant of resident categorization within each of the four collapsed clinical categories. We created function score bins based on breakpoints that recurred in the CART splits, such as 5, 9, and 23. As noted in the proposed rule (83 FR 21048) and above, we dropped cognitive status as a determinant of classification because of the reduced role it played in categorizing residents within the CART-generated groups. Finally, we used the same function score bins to categorize residents within each of the four collapsed clinical categories for both the PT and OT components. As shown in section 3.4.2. of the SNF PDPM technical report (available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.htm), and as explained in the proposed rule, using the proposed case-mix groups for the PT and OT components results in a reduction of 0.005 in the R-squared values for both PT and OT classification models. We stated that this shows that although the proposed case-mix groups improve simplicity by removing one predictor revealed to be less important in categorizing residents (cognitive status) and grouping residents similarly (using the same function score bins) across clinical categories, these decisions have only a minor negative impact on predictive accuracy. These analyses are described in further detail in section 3.4.2. of the SNF PDPM technical report (available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html).

Based on the CART results and the administrative decisions described above, we proposed 16 case-mix groups to classify residents for PT and OT payment. We noted in the proposed rule (83 FR 21048) that this represents a marked reduction in the number of case-mix groups for PT and OT classification under the RCS-I model discussed in the ANPRM. As discussed in the proposed rule and throughout the sections above, after publication of the ANPRM, we received feedback from stakeholders that the RCS-I payment model was overly complex. In particular, commenters expressed concern about the relatively large number of possible combinations of case-mix groups. Based on this feedback, we sought to reduce the number of resident groups in the PT and OT components. First, as discussed in the proposed rule and in this final rule, because we observed similar PT and OT resource utilization patterns in the clinical categories of Non-Orthopedic Surgery and Acute Neurologic, we decided to collapse these categories for the purpose of PT and OT classification. In addition, as discussed in the proposed rule and in this final rule, we replaced the section G-based functional measure from RCS-I with a new functional measure based on section GG items. We found that the inclusion of the section GG-based functional measure in the CART algorithm resulted in case-mix groups in which cognitive function played a less important role in classification. Based on these results, we determined that we could remove cognitive function as a determinant of PT and OT classification without a notable loss in the predictive ability of the payment model, as discussed above. We also consulted with clinicians who advised CMS during development of PDPM, who confirmed the appropriateness of this decision. We stated in the proposed rule that the decisions to collapse Non-Orthopedic Surgery and Acute Neurologic into one clinical category and remove cognitive status resulted in a large reduction in the number of PT and OT case-mix groups, from the 30 in RCS-I to the 16 in the proposed PDPM provided in Table 21 of the proposed rule (and set forth in Table 21 of this final rule). We provided the criteria for each of these groups along with its CMI for both the PT and OT components in Table 21. As shown in Table 21, two factors would be used to classify each resident for PT and OT payment: Clinical category and function score. Each case-mix group corresponds to one clinical category and one function score range. We proposed classifying each SNF resident into one of the 16 groups shown in Table 21 based on these two factors.

To help ensure that payment reflects the average relative resource use at the per diem level, we stated in the proposed rule (83 FR 21048) that CMIs would be set to reflect relative case-mix related differences in costs across groups. We stated that this method helps ensure that the share of payment for each case-mix group would be equal to its share of total costs of the component. We further explained that CMIs for the PT and OT components were calculated based on two factors. One factor was the average per diem costs of a case-mix group relative to the population average. The other factor was the average variable per diem adjustment factor of the group relative to the population average. In this calculation, average per diem costs equaled total PT or OT costs in the group divided by number of utilization days in the group. Similarly, the average variable per diem adjustment factor equaled the sum of variable per diem adjustment factors corresponding to a given component (PT or OT) for all utilization days in the group divided by the number of utilization days in the group. We calculated CMIs such that they equal the ratio of relative average per diem costs for a group to the relative average variable per diem adjustment factor for the group. In this calculation, relative average per diem costs and the relative average variable per diem adjustment factor were weighted by length of stay to account for the different length of stay distributions across case-mix groups (as further discussed in section 3.11.1. of the SNF PDPM technical report, available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html). The relative average variable per diem adjustment factors for a given PT group and the corresponding OT group were the same because residents were classified into the same case-mix group under both components. However, relative average per diem costs were different across the two corresponding PT and OT groups, therefore the resulting CMIs calculated for each group were different, as shown in Table 21. After calculating CMIs as described above, we then applied adjustments to help ensure that the distribution of resources across payment components is aligned with the statutory base rates. We stated that the base rates implicitly allocate resources to case-mix components in proportion to the relative magnitude of the respective component base rates. For example, if the base rate for one component were twice as large as the base rate for another component, this would imply that the component with the larger base rate should receive double the resources of the other component. To ensure that the distribution of resources across payment components was aligned with the statutory base rates, in the proposed rule, we set CMIs such that the average product of the CMI and the variable per diem adjustment factor for a day of care equals 1.0 for each of the five case-mix-adjusted components in PDPM. If the average product of the CMI and the variable per diem adjustment factor for a day of care were different across case-mix components, this would result in allocating resources in a manner inconsistent with the distribution of resources implied by the statutory base rates.

After adjusting the CMIs to align the distribution of resources across payment components with the statutory base rates, a parity adjustment was then applied by multiplying the CMIs by the ratio of case-mix-related payments in RUG-IV over estimated case-mix-related payments in PDPM, as further discussed in section V.J. of the proposed rule. More information on the variable per diem adjustment factors is discussed in section V.D.4. of the proposed rule. The full methodology used to develop CMIs is presented in section 3.11. of the SNF PDPM technical report (available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html).

Table 21—PT and OT Case-Mix Classification Groups

Clinical category	Section GG function score	PT OT case-mix group	PT case-mix index	OT case-mix index
Major Joint Replacement or Spinal Surgery	0-5	TA	1.53	1.49
Major Joint Replacement or Spinal Surgery	6-9	TB	1.69	1.63
Major Joint Replacement or Spinal Surgery	10-23	TC	1.88	1.68
Major Joint Replacement or Spinal Surgery	24	TD	1.92	1.53
Other Orthopedic	0-5	TE	1.42	1.41
Other Orthopedic	6-9	TF	1.61	1.59
Other Orthopedic	10-23	TG	1.67	1.64
Other Orthopedic	24	TH	1.16	1.15
Medical Management	0-5	TI	1.13	1.17
Medical Management	6-9	TJ	1.42	1.44
Medical Management	10-23	TK	1.52	1.54
Medical Management	24	TL	1.09	1.11
Non-Orthopedic Surgery and Acute Neurologic	0-5	TM	1.27	1.30
Non-Orthopedic Surgery and Acute Neurologic	6-9	TN	1.48	1.49
Non-Orthopedic Surgery and Acute Neurologic	10-23	TO	1.55	1.55
Non-Orthopedic Surgery and Acute Neurologic	24	TP	1.08	1.09

We stated in the proposed rule that, under the proposed PDPM, all residents would be classified into one and only one of these 16 PT and OT case-mix groups for each of the two components. We explained that as opposed to the RUG-IV system that determines therapy payments based only on the amount of therapy provided, these groups classify residents based on the two resident characteristics shown to be most predictive of PT and OT utilization: Clinical category and function score. Thus, we believe that the PT and OT case-mix groups better reflect relative resource use of clinically relevant resident subpopulations, and therefore, provide for more appropriate payment under the SNF PPS.

Commenters submitted the following additional comments related to the proposed rule's discussion of the Physical and Occupational Therapy Case-Mix Classification. A discussion of these comments, along with our responses, appears below.

Comment: Several commenters expressed concern that CMS did not include cognition or swallowing disorders as determinants of payment for the OT component. One commenter stated that the removal of cognitive status as a determinant of PT and OT payment will lead to underpayment because cognitive impairment leads to longer recovery time and an increased need for therapy services, particularly occupational therapy.

Response: As discussed in the proposed rule (83 FR 21046) and in section 3.4.2 of the SNF PDPM technical report, cognitive status was initially considered as a determinant of resident classification and payment in the PT and OT components of PDPM. However, after replacing the section G-based function score for PT and OT classification with a function score based on new, IMPACT Act-compliant section GG items, we reran the CART analysis used to develop possible case-mix groups. We found that after including the section GG-based function score, cognitive status played a minimal role in resident classification. As noted in the proposed rule (83 FR 21047), cognition played a role in categorizing less than half of the 18 CART-generated PT groups and only two of the 14 CART-generated OT groups. Based on the reduced role of cognition in resident classification for PT and OT payment, we decided to remove cognitive status as a determinant of payment for these components. This decision also allowed us to substantially reduce the number of case-mix groups for the PT and OT components from the 30 presented in the 2017 ANPRM to the 16 presented in the proposed rule, contributing to a simplification of the payment model, which was requested by a number of commenters responding to the ANPRM. We also confirmed that the decision to remove cognitive status as a determinant of PT and OT classification had only a minor negative impact on predictive accuracy, reducing the R-squared values of the both the PT and OT classification models by only 0.005.

Comment: One commenter expressed concern about the reliability of the cognitive measure used in PDPM.

Response: As detailed in section 3.4.1 of the SNF PDPM technical report, the PDPM cognitive measure was built based on two existing cognitive measures: The Brief Interview for Mental Status (BIMS) and the Cognitive Performance Scale (CPS). Both measures are used in the current RUG-IV system to determine cognitive impairment. BIMS is used when the resident is able to complete the interview, while CPS is used when the resident is unable to complete the interview and the staff assessment has to be conducted. Thus, the PDPM cognitive measure is based on cognitive measures that have been validated and used for years. It combines the existing scores from BIMS and CPS into one scale that can be used to compare cognitive function across all residents.

Comment: Some commenters stated that CMS should consider including comorbidities related to PT or OT utilization, in particular conditions associated with high therapy intensity or duration. Commenters stated coronary artery disease, congestive heart failure, diabetes, chronic obstructive pulmonary disease (COPD), asthma, chronic wounds, depression, swallowing disorders and multiple sclerosis are conditions that could be considered as possible comorbidities for the PT or OT components.

Response: As described in section 3.4.1 of the SNF PDPM technical report, we investigated the impact of a broad list of conditions on PT and OT utilization. These conditions were selected for investigation based on comments received in response to the 2017 ANPRM, clinical input, and a literature search. This broad list included several of the conditions mentioned by commenters, including congestive heart failure, diabetes, depression, and swallowing disorders. To focus on conditions that have non-negligible impact on increasing costs, only those that had a positive impact on PT or OT costs per day of $2 or more were selected for further investigation. None of the conditions mentioned by commenters that were included in this investigation (congestive heart failure, diabetes, depression, swallowing disorders, and multiple sclerosis) met this criterion; therefore, they were not selected for inclusion in the payment model. Additionally, as mentioned in section 3.4.1 of the SNF PDPM technical report, we investigated the impact of an even broader range of MDS items, diagnosis-related groups (DRGs), and hierarchical condition categories (HCCs) on PT and OT utilization. Among the conditions included in this analysis were coronary artery disease, COPD, asthma, and various types of wounds/wound care including wound infection, surgical wounds, and surgical wound care. Based on this analysis, we determined that all of these conditions had either a small or statistically insignificant impact on PT costs per day and OT costs per day. As previously stated, because the current system is heavily based on service provision and most residents are classified into the Ultra-High therapy category, there is currently little variance available in PT and OT costs per day to be explained by the presence of comorbidities. For the foregoing reasons, we do not believe it is appropriate to include the conditions mentioned by commenters as comorbidities for PT or OT payment at this time. However, as care practices change over time, we may consider adding comorbidities that have a strong impact on PT or OT utilization.

Comment: Many commenters supported the proposed separation of the PT and OT components, as compared to the RCS-I model that combined these components into a single component. One commenter questioned if therapy would be covered for pain management and wound care treatments as these types of treatments are not explicitly covered under the clinical categories.

Response: We appreciate the support for the decision to separate the PT and OT components. With regard to the question of therapy coverage for certain conditions, we would note that neither the clinical categories, nor any other aspects of PDPM implementation, should be taken to change any coverage guidelines.

Accordingly, after considering the comments received, for the reasons discussed in the proposed rule and in the final rule, we are finalizing the proposed PT and OT components under the PDPM and our proposals relating to the methodology for classifying residents under the PT and OT components, effective October 1, 2019, with the modifications discussed in this section. As discussed above, in response to comments, rather than requiring providers to record the type of inpatient surgical procedure performed during the prior inpatient hospital stay by coding an ICD-10-PCS code in the second line of item I8000 as we proposed, we will instead require providers to select, as necessary, a surgical procedure category in a sub-item within Item J2000 which would identify the relevant surgical procedure that occurred during the patient's preceding hospital stay and which would augment the patient's PDPM clinical category. For purposes of calculating the function score, all missing values for section GG assessment items will receive zero points. Similarly, the function score will incorporate a new response “10. Not attempted due to environmental limitations” and we will assign it a point value of zero. Furthermore, consistent with a commenter's suggestion, we will adopt MDS item GG0170I1 (Walk 10 feet) as a substitute for retired item GG0170H1 (Does the resident walk), and we will use responses 07: “resident refused,” 09: “not applicable,” 10: “not attempted due to environmental limitations,” or 88: “not attempted due to medical condition or safety concerns” from MDS item GG0170I1 to identify residents who cannot walk.

c. Speech-Language Pathology Case-Mix Classification

As discussed above and in the proposed rule (83 FR 21049), many of the resident characteristics that we found to be predictive of increased PT and OT costs were predictive of lower SLP costs. We stated that as a result of this inverse relationship, using the same set of predictors to case-mix adjust all three therapy components would obscure important differences in variables predicting variation in costs across therapy disciplines and make any model that attempts to predict total therapy costs inherently less accurate. Therefore, we stated that we believe it is appropriate to have a separately adjusted case-mix SLP component that is specifically designed to predict relative differences in SLP costs. As discussed in the proposed rule and in the prior section of this final rule, costs derived from the charges on claims and CCRs on facility cost reports were used as the measure of resource use to develop an alternative payment model. Costs are reflective of therapy utilization as they are correlated to therapy minutes recorded for each therapy discipline.

Following the same methodology we used to identify predictors of PT and OT costs, we explained in the proposed rule that our project team conducted cost regressions with a host of variables from the MDS assessment, prior inpatient claims, and SNF claims that were identified as likely to be predictive of relative increases in SLP costs. The variables were selected with the goal of being as inclusive of the measures recorded on the MDS assessment as possible and also included diagnostic information from the prior inpatient stay. The selection process also incorporated clinical input from TEP panelists, the contractor's clinical staff, and CMS clinical staff. We stated that these initial costs regressions were exploratory and meant to identify a broad set of resident characteristics that are predictive of SLP resource utilization. The results were used to inform which variables should be investigated further and ultimately included in the payment system. A table of all of the variables considered in this analysis appears in the appendix of the SNF PMR technical report that accompanied the ANPRM (available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html).

As we stated in the proposed rule (83 FR 21049), based on these cost regressions, we identified a set of three categories of predictors relevant in predicting relative differences in SLP costs: Clinical reasons for the SNF stay, presence of a swallowing disorder or mechanically-altered diet, and the presence of an SLP-related comorbidity or cognitive impairment. We explained that a model using these predictors to predict SLP costs per day accounted for 14.5 percent of the variation in SLP costs per day, while a very extensive model using 1,016 resident characteristics only predicted 19.3 percent of the variation. We stated that this shows that these predictors alone explain a large share of the variation in SLP costs per day that can be explained with resident characteristics.

As with the proposed PT and OT components, we began with the set of clinical categories identified in Table 14 of the proposed rule (set forth in Table 14 of this final rule) meant to capture general differences in resident resource utilization and ran cost regressions to determine which categories may be predictive of generally higher relative SLP costs. Through this analysis, we found that one clinical category, the Acute Neurologic group, was particularly predictive of increased SLP costs. More detail on this investigation can be found in section 3.5.2. of the SNF PMR technical report that accompanied the ANPRM, available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html. Therefore, to determine the initial resident classification into an SLP group under the proposed PDPM, we stated that residents would first be categorized into one of two groups using the clinical reasons for the resident's SNF stay recorded on the first line of Item I8000 on the MDS assessment: Either the “Acute Neurologic” clinical category or a “Non-Neurologic” group that includes the remaining clinical categories in Table 14 (Major Joint Replacement or Spinal Surgery; Non-Surgical Orthopedic/Musculoskeletal; Orthopedic Surgery (Except Major Joint Replacement or Spinal Surgery); Acute Infections; Cancer; Pulmonary; Non-Orthopedic Surgery; Cardiovascular and Coagulations; and Medical Management).

In addition to the clinical reason for the SNF stay, based on cost regressions and feedback from TEP panelists, we stated in the proposed rule (83 FR 21050) that we also identified the presence of a swallowing disorder or a mechanically-altered diet (which refers to food that has been altered to make it easier for the resident to chew and swallow to address a specific resident need) as a predictor of relative increases in SLP costs. First, we stated that residents who exhibited the signs and symptoms of a swallowing disorder, as identified using K0100Z on the MDS 3.0, demonstrated significantly higher SLP costs than those who did not exhibit such signs and symptoms. Therefore, we considered including the presence of a swallowing disorder as a component in predicting SLP costs. However, when this information was presented during the October 2016 TEP, stakeholders indicated that the signs and symptoms of a swallowing disorder may not be as readily observed when a resident is on a mechanically-altered diet and requested that we also consider evaluating the presence of a mechanically-altered diet, as determined by item K0510C2 on the MDS 3.0, as an additional predictor of increased SLP costs. As we further explained in the proposed rule, our project team conducted this analysis and found that there was an associated increase in SLP costs when a mechanically-altered diet was present. Moreover, we stated that this analysis revealed that while SLP costs may increase when either a swallowing disorder or mechanically-altered diet is present, resident SLP costs increased even more when both of these items were present. More detail on this investigation and these analyses can be found in section 3.5.3. of the SNF PDPM technical report, available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html. As a result, we agreed with the stakeholders that both swallowing disorder and mechanically-altered diet are important components of predicting relative increases in resident SLP costs, and thus, in addition to the clinical categorization, we proposed classifying residents as having either a swallowing disorder, being on a mechanically altered diet, both, or neither for the purpose of classifying the resident under the SLP component. We also noted that we plan to monitor specifically for any increases in the use of mechanically altered diet among the SNF population that may suggest that beneficiaries are being prescribed such a diet based on facility financial considerations, rather than for clinical need.

As a final aspect of the proposed SLP component case-mix adjustment, we explored how SLP costs vary according to cognitive status and the presence of an SLP-related comorbidity. As we explained in the proposed rule, we observed that SLP costs were notably higher for residents who had a mild to severe cognitive impairment as defined by the PDPM cognitive measure methodology described in Table 20 of the proposed rule (set forth in Table 20 of this final rule) or who had an SLP-related comorbidity present. We stated that for each condition or service included as an SLP-related comorbidity, the presence of the condition or service was associated with at least a 43 percent increase in average SLP costs per day. The presence of a mild to severe cognitive impairment was associated with at least a 100 percent increase in average SLP costs per day. Similar to the analysis conducted in relation to the PT and OT components, the project team ran cost regressions on a broad list of possible conditions. As we stated in the proposed rule (83 FR 21050), based on that analysis, and in consultation with stakeholders during our TEPs and clinicians, we identified the conditions listed in Table 22 of the proposed rule (set forth in Table 22 of this final rule) as SLP-related comorbidities which we believe best predict relative differences in SLP costs. As discussed in the proposed rule, we used diagnosis codes on the most recent inpatient claim and the first SNF claim, as well as MDS items on the 5-day assessment for each SNF stay to identify these diagnoses and found that residents with these conditions had much higher SLP costs per day. Further, we stated that rather than accounting for each SLP-related comorbidity separately, all conditions were combined into a single flag. If the resident has at least one SLP-related comorbidity, the combined flag is turned on. We explained in the proposed rule that we combined all SLP-related comorbidities into a single flag because we found that the predictive ability of including a combined SLP comorbidity flag is comparable to the predictive ability of including each SLP-related comorbidity as an individual predictor. Additionally, we stated that using a combined SLP-related comorbidity flag greatly improves the simplicity of the payment model. More detail on these analyses can be found in section 3.5.1. of the SNF PMR technical report that accompanied the ANPRM (available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html).

Aphasia.

CVA, TIA, or Stroke.

Hemiplegia or Hemiparesis.

Traumatic Brain Injury.

Tracheostomy Care (While a Resident).

Ventilator or Respirator (While a Resident).

Laryngeal Cancer.

Apraxia.

Dysphagia.

ALS.

Oral Cancers.

Speech and Language Deficits.

Once each of these variables—clinical reasons for the SNF stay, presence of a swallowing disorder or mechanically-altered diet, and the presence of an SLP-related comorbidity or cognitive impairment—found to be useful in predicting resident SLP costs was identified, as we discussed in the proposed rule (83 FR 21050), we used the CART algorithm, as we discussed above in relation to the PT and OT components, to determine appropriate splits in SLP case-mix groups based on CART output breakpoints using these three variables. We stated we then further refined the SLP case-mix classification groups beyond those produced by the CART algorithm. We used consistent criteria to group residents into 18 payment groups across the two clinical categories determined to be relevant to SLP utilization (Acute Neurologic and Non-Neurologic). These groups simplified the SLP case-mix classification by reducing the number of groups while maintaining the CART predictive power in terms of R-squared. This methodology and the results of our analysis are more thoroughly explained in sections 3.4.2. and 3.5.2. of the SNF PMR technical report that accompanied the ANPRM (available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html).

Under the original RCS-I SLP component, a resident could be classified into one of 18 possible case-mix groups. Comments received in response to the ANPRM expressed concern over the complexity of the payment model due to the high number of possible combinations of case-mix groups. We stated in the proposed rule (83 FR 21051) that, to reduce the number of possible SLP case-mix groups, we simplified the consistent splits model selected for RCS-I. To accomplish this, we combined clinical category (Acute Neurologic or Non-Neurologic), cognitive impairment, and the presence of an SLP-related comorbidity into a single predictor due to the clinical relationship between acute neurologic conditions, cognition, and SLP comorbidities. We explained in the proposed rule that these three predictors are highly interrelated as acute neurologic conditions may often result in cognitive impairment or SLP-related comorbidities such as speech and language deficits. As we discussed in the proposed rule, using this combined variable along with presence of a swallowing disorder or mechanically-altered diet results in 12 groups. We compared the predictive ability of the simplified model with more complex classification options, including the original RCS-I SLP model. We explained that regression results showed that the reduction in case-mix groups by collapsing independent variables had little to no effect on payment accuracy. Specifically, we noted that the proposed PDPM SLP model has an R-squared value almost identical to that of the original RCS-I SLP model, while reducing the number of resident groups from 18 to 12. Therefore, we determined that 12 case-mix groups would be necessary to classify residents adequately in terms of their SLP costs in a manner that captures sufficient variation in SLP costs without creating unnecessarily granular separations. More information on this analysis can be found in section 3.5.2. of the SNF PDPM technical report (available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html). We provided the criteria for each of these groups along with its CMI in Table 23 of the proposed rule (set forth in Table 23 of this final rule).

To help ensure that payment reflects the average relative resource use at the per diem level, we stated in the proposed rule (83 FR 21051) that CMIs would be set to reflect relative case-mix related differences in costs across groups. We stated that this method helps ensure that the share of payment for each case-mix group would be equal to its share of total costs of the component. We further explained that CMIs for the SLP component were calculated based on the average per diem costs of a case-mix group relative to the population average. Relative average differences in costs were weighted by length of stay to account for the different length of stay distributions across case-mix groups (as further discussed in section 3.11.1. of the SNF PDPM technical report, available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html). In this calculation, average per diem costs equaled total SLP costs in the group divided by number of utilization days in the group. Because the SLP component does not have a variable per diem schedule (as further discussed in section 3.9.1. of the SNF PDPM technical report, available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html), variable per diem adjustment factors were not involved in SLP CMI calculation. We further stated that a parity adjustment was then applied by multiplying the CMI by the ratio of case-mix-related payments in RUG-IV over estimated case-mix-related payments in PDPM, as further discussed in section V.J. of the proposed rule. We stated that this method helps ensure that the share of payment for each case-mix group is equal to its share of total costs of the component and that PDPM is budget neutral relative to RUG-IV. The full methodology used to develop CMIs is presented in section 3.11. of the SNF PDPM technical report (available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html).

TABLE 23—SLP Case-Mix Classification Groups

Presence of acute neurologic condition, SLP-related comorbidity, or cognitive impairment	Mechanically altered diet or swallowing disorder	SLP case-mix group	SLP case-mix index
None	Neither	SA	0.68
None	Either	SB	1.82
None	Both	SC	2.66
Any one	Neither	SD	1.46
Any one	Either	SE	2.33
Any one	Both	SF	2.97
Any two	Neither	SG	2.04
Any two	Either	SH	2.85
Any two	Both	SI	3.51
All three	Neither	SJ	2.98
All three	Either	SK	3.69
All three	Both	SL	4.19

As with the PT and OT components, we stated that all residents would be classified into one and only one of these 12 SLP case-mix groups under the PDPM. We explained that, as opposed to the RUG-IV system that determines therapy payments based only on the amount of therapy provided, under the PDPM, residents would be classified into SLP case-mix groups based on resident characteristics shown to be predictive of SLP utilization. Thus, we stated that believe the SLP case-mix groups will provide a better measure of resource use and will provide for more appropriate payment under the SNF PPS.

We invited comments on the approach we proposed above to classify residents for SLP payment under the proposed PDPM.

Commenters submitted the following comments related to the proposed rule's discussion of the classification of residents for SLP payment under the PDPM. A discussion of these comments, along with our responses, appears below.

Comment: Some commenters agreed with the SLP-related patient classifiers. Some commenters suggested using a different assessment of cognition than that currently used on the MDS, such as the Montreal Cognitive Assessment (MOCA). One commenter expressed concerns regarding the use of cognition as a first tier classification criterion, as changes in cognition can be difficult to identify and this could impact on the possibility of late or missed IPAs. This commenter suggested moving cognition into the second tier classification criteria.

Response: We appreciate the support for the SLP component classification criteria. With regard to the comment on using a different assessment for assessing cognition, we are not opposed to this idea and would encourage stakeholders to work with CMS in developing potential revisions to the MDS to improve care planning and management. That being said, as the MOCA is not currently in use on the MDS, we must utilize the data and assessment tools to which we currently have access. Finally, with regard to the concern about the interplay between cognition and the IPA, we expect that this concern would be addressed by having the IPA be completed on an optional basis, as described in section V.D.1 of this final rule.

Comment: One commenter expressed concern that having a separate SLP component could result in the overutilization of SLP services, specifically for treating cognitive impairments. The commenter advised CMS to limit the overutilization of SLP services for cognitive impairment issues.

Response: As discussed above, we found that cognitive impairment is a relevant characteristic in predicting SLP resource utilization and costs. However, we understand the concern regarding the potential for providers to overutilize SLP services in certain instances and will monitor the use of SLP services under PDPM to identify any potential consequences of using this payment classifier as part of the SLP component.

Comment: A commenter questioned the accuracy of using the same primary diagnosis to assign clinical category across the PT, OT, and SLP components. This commenter states that multiple diagnoses can contribute to the reason for the SNF stay and proposes distinguishing between PT/OT and SLP diagnoses. Specifically, the commenter suggests allowing providers to enter the clinical reason for PT/OT services in the first two lines of MDS item I8000 and the clinical reason for SLP services in the third line of item I8000. This commenter points to our decision to separate therapy disciplines into different payment components based on our observation that different sets of resident characteristics were predictive of PT and OT costs, on one hand, and SLP costs, on the other. Given that utilization of PT and OT resources and utilization of SLP services are explained by a different set of predictors, this commenter concludes that the clinical reasons for receiving SLP services are distinct from those motivating PT/OT services.

Response: As detailed in the proposed rule (83 FR 21043) and section 3.4.1 of the SNF PDPM technical report, when constructing the ten clinical categories, we explored conditions that are clinically relevant to general SNF resource utilization. Within each component, we further consolidated the ten clinical categories into groups that have significant impact on component-specific resource utilization. We found that the clinical reason for a SNF stay as represented by the clinical categories was highly predictive of PT, OT, and SLP utilization, and thus we do not believe it is necessary to enter separate clinical reasons for PT/OT and SLP services, as suggested by the commenter. For this reason, we believe it is appropriate to include the clinical categories as determinants of resident classification and payment for all three components. We would also emphasize that clinical category is the only predictor shared by the PT/OT and SLP components. The other independent variables are unique to the PT and OT or SLP components and capture other clinical reasons for PT/OT and SLP services. As a result, in many cases, a resident's cognitive status and the presence of SLP-related comorbidities may be as relevant as primary diagnosis in determining resident classification and payment.

Comment: A few commenters stated that the proposed SLP-related comorbidity list is an incomplete reflection of all comorbidities that require SLP treatment. One commenter stated that the SLP comorbidity list should include progressive neurologic disorders that increase SLP resource use. This commenter suggests relabeling the “ALS” MDS checkbox item as “Progressive Neurologic Diseases” and updating the MDS manual definition for this item to meet the criteria of specific progressive neurologic diseases.

Response: We appreciate commenters' concerns regarding additional conditions that may be related to SLP utilization. We may consider adding conditions that have a demonstrated relationship to SLP resource use in future revisions to the payment model. To examine the impact of PDPM on residents with chronic neurological conditions, we included this subpopulation in our resident impact analysis and found that PDPM is estimated to slightly increase the payment associated with these residents.

Comment: Some commenters agreed with the use of mechanically altered diet as a payment classifier. One commenter requested that CMS provide evidence that a mechanically altered diet is associated with higher SLP utilization than other nutritional approaches such as personal assistance with feeding. One commenter requested that CMS monitor the use of mechanically altered diets under PDPM to identify any potentially inappropriate use of such diets. One commenter stated that overutilization of such diets can have negative repercussions for patient care.

Response: As described in section 3.5.1 and 3.5.2 of the SNF PMR technical report, besides mechanically altered diet, we additionally explored feeding tube as a determinant of classification and payment for the SLP component. We used CART to test several SLP models with different variables related to swallowing and nutritional approach. This investigation found that mechanically altered diet notably increased the predictive power of the models, whereas feeding tube only had a small impact on predictive ability. While feeding tube was associated with an increase in SLP costs per day, we did not include feeding tube in the payment model because it only had a small impact on the predictive accuracy of the model relative to mechanically altered diet. We also explored the MDS item Eating Self-Performance (G0110H1) as a potential predictor of SLP utilization. While increased eating dependence was associated with higher SLP utilization, when we included Eating Self-Performance as an independent variable in the CART analysis used to explore possible case-mix groups, Eating Self-Performance was only selected as a determinant of classification for half of the 18 groups created by the CART algorithm. As a result, we determined that we could remove Eating Self-Performance from the SLP classification without notably sacrificing predictive ability. As shown in section 3.5.2 of the SNF PMR technical report, removing Eating Self-Performance and combining various independent variables to simplify the classification reduced the R-squared value of the classification by only 0.005. As a result, this classification was used as the basis for the proposed PDPM SLP component.

With regard to the possibility of some providers prescribing mechanically altered diets inappropriately or the possibility of overutilization, we do plan to monitor the use of these diets as part of our general PDPM monitoring strategy.

Accordingly, after considering the comments received, for the reasons discussed in the proposed rule and in this final rule, we are finalizing, without modification, the proposed SLP component of PDPM and our proposals relating to the classification of residents under the SLP component.

d. Nursing Case-Mix Classification

As we explained in the proposed rule (83 FR 21051 through 21052), the RUG-IV classification system first divides residents into “rehabilitation residents” and “non-rehabilitation residents” based on the amount of therapy a resident receives. We stated that differences in nursing needs can be obscured for rehabilitation residents, where the primary driver of payment classification is the intensity of therapy services that a resident receives. For example, for two residents classified into the RUB RUG-IV category, which would occur on the basis of therapy intensity and ADL score alone, the nursing component for each of these residents would be multiplied by a CMI of 1.56. We stated that this reflects that residents in that group were found, during our previous Staff time measurement (STM) work, to have nursing costs 56 percent higher than residents with a 1.00 index. We noted that while this CMI also includes adjustments made in FY 2010 and FY 2012 for budget-neutrality purposes, what is clear is that two residents, who may have significantly different nursing needs, are nevertheless deemed to have the very same nursing costs, and SNFs would receive the same nursing payment for each. Given the discussion above and in the proposed rule, which noted that approximately 60 percent of resident days are billed using one of three Ultra-High Rehabilitation RUGs (two of which have the same nursing index), we stated that the current case-mix model effectively classifies a significant portion of SNF therapy residents as having exactly the same degree of nursing needs and requiring exactly the same amount of nursing resources. As such, we stated we believed that further refinement of the case-mix model would be appropriate to better differentiate among patients, particularly those who receive therapy services with different nursing needs.

We further explained in the proposed rule (83 FR 21052) that an additional concern in the RUG-IV system is the use of therapy minutes to determine not only therapy payments but also nursing payments. For example, residents classified into the RUB RUG fall in the same ADL score range as residents classified into the RVB RUG. The only difference between those residents is the number of therapy minutes that they received. However, as we stated in the proposed rule, the difference in payment that results from this difference in therapy minutes impacts not only the RUG-IV therapy component but also the nursing component: Nursing payments for RUB residents are 40 percent higher than nursing payments for RVB residents. We stated that as a result of this feature of the RUG-IV system, the amount of therapy minutes provided to a resident is one of the main sources of variation in nursing payments, while other resident characteristics that may better reflect nursing needs play a more limited role in determining payment.

As discussed in the proposed rule (83 FR 21052), the more nuanced and resident-centered classifications in current RUG-IV non-rehabilitation categories are obscured under the current payment model, which utilizes only a single RUG-IV category for payment purposes and has over 90 percent of resident days billed using a rehabilitation RUG. The RUG-IV non-rehabilitation groups classify residents based on their ADL score, the use of extensive services, the presence of specific clinical conditions such as depression, pneumonia, or septicemia, and the use of restorative nursing services, among other characteristics. These characteristics are associated with nursing utilization, and the STRIVE study accounted for relative differences in nursing staff time across groups. Therefore, we proposed to use the existing RUG-IV methodology for classifying residents into non-rehabilitation RUGs to develop a proposed nursing classification that helps ensure nursing payment reflects expected nursing utilization rather than therapy utilization.

For example, in the proposed rule (83 FR 21052), we considered two residents. The first patient classifies into the RUB rehabilitation RUG (on the basis of the resident's therapy minutes) and into the CC1 non-rehabilitation RUG (on the basis of having pneumonia), while the second classifies into the RUB rehabilitation RUG (on the basis of the resident's therapy minutes) and the HC1 non-rehabilitation RUG (on the basis of the resident having quadriplegia and a high ADL score). Under the current RUG-IV based payment model, the billing for both residents would utilize only the RUB rehabilitation RUG, despite clear differences in their associated nursing needs and resident characteristics. We proposed an approach where, for the purpose of determining payment under the nursing component, the first resident would be classified into CC1, while the second would be classified into HC1 under the PDPM. We stated that believe classifying the residents in this manner for payment purposes would capture variation in nursing costs in a more accurate and granular way than relying on the rehabilitation RUG's nursing CMI.

While resident classification in the proposed PDPM nursing component is guided by RUG-IV methodology, we proposed to make several modifications to the RUG-IV nursing RUGs and classification methodology under the proposed PDPM. First, we proposed under the PDPM to reduce the number of nursing RUGs by decreasing distinctions based on function. We stated that under RUG-IV, residents with a serious medical condition/service such as septicemia or respiratory therapy are classified into one of eight nursing RUGs in the Special Care High category. The specific RUG into which a resident is placed depends on the resident's ADL score and whether the resident is depressed. RUG-IV groups ADL score into bins for simplicity (for example, 2-5 and 6-10). For example, under RUG-IV, a resident in the Special Care High category who has depression and an ADL score of 3 would fall into the 2-5 ADL score bin, and therefore, be classified into the HB2 RUG, which corresponds to Special Care High residents with depression and an ADL score between 2 and 5 (a mapping of clinical traits and ADL score to RUG-IV nursing groups is shown in the appendix of the SNF PDPM technical report, available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html). To explore options to reduce the number of nursing RUGs, we explained in the proposed rule that we compared average nursing utilization across all 43 RUG-IV nursing RUGs. The dependent variable used in this investigation was the average wage-weighted staff time (WWST) for each nursing RUG from the STRIVE study. WWST is a measure of nursing resource utilization used in the STRIVE study. As discussed in more detail in the proposed rule (83 FR 21052) and in section 3.2.1. of the PDPM technical report (available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html), we were unable to construct a measure of nursing utilization based on current data because facilities do not report resident-specific nursing costs. As discussed in the proposed rule, we observed that nursing resource use as measured by WWST does not vary markedly between nursing case-mix groups defined by contiguous ADL score bins (for example, 11-14 and 15-16) but otherwise sharing the same clinical traits (for example, classified into Special Care High and depressed). We explained that this suggests that collapsing contiguous ADL score bins for RUGs that are otherwise defined by the same set of clinical traits is unlikely to notably affect payment accuracy. Section 3.6.1. of the SNF PDPM technical report (available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html) provides more detail on this analysis.

In the Special Care High, Special Care Low, Clinically Complex, and Reduced Physical Function classification groups (RUGs beginning with H, L, C, or P), for nursing groups that were otherwise defined with the same clinical traits (for example, extensive services, medical conditions, depression, restorative nursing services received), we proposed to combine the following pairs of second characters due to their contiguous ADL score bins: (E, D) and (C, B). These characters correspond to ADL score bins (15 to 16, 11 to 14) and (6 to 10, 2 to 5), respectively. We observed that nursing utilization did not vary notably across these contiguous ADL score bins; therefore, we stated that we believe it is appropriate to collapse pairs of RUGs in these classification groups that correspond to contiguous ADL score bins but are otherwise defined by the same clinical traits. For example, HE2 and HD2, which are both in the Special Care High group and both indicate the presence of depression, would be collapsed into a single nursing case-mix group. Similarly, we stated that PC1 and PB1 (Reduced Physical Function and 0 to 1 restorative nursing services) also would be combined into a single nursing case-mix group. Section 3.6.1. of the SNF PDPM technical report (available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html) provides more detail on this analysis. In the Behavioral and Cognitive Performance classification group (RUGs beginning with B), for RUGs that are otherwise defined by the same number of restorative nursing services (0 to1 or 2 or more), we proposed to combine RUGs with the second character B and A, which correspond to contiguous ADL score bins 2 to 5 and 0 to 1, respectively. We observed that nursing utilization did not vary notably across these contiguous ADL score bins; therefore, we stated that we believe it is appropriate to collapse pairs of RUGs in this classification group that correspond to contiguous ADL score bins but are otherwise defined by the same clinical traits. In other words, BB2 and BA2 would be combined into a single nursing group, and BB1 and BA1 would also be combined into a single nursing group. Section 3.6.1. of the SNF PDPM technical report (available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html) provides more detail on this analysis. We proposed to maintain CA1, CA2, PA1, and PA2 as separate case-mix groups under the nursing component of the PDPM. We observed that these RUGs do not share similar levels of nursing resource use with RUGs in adjacent ADL score bins that are otherwise defined by the same clinical traits (for example, medical conditions, depression, restorative nursing services received). Rather, we noted that CA1, CA2, PA1, and PA2 are associated with distinctly lower nursing utilization compared to RUGs that otherwise have the same clinical traits (for example, medical conditions, depression, restorative nursing services received) but higher ADL score bins. Section 3.6.1. of the SNF PDPM technical report (available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html) provides more detail on this analysis. We further stated that ES3, ES2, and ES1 also would be maintained as separate case-mix groups under the nursing component of the proposed PDPM because, although they are defined by the same ADL score bin, they are defined by different clinical traits unlike the pairs of RUGs that were combined. Specifically, ES3, ES2, and ES1 are defined by different combinations of extensive services. We stated that we believe collapsing case-mix groups based on ADL score for the RUGs specified above would reduce model complexity by decreasing the number of nursing case-mix groups from 43 to 25, which thereby decreases the total number of possible combinations of case-mix groups under the proposed PDPM. Table 26 of the proposed rule (set forth in Table 26 of this final rule) shows the proposed 25 case-mix groups for nursing payment. Section 3.6.1. of the SNF PDPM technical report (available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html) provides more detail on the analyses and data supporting these proposals.

As explained in the proposed rule (83 FR 21053), the second modification to the RUG-IV nursing classification methodology would update the nursing ADL score to incorporate section GG items. Currently, the RUG-IV ADL score is based on four late-loss items from section G of MDS 3.0: eating, toileting, transfer, and bed mobility. We stated that under the proposed PDPM, these section G items would be replaced with an eating item, a toileting item, three transfer items, and two bed mobility items from the admission performance assessment of section GG. In contrast to the RUG-IV ADL score, the proposed PDPM score assigns higher points to higher levels of independence. Therefore, an ADL score of 0 (independent) corresponds to a section GG-based function score of 16, while an ADL score of 16 (dependent) corresponds to a section GG-based function score of 0. We explained that this scoring methodology is consistent with the proposed PDPM PT and OT function score, as well as functional scores in other care settings, such as the IRF PPS. The proposed nursing scoring methodology also assigns 0 points when an activity cannot be completed (“Resident refused,” “Not applicable,” “Not attempted due to medical condition or safety concerns”). As described in section V.D.3.c. (PT and OT Case-Mix Classification) of the proposed rule, grouping these responses with “dependent” aligns with clinical expectations of resource utilization for residents who cannot complete an ADL activity. The proposed scoring methodology is shown in Table 24 of the proposed rule (set forth in Table 24 of this final rule). As discussed in section V.D.3.c. of the proposed rule, section GG measures functional areas with more than one item, which results in substantial overlap between the two bed mobility items and the three transfer items. To address overlap, we proposed to calculate an average score for each of these related items. That is, we stated we would average the scores for the two bed mobility items and for the three transfer items. This averaging approach was also used in the proposed PT and OT function scores and is illustrated in Table 25 of the proposed rule (set forth in Table 25 of this final rule). We stated that the final score sums the average bed mobility and transfer scores with eating and toileting scores, resulting in a nursing function score that ranges from 0 to 16.

Table 24—Nursing Function Score Construction

Response	ADL score
05, 06—Set-up assistance, Independent	4
04—Supervision or touching assistance	3
03—Partial/moderate assistance	2
02—Substantial/maximal assistance	1
01, 07, 09, 88—Dependent, Refused, N/A, Not Attempted	0

Table 25—Section GG Items Included in Nursing Functional Measure

Section GG item	ADL score
GG0130A1—Self-care: Eating	0-4.
GG0130C1—Self-care: Toileting Hygiene	0-4.
GG0170B1—Mobility: Sit to lying	0-4 (average of 2 items).
GG0170C1—Mobility: Lying to sitting on side of bed
GG0170D1—Mobility: Sit to stand	0-4 (average of 3 items).
GG0170E1—Mobility: Chair/bed-to-chair transfer
GG0170F1—Mobility: Toilet transfer

In addition to proposing to replace the nursing ADL score with a function score based on section GG items and to collapse certain nursing RUGs, we also proposed (83 FR 21054) to update the existing nursing CMIs using the STRIVE staff time measurement data that were originally used to create these indexes. We explained that under the current payment system, non-rehabilitation nursing indexes were calculated to capture variation in nursing utilization by using only the staff time collected for the non-rehabilitation population. We stated we believe that, to provide a more accurate reflection of the relative nursing resource needs of the SNF population, the nursing indexes should reflect nursing utilization for all residents. To accomplish this, we stated in the proposed rule that we replicated the methodology described in the FY 2010 SNF PPS rule (74 FR 22236 through 22238) but classified the full STRIVE study population under non-rehabilitation RUGs using the RUG-IV classification rules. The methodology set forth in the proposed rule for updating resource use estimates for each nursing RUG proceeded according to the following steps:

(1) Calculate average wage-weighted staff time (WWST) for each STRIVE study resident using FY 2015 SNF wages.

(2) Assign the full STRIVE population to the appropriate non-rehabilitation RUG.

(3) Apply sample weights to WWST estimates to allow for unbiased population estimates. The reason for this weighting is that the STRIVE study was not a random sample of residents. Certain key subpopulations, such as residents with HIV/AIDS, were over-sampled to ensure that there were enough residents to draw conclusions on the subpopulations' resource use. As a result, STRIVE researchers also developed sample weights, equal to the inverse of each resident's probability of selection, to permit calculation of unbiased population estimates. Applying the sample weights to a summary statistic results in an estimate that is representative of the actual population. The sample weight method is explained in Phase I of the STRIVE study. A link to the STRIVE study is available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​TimeStudy.html.

(4) Smooth WWST estimates that do not match RUG hierarchy in the same manner as the STRIVE study. RUG-IV, from which the nursing RUGs are derived, is a hierarchical classification in which payment should track clinical acuity. It is intended that residents who are more clinically complex or who have other indicators of acuity, including a higher ADL score, depression, or restorative nursing services, would receive higher payment. When STRIVE researchers estimated WWST for each RUG, several inversions occurred because of imprecision in the means. These are defined as WWST estimates that are not in line with clinical expectations. The methodology used to smooth WWST estimates is explained in Phase II of the STRIVE study. A link to the STRIVE study is available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​TimeStudy.html.

(5) Calculate nursing indexes, which reflect the average WWST for each of the 25 nursing case-mix groups divided by the average WWST for the study population used throughout our research. To impute WWST for each stay in the population, we assigned each resident the average WWST of the collapsed nursing RUG into which they are categorized. To derive the average WWST of each collapsed RUG, we first estimate the average WWST of the original 43 nursing RUGs based on steps 1 through 4 above, then calculate a weighted mean of the average WWST of the two RUGs that form the collapsed RUG. More details on this analysis can be found in section 3.6.3. of the SNF PDPM technical report (available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html).

Through this refinement, we stated that we believe the nursing indexes under the proposed PDPM better reflect the varied nursing resource needs of the full SNF population. In Table 26 of the proposed rule (set forth in Table 26 of this final rule), we provided the nursing indexes under the proposed PDPM.

To help ensure that payment reflects the average relative resource use at the per diem level, we stated that the nursing CMIs would be set to reflect case-mix related relative differences in WWST across groups. We further stated that Nursing CMIs would be calculated based on the average per diem nursing WWST of a case-mix group relative to the population average. In this calculation, average per diem WWST equaled total WWST in the group divided by number of utilization days in the group. We further explained that because the nursing component does not have a variable per diem schedule (as further discussed in section 3.9.1. of the SNF PDPM technical report, available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html), variable per diem adjustment factors were not involved in nursing CMI calculation. We then applied a parity adjustment by multiplying the CMI by the ratio of case-mix-related payments in RUG-IV over estimated case-mix-related payments in PDPM, as discussed further in section V.J. of the proposed rule. The full methodology used to develop CMIs is presented in section 3.11. of the SNF PDPM technical report (available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html).

Table 26—Nursing Indexes Under PDPM Classification Model

RUG-IV nursing RUG	Extensive services	Clinical conditions	Depression	Number of restorative nursing services	GG-based function score	PDPM nursing case-mix group	Nursing case-mix index
ES3	Tracheostomy & Ventilator				0-14	ES3	4.04
ES2	Tracheostomy or Ventilator				0-14	ES2	3.06
ES1	Infection				0-14	ES1	2.91
HE2/HD2		Serious medical conditions e.g. comatose, septicemia, respiratory therapy	Yes		0-5	HDE2	2.39
HE1/HD1		Serious medical conditions e.g. comatose, septicemia, respiratory therapy	No		0-5	HDE1	1.99
HC2/HB2		Serious medical conditions e.g. comatose, septicemia, respiratory therapy	Yes		6-14	HBC2	2.23
HC1/HB1		Serious medical conditions e.g. comatose, septicemia, respiratory therapy	No		6-14	HBC1	1.85
LE2/LD2		Serious medical conditions e.g. radiation therapy or dialysis	Yes		0-5	LDE2	2.07
LE1/LD1		Serious medical conditions e.g. radiation therapy or dialysis	No		0-5	LDE1	1.72
LC2/LB2		Serious medical conditions e.g. radiation therapy or dialysis	Yes		6-14	LBC2	1.71
LC1/LB1		Serious medical conditions e.g. radiation therapy or dialysis	No		6-14	LBC1	1.43
CE2/CD2		Conditions requiring complex medical care e.g. pneumonia, surgical wounds, burns	Yes		0-5	CDE2	1.86
CE1/CD1		Conditions requiring complex medical care e.g. pneumonia, surgical wounds, burns	No		0-5	CDE1	1.62
CC2/CB2		Conditions requiring complex medical care e.g. pneumonia, surgical wounds, burns	Yes		6-14	CBC2	1.54
CA2		Conditions requiring complex medical care e.g. pneumonia, surgical wounds, burns	Yes		15-16	CA2	1.08
CC1/CB1		Conditions requiring complex medical care e.g. pneumonia, surgical wounds, burns	No		6-14	CBC1	1.34
CA1		Conditions requiring complex medical care e.g. pneumonia, surgical wounds, burns	No		15-16	CA1	0.94
BB2/BA2		Behavioral or cognitive symptoms		2 or more	11-16	BAB2	1.04
BB1/BA1		Behavioral or cognitive symptoms		0-1	11-16	BAB1	0.99
PE2/PD2		Assistance with daily living and general supervision		2 or more	0-5	PDE2	1.57
PE1/PD1		Assistance with daily living and general supervision		0-1	0-5	PDE1	1.47
PC2/PB2		Assistance with daily living and general supervision		2 or more	6-14	PBC2	1.21
PA2		Assistance with daily living and general supervision		2 or more	15-16	PA2	0.70
PC1/PB1		Assistance with daily living and general supervision		0-1	6-14	PBC1	1.13
PA1		Assistance with daily living and general supervision		0-1	15-16	PA1	0.66

As with the previously discussed components, we stated that all residents would be classified into one and only one of these 25 nursing case-mix groups under the proposed PDPM. As explained in the proposed rule (83 FR 21055), we also used the STRIVE data to quantify the effects of an HIV/AIDS diagnosis on nursing resource use. We controlled for case mix by including the proposed PDPM resident groups (in this case, the nursing RUGs) as independent variables. The results showed that even after controlling for nursing RUG, HIV/AIDS status is associated with a positive and significant increase in nursing utilization. Based on the results of regression analyses, we found that wage-weighted nursing staff time is 18 percent higher for residents with HIV/AIDS. (The estimate of average wage-weighted nursing staff time for the SNF population was adjusted to account for the deliberate over-sampling of certain sub-populations in the STRIVE study. Specifically, we applied sample weights from the STRIVE dataset equal to the inverse of each resident's probability of selection to permit calculation of an unbiased estimate.) Based on these findings, as discussed in the proposed rule, we concluded that the proposed PDPM nursing groups may not fully capture the additional nursing costs associated with HIV/AIDS residents. More information on this analysis can be found in section 3.8.2. of the SNF PDPM technical report (available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html). Thus, as part of the case-mix adjustment of the nursing component, we proposed an 18 percent increase in payment for the nursing component for residents with HIV/AIDS. We stated that this adjustment would be applied based on the presence of ICD-10-CM code B20 on the SNF claim. In cases where a resident is coded as having this diagnosis, we stated that the nursing component per diem rate for this resident would be multiplied by 1.18, to account for the 18 percent increase in nursing costs for residents with this diagnosis. We also discussed this proposal, as well as its relation to the existing AIDS add-on payment under RUG-IV, in section V.I. of the proposed rule.

We invited comments on the approach we proposed to classify residents for nursing payment under the proposed PDPM.

Commenters submitted the following comments related to the proposed rule's discussion of the classification of residents for nursing payment under the PDPM. A discussion of these comments, along with our responses, appears below.

Comment: One commenter supported the nursing case-mix classification model that would be used under PDPM, specifically citing the nursing function score refinements and using a separate function score for the therapy components than would be used for the nursing component. This commenter also requested that CMS consider aligning the nursing classification under PDPM with certain hospice criteria. Finally, one commenter expressed concern regarding the collapsing of nursing groups to only 25 groups and that these groups may not accurately account for the variety of patient conditions.

Response: We appreciate the support for the nursing component classification criteria. We can examine the hospice criteria specified by the commenter for future consideration. With regard to the comment on the collapsed nursing groups, we believe that, given that we collapsed groups primarily based on functional score bins and did not collapse any of the general nursing group categories (such as extensive services and special care high), we believe that the level of granularity in the nursing classifications is not significantly impacted. As stated in the proposed rule (83 FR 21052) and in section 3.6.1 of the SNF PDPM technical report, we collapsed groups based on function score due to the observation that among nursing RUGs defined by the same clinical traits, nursing resource use is similar across contiguous functional score bins (for example, 11-14 and 15-16). Since WWST does not vary markedly between nursing RUGs defined by contiguous functional score bins, collapsing groups based on functional score bins simplifies the payment model without a notable loss in accuracy. Therefore, we believe that 25 nursing rugs sufficiently captures variation in patient conditions.

Comment: Several commenters questioned the appropriateness of using staff-time measurement data from the STRIVE study to estimate relative differences in nursing utilization across the nursing groups given the age of the data, methodological flaws in the collection of therapy minutes, and small sample sizes for certain resident groups used to estimate CMIs. Additionally, one commenter stated that the STRIVE study underestimates the nursing needs of residents by only measuring the usual nursing time provided to residents in the sampled homes. The commenter further stated that the STRIVE study did not take into account nursing time needed to assure resident safety and maintain resident well-being. The commenter expressed concern that basing nursing payment on STRIVE data will provide inadequate reimbursement, which will result in understaffing. A couple of commenters recommended replacing STRIVE with the Schnelle et al. 2016 simulation model to estimate nursing resource requirements.

Response: Unlike the therapy and NTA charges, nursing charges are reported on SNF claims as part of routine revenue centers, which also include non-case mix services such as room and board, rather than revenue centers specific to nursing. Due to the lack of resident-specific nursing charges, we used WWST from STRIVE data as a measure of nursing resource use in limited instances. Specifically, STRIVE data was not used to select determinants of payment for the nursing component. We only used STRIVE data to update case-mix indexes for the nursing component, so that nursing CMIs were calculated based on the entire SNF population rather than only on non-rehabilitation residents. We conducted a series of investigations into possible changes in resident characteristics from the time of the STRIVE study (2006) to fiscal year 2014 to determine if resident characteristics had changed in a manner that would suggest it would not be appropriate to use data from the STRIVE study in designing payment alternatives. The resident characteristics investigated include, but not limited to, most common Major Diagnostic Categories (MDC), percent of stays with complications or comorbidities in the qualifying inpatient stay, and frequency of MDS section I active diagnoses. The result of the investigations suggest that although there are small changes in prior inpatient hospital stay and SNF stay lengths, there have not been notable changes in resident characteristics or acuity over time. Given the stability of resident characteristics over time, there is no strong evidence of change in the relative resource utilization pattern among nursing groups since the time of STRIVE study in 2006.

In response to the concern about the methodology in collecting therapy minutes, we note that we only used nursing time to estimate CMIs for the nursing component under PDPM. Because therapy minutes were not included in the nursing staff time measure, concerns about how the STRIVE study collected therapy utilization data are not relevant to our use of STRIVE data to estimate nursing CMIs under PDPM.

As for the comments on the small sample sizes of certain resident groups in the STRIVE study, as detailed in section 4.1.2 of the STRIVE Phase II Report available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​TimeStudy.html, the STRIVE study used several procedures to address these concerns. First, STRIVE researchers deliberately over-sampled certain vulnerable resident groups to obtain more robust estimates of resource utilization for these subpopulations. Second, the STRIVE authors applied sample weights to obtain reliable population estimates. Because the proportion of facilities included in the study varied from state to state, the study population was not truly random. To account for this, the study developed sampling weights equal to the inverse of a resident's probability of selection for inclusion in the study population. The use of sampling weights allows the calculation of unbiased population estimates from the sample data, as described in section 4.1.2 of the STRIVE Phase II Report.

With regard to the comment stating concerns about how the STRIVE study measured nursing time, it is unclear what the commenter means by “usual nursing time” and “nursing time needed to assure resident safety and maintain resident well-being.” As discussed in the STRIVE Phase I and Phase II reports available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​TimeStudy.html, the STRIVE study collected three kinds of staff time: Resident Specific Time (RST), Non-Resident Specific Time (NRST), and Non-Study Time (NST). It was not appropriate to include NST in the dependent variable used to measure nursing utilization because these minutes did not benefit residents in the study population. As for NRST, while these minutes did benefit the study population, there are numerous methodological issues involved in including these minutes in the dependent variable. As noted in the STRIVE Phase II Report available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​TimeStudy.html, for many types of NRST, it is not clear how to allocate these non-resident specific minutes to specific residents. The STRIVE authors note that during development of RUG-III, NRST was allocated to individual residents in proportion to a resident's RST, based on the assumption that a resident's utilization of NRST was proportional to their utilization of RST. However, as the STRIVE authors note, this assumption may not be accurate. Accurate allocation of NRST would have involved extensive additional data collection that was beyond the scope of the STRIVE study. Without confidence in the allocation methodology, including NRST in the dependent variable for nursing would have introduced substantial noise into the dependent variable that could obscure the relationships between resident characteristics and resource utilization. As a result, the STRIVE authors decided to set relative payment weights based on RST alone. However, we disagree with the commenter if they are suggesting that excluding NRST leads us to underestimate nursing utilization. As noted in the STRIVE Phase II Report, the STRIVE study was only used to allocate nursing resources based on estimated relative resource utilization. It did not determine aggregate nursing resources, which are largely determined based on the methodology for setting and updating the federal per diem rates as specified in the Act. Therefore, it is incorrect to assert that relying on the STRIVE data for case-mix adjustment leads to inadequate nursing reimbursement since STRIVE is used to determine allocation of nursing resources rather than total nursing resources.

In response to the alternative data source proposed by commenters, the Schnelle et al. simulation model estimates resource use for nurse aides only; therefore, it is not a comprehensive or appropriate measure of nursing utilization.

Accordingly, after considering the comments received, for the reasons discussed in the proposed rule and in this final rule, we are finalizing our proposals, without modification, relating to the methodology, as described in this section, for classifying patients under the nursing component of PDPM.

e. Non-Therapy Ancillary Case-Mix Classification

Under the current SNF PPS, payments for NTA costs incurred by SNFs are incorporated into the nursing component. This means that the CMIs used to adjust the nursing component of the SNF PPS are intended to reflect not only differences in nursing resource use but also NTA costs. However, as we explained in the proposed rule (83 FR 21055), there have been concerns that the current nursing CMIs do not accurately reflect the basis for or the magnitude of relative differences in resident NTA costs. In its March 2016 Report to Congress, MedPAC wrote: “Almost since its inception, the SNF PPS has been criticized for encouraging the provision of unnecessary rehabilitation therapy services and not accurately targeting payments for nontherapy ancillary (NTA) services such as drugs (Government Accountability Office 2002, Government Accountability Office 1999, White et al. 2002)” (available at http://medpac.gov/​docs/​default-source/​reports/​chapter-7-skilled-nursing-facility-services-march-2016-report-.pdf). In the proposed rule, we stated that while the proposed PT, OT, and SLP components were designed to address the issue related to provision of therapy services raised by MedPAC above, the proposed NTA component was designed to address the issue related to accurately targeting payments for NTA services—specifically, that the current manner of using the RUG-IV case-mix system to determine NTA payment levels inadequately adjusts for relative differences in resident NTA costs.

As noted in the quotation from MedPAC above, MedPAC is not the only group to offer this critique of the SNF PPS. We stated in the proposed rule that just as the aforementioned criticisms that MedPAC cited have existed almost since the inception of the SNF PPS itself, ideas for addressing this concern have a similarly long history. In response to comments on the 1998 interim final rule which served to establish the SNF PPS, we published a final rule on July 30, 1999 (64 FR 41644). In this 1998 interim final rule, we acknowledged the commenters' concerns about the new system's ability to account accurately for NTA costs, stating that there were a number of comments expressing concern with the adequacy of the PPS rates to cover the costs of ancillary services other than occupational, physical, and speech therapy (non-therapy ancillaries), including such things as drugs, laboratory services, respiratory therapy, and medical supplies. We stated in the 1998 interim final rule that prescription drugs or medication therapy were frequently noted areas of concern due to their potentially high cost for particular residents. Some commenters suggested that the RUG-III case-mix classification methodology did not adequately provide for payments that account for the variation in, or the real costs of, these services provided to their residents. (64 FR 41647)

In response to those comments, we stated in the 1998 interim final rule that “we are funding substantial research to examine the potential for refinements to the case-mix methodology, including an examination of medication therapy, medically complex patients, and other nontherapy ancillary services” (64 FR 41648). In the FY 2019 SNF PPS proposed rule (83 FR 21055 through 21056), we proposed a methodology that we believe would case-mix adjust SNF PPS payments more appropriately to reflect differences in NTA costs.

Following the same methodology we used for the proposed PT, OT, and SLP components, the project team ran cost regression models to determine which resident characteristics may be predictive of relative increases in NTA costs. As explained in the proposed rule, the three categories of cost-related resident characteristics identified through this analysis were resident comorbidities, the use of extensive services (services provided to residents that are particularly expensive and/or invasive), and resident age. However, as discussed in the proposed rule, we removed age from further consideration as part of the NTA component based on concerns shared by TEP panelists during the June 2016 TEP. Particularly, some panelists expressed concern that including age as a determinant of NTA payment could create access issues for older populations. Additionally, we state that the CART algorithm used to explore potential resident groups for the NTA component only selected age as a determinant of classification for 2 of the 7 groups created. We noted that we also tested a classification option that used age as a determinant of classification for every NTA group. This only led to a 5 percent increase in the R-squared value of the NTA classification. More information on these analyses can be found in section 3.7.1. of the SNF PMR technical report available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html.

As we explained in the proposed rule (83 FR 21056), with regard to capturing comorbidities and extensive services associated with high NTA utilization, we used multiple years of data (FY 2014 to FY 2017) to estimate the impact of comorbidities and extensive services on NTA costs. This was in response to comments on the ANPRM that the design of the NTA component should be more robust and remain applicable in light of potential changes in the SNF population and care practices over time. We explained in the proposed rule that conditions and services were defined in three ways. First, clinicians identified MDS items that correspond to conditions/extensive services likely related to NTA utilization. However, we stated that since many conditions/extensive services related to NTA utilization are not included on the MDS assessment, we then mapped ICD-10 diagnosis codes from the prior inpatient claim, the first SNF claim, and section I8000 of the 5-day MDS assessment to condition categories from the Part C risk adjustment model (CCs) and the Part D risk adjustment model (RxCCs). The CCs and RxCCs define conditions by aggregating related diagnosis codes into a single condition flag. We use the condition flags defined by the CCs and RxCCs to predict Part A and B expenditures or Part D expenditures, respectively for Medicare beneficiaries. The predicted relationship between the conditions defined in the respective models and Medicare expenditures is then used to risk-adjust capitated payments to Part C and Part D sponsors. Similarly, we explained that our comorbidities investigation aimed to use a comprehensive list of conditions and services to predict resource utilization for beneficiaries in Part A-covered SNF stays. As we stated in the proposed rule, ultimately, the predicted relationship between these conditions/services and utilization of NTA services would be used to case-mix adjust payments to SNF providers, in a process similar to risk adjustment of capitated payments. Given these similarities, we decided to use the diagnosis-defined conditions from the Part C and Part D risk adjustment models to define conditions and services that were not defined on the MDS. Because the CCs were developed to predict utilization of Part A and B services, while the RxCCs were developed to predict Part D drug costs, the largest component of NTA costs, we stated that believe using both sources allows us to define the conditions and services potentially associated with NTA utilization more comprehensively. Lastly, we used ICD-10 diagnosis codes to define additional conditions that clinicians who advised CMS during PDPM development identified as being potentially associated with increased NTA service utilization but are not fully reflected in either the MDS or the CCs/RxCCs. The resulting list was meant to encompass as many diverse and expensive conditions and extensive services as possible from the MDS assessment, the CCs, the RxCCs, and diagnoses. As discussed in the proposed rule, using cost regressions, we found that certain comorbidity conditions and extensive services were highly predictive of relative differences in resident NTA costs. These conditions and services were identified in Table 27 of the proposed rule (set forth in Table 27 in this final rule). More information on this analysis can be found in section 3.7.1. of the SNF PDPM technical report available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html. We noted in the proposed rule that certain conditions that were associated with higher NTA utilization were nevertheless excluded from the list because of clinical concerns. Esophageal reflux was excluded because it is a very common condition in the SNF population and clinicians noted that coding can be discretionary. Migraine headache was also excluded due to clinicians' concerns about coding reliability. Additionally, we noted that clinicians stated that in many cases migraine headache is not treated by medication, the largest component of NTA costs.

Having identified the list of relevant conditions and services for adjusting NTA payments, in the proposed rule (83 FR 21056 through 21057), we considered different options for how to capture the variation in NTA costs explained by these identified conditions and services. We stated that one such method would be merely to count the number of comorbidities and services a resident receives and assign a score to that resident based on this count. We found that this option accounts for the additive effect of having multiple comorbidities and extensive services but did not adequately reflect the relative differences in the impact of certain higher-cost conditions and services. We also considered a tier system similar to the one used in the IRF PPS, where SNF residents would be placed into payment tiers based on the costliest comorbidity or extensive service. However, we found that this option did not account for the additive effect noted above. To address both of these issues, we proposed basing a resident's NTA score, which would be used to classify the resident into an NTA case-mix classification group, on a weighted-count methodology. Specifically, as shown in Table 27, each of the comorbidities and services that factored into a resident's NTA classification was assigned a certain number of points based on its relative impact on a resident's NTA costs. Those conditions and services with a greater impact on NTA costs were assigned more points, while those with less of an impact were assigned fewer points. The relative impacts are estimated based the coefficients of an ordinary least squares (OLS) regression that used the selected conditions and extensive services to predict NTA costs per day. Points were assigned by grouping together conditions and extensive services with similar OLS regression estimates. More information on this methodology and analysis can be found in section 3.7.1. of the SNF PDPM technical report available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html. We stated that the effect of this methodology was that the NTA component would adequately reflect relative differences in the NTA costs for each condition or service, as well as the additive effect of having multiple comorbidities.

We stated in the proposed rule (83 FR 21057) that a resident's total comorbidity score, which would be the sum of the points associated with all of a resident's comorbidities and services, would be used to classify the resident into an NTA case-mix group. For conditions and services where the source is indicated as MDS item I8000, SNF PDPM NTA Comorbidity Mapping (which accompanied the FY 2019 SNF PPS proposed rule) (available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html) provides a crosswalk between the listed condition and the ICD-10-CM codes which may be coded to qualify that condition to serve as part of the resident's NTA classification. MDS item I8000 is an open-ended item in the MDS assessment where the assessment provider can fill in additional active diagnoses that are not explicitly on the MDS for the resident in the form of ICD-10 codes. In the case of Parenteral/IV Feeding, we stated that we observed that NTA costs per day increase as the amount of intake through parenteral or tube feeding increases. For this reason, we proposed to separate this item into a high intensity item and a low intensity item, similar to how it is defined in the RUG-IV system. In order for a resident to qualify for the high intensity category, the percent of calories taken in by the resident by parenteral or tube feeding, as reported in item K0710A2 on the MDS 3.0, must be greater than 50 percent. We further stated that in order to qualify for the low intensity category, the percent of calories taken in by the resident by parenteral or tube feeding, as reported in item K0710A2 on the MDS 3.0, must be greater than 25 percent but less than or equal to 50 percent, and the resident must receive an average fluid intake by IV or tube feeding of at least 501cc per day, as reported in item K0710B2 of the MDS 3.0.

We also noted that the source of the HIV/AIDS diagnosis is listed as the SNF claim. We explained in the proposed rule that this is because 16 states have state laws that prevent the reporting of HIV/AIDS diagnosis information to CMS through the current assessment system and/or prevent CMS from seeing such diagnosis information within that system, should that information be mistakenly reported. We noted that the states are Alabama, Alaska, California, Colorado, Connecticut, Idaho, Illinois, Massachusetts, Nevada, New Hampshire, New Jersey, New Mexico, South Carolina, Texas, Washington, and West Virginia. Given this restriction, it would not be possible to have SNFs utilize the MDS 3.0 as the vehicle to report HIV/AIDS diagnosis information for purposes of determining a resident's NTA classification. We noted that the current SNF PPS uses a claims reporting mechanism as the basis for the temporary AIDS add-on payment which exists under RUG-IV. To address the issue discussed above with respect to reporting of HIV/AIDS diagnosis information under the proposed PDPM, we proposed to utilize this existing claims reporting mechanism to determine a resident's HIV/AIDS status for the purpose of NTA classification. More specifically, we explained that HIV/AIDS diagnosis information reported on the MDS would be ignored by the GROUPER software used to classify a resident into an NTA case-mix group. Instead, we stated that providers would be instructed to locate the HIPPS code provided to the SNF on the validation report associated with that assessment and report it to CMS on the associated SNF claim. Following current protocol, the provider would then enter ICD-10-CM code B20 on the associated SNF claim as if it were being coded to receive payment through the current AIDS add-on payment. The PRICER software, which we use to determine the appropriate per diem payment for a provider based on their wage index and other factors, would make the adjustment to the resident's NTA case-mix group based on the presence of the B20 code on the claim, as well as adjust the associated per diem payment based on the adjusted resident HIPPS code. Again, we noted that this methodology follows the same logic that the SNF PPS currently uses to pay the temporary AIDS add-on adjustment but merely changes the target and type of adjustment from the SNF PPS per diem to the NTA component of the proposed PDPM. We explained that the difference is that while under the current system, the presence of the B20 code would lead to a 128 percent increase in the per diem rate, under the proposed PDPM, the presence of the B20 code would mean the addition of 8 points (as determined by the OLS regression described above) to the resident's NTA score, the categorization of the resident into the appropriate NTA group, and an adjustment to the nursing component, as described in section V.D.3.d. of the proposed rule. Section 1888(e)(12) of the Act enacted a temporary 128 percent increase in the PPS per diem payment for SNF residents with HIV/AIDS and stipulated that the temporary adjustment was to be applied only until the Secretary certifies that there is an appropriate case-mix adjustment to compensate for the increased costs associated with this population. As we explained in the proposed rule, based on this language, we conducted an analysis similar to that used to determine the HIV/AIDS add-on for the nursing component to examine the adequacy of payment for ancillary services (all non-nursing services: PT, OT, SLP, and NTA) for residents with HIV/AIDS under the proposed PDPM. This analysis determined that after accounting for the 8 points assigned for HIV/AIDS in the NTA component and controlling for case-mix classification across the three therapy components and NTA component, HIV/AIDS was not associated with an increase in ancillary costs. We noted that nursing costs were not included in this regression because we separately investigated the increased nursing utilization associated with HIV/AIDS, as described in section V.D.3.d. of the proposed rule. Based on the results of this investigation, we concluded that the four ancillary case-mix components (PT, OT, SLP, and NTA) adequately reimburse costs associated with residents with HIV/AIDS. Therefore, we stated that we do not believe an HIV/AIDS add-on is warranted for the ancillary cost components. More information on this analysis can be found in section 3.8.2. of the PDPM technical report available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html.

Table 27 provides the proposed list of conditions and extensive services that would be used for NTA classification, the source of that information, and the associated number of points for that condition.

Table 27—Conditions and Extensive Services Used for NTA Classification

Condition/extensive service	Source	Points
HIV/AIDS	SNF Claim	8
Parenteral IV Feeding: Level High	MDS Item K0510A2, K0710A2	7
Special Treatments/Programs: Intravenous Medication Post-admit Code	MDS Item O0100H2	5
Special Treatments/Programs: Ventilator or Respirator Post-admit Code	MDS Item O0100F2	4
Parenteral IV feeding: Level Low	MDS Item K0510A2, K0710A2, K0710B2	3
Lung Transplant Status	MDS Item I8000	3
Special Treatments/Programs: Transfusion Post-admit Code	MDS Item O0100I2	2
Major Organ Transplant Status, Except Lung	MDS Item I8000	2
Active Diagnoses: Multiple Sclerosis Code	MDS Item I5200	2
Opportunistic Infections	MDS Item I8000	2
Active Diagnoses: Asthma COPD Chronic Lung Disease Code	MDS Item I6200	2
Bone/Joint/Muscle Infections/Necrosis—Except Aseptic Necrosis of Bone	MDS Item I8000	2
Chronic Myeloid Leukemia	MDS Item I8000	2
Wound Infection Code	MDS Item I2500	2
Active Diagnoses: Diabetes Mellitus (DM) Code	MDS Item I2900	2
Endocarditis	MDS Item I8000	1
Immune Disorders	MDS Item I8000	1
End-Stage Liver Disease	MDS Item I8000	1
Other Foot Skin Problems: Diabetic Foot Ulcer Code	MDS Item M1040B	1
Narcolepsy and Cataplexy	MDS Item I8000	1
Cystic Fibrosis	MDS Item I8000	1
Special Treatments/Programs: Tracheostomy Care Post-admit Code	MDS Item O0100E2	1
Active Diagnoses: Multi-Drug Resistant Organism (MDRO) Code	MDS Item I1700	1
Special Treatments/Programs: Isolation Post-admit Code	MDS Item O0100M2	1
Specified Hereditary Metabolic/Immune Disorders	MDS Item I8000	1
Morbid Obesity	MDS Item I8000	1
Special Treatments/Programs: Radiation Post-admit Code	MDS Item O0100B2	1
Highest Stage of Unhealed Pressure Ulcer—Stage 4	MDS Item M0300X1	1
Psoriatic Arthropathy and Systemic Sclerosis	MDS Item I8000	1
Chronic Pancreatitis	MDS Item I8000	1
Proliferative Diabetic Retinopathy and Vitreous Hemorrhage	MDS Item I8000	1
Other Foot Skin Problems: Foot Infection Code, Other Open Lesion on Foot Code, Except Diabetic Foot Ulcer Code	MDS Item M1040A, M1040B, M1040C	1
Complications of Specified Implanted Device or Graft	MDS Item I8000	1
Bladder and Bowel Appliances: Intermittent Catheterization	MDS Item H0100D	1
Inflammatory Bowel Disease	MDS Item I8000	1
Aseptic Necrosis of Bone	MDS Item I8000	1
Special Treatments/Programs: Suctioning Post-admit Code	MDS Item O0100D2	1
Cardio-Respiratory Failure and Shock	MDS Item I8000	1
Myelodysplastic Syndromes and Myelofibrosis	MDS Item I8000	1
Systemic Lupus Erythematosus, Other Connective Tissue Disorders, and Inflammatory Spondylopathies	MDS Item I8000	1
Diabetic Retinopathy—Except Proliferative Diabetic Retinopathy and Vitreous Hemorrhage	MDS Item I8000	1
Nutritional Approaches While a Resident: Feeding Tube	MDS Item K0510B2	1
Severe Skin Burn or Condition	MDS Item I8000	1
Intractable Epilepsy	MDS Item I8000	1
Active Diagnoses: Malnutrition Code	MDS Item I5600	1
Disorders of Immunity—Except: RxCC97: Immune Disorders	MDS Item I8000	1
Cirrhosis of Liver	MDS Item I8000	1
Bladder and Bowel Appliances: Ostomy	MDS Item H0100C	1
Respiratory Arrest	MDS Item I8000	1
Pulmonary Fibrosis and Other Chronic Lung Disorders	MDS Item I8000	1

Given the NTA scoring methodology described in the proposed rule (83 FR 21058 through 21059) and above, and following the same methodology used for the PT, OT, and SLP components, we used the CART algorithm to determine the most appropriate splits in resident NTA case-mix groups. This methodology is more thoroughly explained in sections 3.4.2. and 3.7.2. of the SNF PDPM technical report available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html. Based on the breakpoints generated by the CART algorithm, we determined that 6 case-mix groups would be necessary to classify residents adequately in terms of their NTA costs in a manner that captures sufficient variation in NTA costs without creating unnecessarily granular separations. As discussed in the proposed rule, we made certain administrative decisions that further refined the NTA case-mix classification groups beyond those produced through use of the CART algorithm but maintained the CART output predictive accuracy. We explained that the proposed NTA case-mix classification departs from the CART comorbidity score bins in grouping residents with a comorbidity score of 1 with residents with scores of 2 instead of with residents with scores of 0. This is to maintain the distinction between residents with no comorbidities and the rest of the population. In addition, we grouped residents with a score of 5 together with residents with scores of 3 to 4 based on their similarity in average NTA costs per day. More information on this analysis can be found in section 3.7.2. of the SNF PDPM technical report available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html. We provided the criteria for each of these groups along with its CMI in Table 28 of the proposed rule (set forth in Table 28 of this final rule).

We stated in the proposed rule (83 FR 21059) that to help ensure payment reflects the average relative resource use at the per diem level, CMIs would be set to reflect relative case-mix related differences in costs across groups. We further stated that this method helps ensure that the share of payment for each case-mix group would be equal to its share of total costs of the component. CMIs for the NTA component were calculated based on two factors. One factor was the average per diem costs of a case-mix group relative to the population average. The other factor was the average variable per diem adjustment factor of the group relative to the population average. In this calculation, average per diem costs equaled total NTA costs in the group divided by number of utilization days in the group. Similarly, the average variable per diem adjustment factor equaled the sum of NTA variable per diem adjustment factors for all utilization days in the group divided by the number of utilization days in the group. We calculated CMIs such that they equaled the ratio of relative average per diem costs for a group to the relative average variable per diem adjustment factor for the group. In this calculation, relative average per diem costs and the relative average variable per diem adjustment factor were weighted by length of stay to account for the different length of stay distributions across case-mix groups (as further discussed in section 3.11.1. of the SNF PDPM technical report, available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html). After calculating CMIs as described above, we then applied adjustments to ensure that the distribution of resources across payment components was aligned with the statutory base rates as discussed in section V.D.3.b. of the proposed rule. We also applied a parity adjustment by multiplying the CMIs by the ratio of case-mix-related payments in RUG-IV over estimated case-mix-related payments in PDPM, as further discussed in section V.J. of the proposed rule. More information on the variable per diem adjustment factor is discussed in section V.D.4. of the proposed rule. The full methodology used to develop CMIs is presented in section 3.11. of the SNF PDPM technical report (available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html).

Table 28—NTA Case-Mix Classification Groups

NTA score range	NTA case-mix group	NTA case-mix index
12+	NA	3.25
9-11	NB	2.53
6-8	NC	1.85
3-5	ND	1.34
1-2	NE	0.96
0	NF	0.72

We stated in the proposed rule (83 FR 21059) that as with the previously discussed components, all residents would be classified into one and only one of these 6 NTA case-mix groups under the proposed PDPM. We explained that the proposed PDPM would create a separate payment component for NTA services, as opposed to combining NTA and nursing into one component as in the RUG-IV system. This separation would allow payment for NTA services to be based on resident characteristics that predict NTA resource utilization rather than nursing staff time. Thus, we stated that we believe the proposed NTA case-mix groups would provide a better measure of resource utilization and lead to more accurate payments under the SNF PPS.

We invited comments on the approach proposed above to classify residents for NTA payment under the proposed PDPM.

Commenters submitted the following comments related to the proposed rule's discussion of the classification of residents for NTA payment under the PDPM. A discussion of these comments, along with our responses, appears below.

Comment: A few commenters recommended CMS include additional conditions as comorbidities for NTA classification and payment, including: Parkinson's disease, non-refractory epilepsy/seizure disorders, and mental health conditions that bear a strong relationship to NTA utilization. One commenter recommended that CMS include all conditions associated with higher NTA costs, not only the 50 costliest comorbidities. Another commenter suggested implementing a periodic review process to update the NTA comorbidity list based on changes in care practices. One commenter recommended CMS add cardio-respiratory failure and shock, respiratory arrest, pulmonary fibrosis or other chronic lung disorders, oxygen therapy, and non-invasive ventilation (for example, BiPAP/CPAP) to the NTA comorbidity list, as these conditions/services reflect medical complexity and high acuity. Another commenter stated that NTA comorbidities should include wound care and all pressure ulcers, not only stage 4 pressure ulcers.

Response: As described in section 3.7.1 of the SNF PDPM technical report, we investigated a broad list of conditions and services as potential NTA comorbidities, defined using MDS items, ICD-10-CM diagnoses, and CCs and RxCCs from the Medicare Parts C and D risk adjustment models. We used MDS item I5300 to identify residents with Parkinson's disease, RxCC 164 to identify residents with non-refractory epilepsy, CC 84 to identify residents with cardio-respiratory failure and shock, CC 83 to identify residents with respiratory arrest, CC 112 and RxCC 227 to identify residents with pulmonary fibrosis or other chronic lung disorders, MDS item M1200F to identify residents receiving wound care, and MDS item M0300X1 to identify residents with a pressure ulcer. For mental health conditions, we used RxCC 135 to identify residents with anxiety, RxCC 133 to identify residents with specified anxiety, personality, and behavior disorders, RxCC 132 and 134 to identify residents with depression, CC 58 to identify residents with Major Depressive, Bipolar, and Paranoid Disorder, CC 57, RxCC 130 to identify residents with schizophrenia, CC 54 to identify residents with drug or alcohol psychosis, and CC 55 to identify residents with drug or alcohol dependence. Neither Parkinson's disease, non-refractory epilepsy, pulmonary fibrosis or other chronic lung disorders, nor any mental health condition were among the top 50 costliest conditions/services in terms of NTA utilization. Non-refractory epilepsy was associated with an increase of about $1.60 in NTA costs per day, while Parkinson's disease was associated with an increase of about $2.50 in NTA costs per day and pulmonary fibrosis or other chronic lung disorders were associated with an increase of about $4 per day in NTA costs. Wound care was associated with an increase of about $2 in NTA costs per day, while stage 3 pressure ulcers (the next highest level of severity after stage 4) were associated with an increase of about $1 in NTA costs per day. Among mental health conditions, major depression was the most costly and associated with an increase of about $4 per day in NTA costs. The other mental health conditions were associated with less than $2 in NTA costs per day. In contrast, the least costly comorbidity included in the final list of included comorbidities for NTA classification and payment was associated with an increase of about $4.50 in NTA costs per day. Therefore, these conditions were not included as NTA comorbidities. On the other hand, cardio-respiratory failure and shock, as well as respiratory arrest were found to be among the 50 costliest conditions in terms of NTA utilization. Therefore, these two conditions were included in the final list of NTA comorbidities. As for oxygen therapy and non-invasive ventilation such as BiPAP and CPAP, clinicians advised CMS that it was not appropriate to include these services in the payment model because their inclusion would likely lead to inappropriate provision of these services in excess of clinical need. We do not believe it is appropriate to include conditions/services that do not have a notable impact on NTA costs per day, and therefore, we only included the 50 costliest comorbidities.

Comment: A commenter states that the points assigned to ventilator care should be much higher because this service requires 24-hour assistance. Additionally, this commenter requests CMS modify the term “ventilator/respirator” to only “ventilator” as the term “respirator” is outdated. Another commenter recommended further evaluation of the proposed point assignment, particularly for pressure ulcers, diabetic ulcers, respiratory failure, severe burns, multi-drug resistant organisms, and morbid obesity. According to the commenter, these items require higher resource utilization compared to other conditions/services that are assigned the same number of points.

Response: As described in section 3.7.1 of the SNF PDPM technical report, after determining the 50 costliest comorbidities in terms of NTA utilization, we ran an OLS regression to estimate the increase in NTA costs associated with each included condition or service. We then assigned points to each condition/service in proportion to the associated increase in NTA costs by dividing the coefficient for each condition or service by 10 and then rounding to the nearest integer. Based on this procedure, we assigned 4 points to ventilator/respirator care to reflect our finding that this service was associated with an increase of about $40 in NTA costs per day. Using the same procedure, we assigned 1 point to stage 4 pressure ulcers, diabetic foot ulcers, respiratory failure, severe burns, multi-drug resistant organisms, and morbid obesity as each of these conditions was associated with an increase of roughly $10 in NTA costs per day. Therefore, our analysis does not support increasing the points assigned to these services. The nomenclature used to refer to ventilator/respirator care under PDPM is consistent with the description of this service on the current version of the MDS 3.0 assessment. We appreciate the feedback on the appropriateness of the current name and will consider modifying the name of this item as appropriate to reflect current usage.

Comment: One commenter states that given the theoretical maximum NTA score is 83, the highest NTA score bin should not be 12+. This commenter suggests creating additional score bins at the upper end of the score, such as 12-14, 15-17, and 18+, to more accurately reflect residents with highly complex conditions and multiple extensive services.

Response: While it is true that some stays have very high NTA costs, we find that stays with an NTA comorbidity score of 12 or above are very rare (about 1 percent of all stays). As the number of stays included in each group declines, the magnitude of the standard error associated with the estimate of a group's resource utilization increases, raising concerns about the precision of these estimates. For the foregoing reasons, we do not believe it is appropriate to add additional NTA groups to include residents with extraordinarily high NTA utilization at this time. We will also consider revisiting both the list of included NTA comorbidities and the points assigned to each condition/service based on changes in the resident population and care practices over time.

Comment: Another commenter expressed concern that NTA costs, especially high-cost cases, are impossible to predict through use of existing administrative data due to the small sample size of these high-cost outliers. Since PDPM was developed on data that may fail to account for high-cost outliers, the commenter believes that PDPM is not sufficient to explain NTA utilization and will underpay the actual high-cost cases that cannot be predicted. One commenter questioned the validity of current NTA data, stating that providers do not accurately record NTA costs because these services are not important determinants of payment under RUG-IV. As a result, current data may underestimate NTA costs. Therefore, PDPM may not accurately reimburse NTA utilization.

Response: As shown in section 3.7.1 of the SNF PDPM technical report, average NTA costs per day by comorbidity score varies from around $30 to near $180, which indicates the data being used captures great variations of NTA costs and includes many expensive cases. The NTA comorbidity list as shown in Table 27 of the proposed rule (83 FR 21058) captures comorbidities and services with high NTA costs. Moreover the selected comorbidities and services meet clinical expectations of conditions that are expected to require high NTA utilization. Although the data available may be limited in capturing high-cost cases due to the small sample size of less common comorbidities, the proposed rule (83 FR 21073 through 21077) and section 3.12 of the SNF PDPM technical report show that payments for beneficiaries with high NTA costs will increase notably under PDPM compared with RUG IV. In particular, our impact analysis finds that payment increases by 27.2 percent for residents with 12 or more conditions under PDPM compared to RUG-IV.

Regarding the concern that current administrative data may not fully capture NTA utilization for the SNF population, first, as described in Section 3.2.2. of the SNF PDPM technical report, we checked the quality of self-reported NTA utilization data by comparing charges from cost reports and charges from claims and verifying that these were generally consistent. Second, we used four years of data (FYs 2014-2017) to identify the conditions and services associated with high NTA utilization and assign points to these comorbidities reflective of their impact on resource use. Using several years of data addresses a key concern of commenters responding to the 2017 ANPRM and ensures a higher level of robustness compared to using a single year of data. Third, if NTA utilization is indeed underreported overall, this should not affect relative NTA resource use across different types of residents, therefore PDPM should still assign payment appropriately based on observed relative differences in NTA utilization. Fourth, clinicians reviewed the proposed NTA classification and verified that it accords with clinical expectations regarding conditions and services that are associated with high NTA utilization. Finally, as SNF care practices and reporting patterns change in response to the new payment model and other factors, we will consider revising elements of PDPM, including the NTA comorbidities, to reflect changes in relative resource use.

Comment: One commenter requested clarification on the NTA comorbidity list change from RCS-I to PDPM.

Response: The change in the comorbidity list from RCS-I to PDPM is due to the following: first, we used 4 years of data (FY2014-FY2017) under PDPM instead of a single year of data under RCS-I to make the list more robust to changes in the SNF population and care practices over time; second, we added Part D condition categories to better capture conditions associated with high medication costs; finally, we expanded the list to the top 50 comorbidities to include more conditions.

Comment: One commenter recommended that PDPM include an NTA default category to accommodate new conditions and services for greater flexibility.

Response: We are not clear on how such a default category would operate, nor what level of reimbursement would be appropriate to set for the addition of new conditions and services. We would need additional information on how such a default category could be constructed.

Comment: One commenter expressed concern regarding access to novel therapies, and encouraged CMS to consider adding a new technology add-on payment, similar to that done for inpatient hospitals, and make additional payments to SNFs when new treatment options become available. One commenter also stated that PDPM does not account for new classes of expensive medications.

Response: The points associated with each NTA comorbidity under the PDPM are based on existing cost data, which may be updated in future years to reflect the costs of new technologies and treatments or new classes of medications. Rather than merely incentivizing new treatments, we expect providers to utilize the best treatments for a given patient, which may or may not be more costly than existing treatments. Further, we note that the inpatient hospital PPS's new technology add-on payment is specifically authorized by sections 1886(d)(5)(K) and (L) of the Act, whereas no similar statutory authority exists under the SNF PPS.

Comment: One commenter expressed concern that using a patient's NTA score as a first tier classification criterion could put providers at risk of late or missing IPAs.

Response: As discussed in section V.D of this final rule, the IPA would be an optional assessment and, as such, not susceptible to late or missed assessment penalties.

Accordingly, after considering the comments received, for the reasons discussed in the proposed rule and in this final rule, we are finalizing the proposed NTA component of the PDPM and the proposed classification methodology for the NTA component, without modification.

f. Payment Classifications Under PDPM

RUG-IV classifies each resident into a single RUG, with a single payment for all services. By contrast, the PDPM classifies each resident into five components (PT, OT, SLP, NTA, and nursing) and provide a single payment based on the sum of these individual classifications. The payment for each component would be calculated by multiplying the CMI for the resident's group first by the component federal base payment rate, then by the specific day in the variable per diem adjustment schedule (as discussed in section V.D.4 of the proposed rule and in section V.D.4 of this final rule). Additionally, for residents with HIV/AIDS indicated on their claim, the nursing portion of payment would be multiplied by 1.18 (as discussed in section V.D.3.d. of the proposed rule and section V.H of this final rule). These payments would then be added together along with the non-case-mix component payment rate to create a resident's total SNF PPS per diem rate under the PDPM. This section describes how two hypothetical residents would be classified into payment groups under the current RUG-IV model and PDPM. To begin, consider two residents, Resident A and Resident B, with the resident characteristics identified in Table 29.

Table 29—Hypothetical Resident Characteristics

Resident characteristics	Resident A	Resident B
Rehabilitation Received?	Yes	Yes.
Therapy Minutes	730	730.
Extensive Services	No	No.
ADL Score	9	9.
Clinical Category	Acute Neurologic	Major Joint Replacement.
PT and OT Function Score	10	10.
Nursing Function Score	7	7.
Cognitive Impairment	Moderate	Intact.
Swallowing Disorder?	No	No.
Mechanically Altered Diet?	Yes	No.
SLP Comorbidity?	No	No.
Comorbidity Score	7 (IV Medication and DM)	1 (Chronic Pancreatitis).
Other Conditions	Dialysis	Septicemia.
Depression?	No	Yes.

Currently under the SNF PPS, Resident A and Resident B would be classified into the same RUG-IV group. They both received rehabilitation, did not receive extensive services, received 730 minutes of therapy, and have an ADL score of 9. This places the two residents into the “RUB” RUG-IV group and SNFs would be paid at the same rate, despite the many differences between these two residents in terms of their characteristics, expected care needs, and predicted costs of care.

Under the PDPM, however, these two residents would be classified very differently. With regard to the PT and OT components, Resident A would fall into group TO, as a result of his categorization in the Acute Neurologic group and a function score within the 10 to 23 range. Resident B, however, would fall into group TC for the PT and OT components, as a result of his categorization in the Major Joint Replacement group and a function score within the 10 to 23 range. For the SLP component, Resident A would be classified into group SH, based on his categorization in the Acute Neurologic group, the presence of moderate cognitive impairment, and the presence of Mechanically-Altered Diet, while Resident B would be classified into group SA, based on his categorization in the Non-Neurologic group, the absence of cognitive impairment or any SLP-related comorbidity, and the lack of any swallowing disorder or mechanically-altered diet. For the Nursing component, following the existing nursing case-mix methodology, Resident A would fall into group LBC1, based on his use of dialysis services and a nursing function score of 7, while Resident B would fall into group HBC2, due to the diagnosis of septicemia, presence of depression, and a nursing function score of 7. Finally, with regard to NTA classification, Resident A would be classified in group NC, with an NTA score of 7, while Resident B would be classified in group NE, with an NTA score of 1. This demonstrates that, under the PDPM, more aspects of a resident's unique characteristics and needs factor into determining the resident's payment classification, which makes for a more resident-centered case-mix model while also eliminating, or greatly reducing, the number of service-based factors which are used to determine the resident's payment classification. Because this system is based on specific resident characteristics predictive of resource utilization for each component, we expect that payments will be better aligned with resident need.

4. Variable per Diem Adjustment Factors and Payment Schedule

Section 1888(e)(4)(G)(i) of the Act provides that payments must be adjusted for case mix, based on a resident classification system which accounts for the relative resource utilization of different types of residents. Additionally, section 1888(e)(1)(B) of the Act specifies that payments to SNFs through the SNF PPS must be made on a per-diem basis. Currently under the SNF PPS, each RUG is paid at a constant per diem rate, regardless of how many days a resident is classified in that particular RUG. However, we explained in the proposed rule (83 FR 21060) that during the course of the SNF PMR project, analyses on cost over the stay for each of the case-mix adjusted components revealed different trends in resource utilization over the course of the SNF stay. These analyses utilized costs derived from claim charges as a measure of resource utilization. Costs were derived by multiplying charges from claims by the CCRs on facility-level costs reports. As described in section V.B.3.b. of the proposed rule, costs better reflect differences in the relative resource use of residents as opposed to charges, which partly reflect decisions made by providers about how much to charge payers for certain services. In examining costs over a stay, we stated we found that for certain categories of SNF services, notably PT, OT and NTA services, costs declined over the course of a stay. Based on the claim submission schedule and variation in the point during the month when a stay began, we were able to estimate resource use for a specific day in a stay. Facilities are required to submit monthly claims. Each claim covers the period from the first day during the month a resident is in the facility to the end of the month. If a resident was admitted on the first day of the month, remains in the facility, and continues to have Part A SNF coverage until the end of the month, the claim for that month will include all days in the month. However, if a resident is admitted after the first day of the month, the first claim associated with the resident's stay will be shorter than a month. We stated in the proposed rule that to estimate resource utilization for each day in the stay, we used the marginal estimated cost from claims of varying length based on random variation in the day of a month when a stay began. Using this methodology, we observed a decline in the marginal estimated cost of each additional day of SNF care over the course of the stay. We further stated that to supplement this analysis, we also looked at changes in the number of therapy minutes reported in different assessments throughout the stay. Because therapy minutes are recorded on the MDS, the presence of multiple assessments throughout the stay provided information on changes in resource use. For example, it was clear whether the number of therapy minutes a resident received changed from the 5-day assessment to the 14-day assessment. We explained that the results from this analysis were consistent with the cost from claims analysis and showed that, on average, the number of therapy minutes is lower for assessments conducted later in the stay. This finding was consistent across different lengths of stay. More information on these analyses can be found in section 3.9. of the SNF PDPM technical report and section 3.9. of the SNF PMR technical report that accompanied the ANPRM, both available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html.

As discussed in the proposed rule (83 FR 21060 through 21061), analyses of the SLP component revealed that the per diem costs remain relatively constant over time, while the PT, OT, and NTA component cost analyses indicate that the per diem cost for these three components decline over the course of the stay. We stated in the proposed rule that in the case of the PT and OT components, costs start higher at the beginning of the stay and decline slowly over the course of the stay. By comparison, the NTA component cost analyses indicated significantly increased NTA costs at the beginning of a stay that then drop to a much lower level that holds relatively constant over the remainder of the SNF stay. This is consistent with how most SNF drug costs are typically incurred at the outset of a SNF stay. We stated that these results indicate that resource utilization for PT, OT, and NTA services changes over the course of the stay. More information on these analyses can be found in section 3.9.1. of the SNF PMR technical report that accompanied the ANPRM available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html. As we stated in the proposed rule, we were unable to assess potential changes in the level of nursing costs over a resident's stay, in particular because nursing charges are not separately identifiable in SNF claims, and nursing minutes are not reported on the MDS assessments. However, stakeholders (industry representatives and clinicians) at multiple TEPs indicated that nursing costs tend to remain relatively constant over the course of a resident's stay.

We explained in the proposed rule that constant per diem rates, by definition, do not track variations in resource use throughout a SNF stay. We stated we believe this may lead to too few resources being allocated for SNF providers at the beginning of a stay. Given the trends in resource utilization over the course of a SNF stay discussed above, and that section 1888(e)(4)(G)(i) of the Act requires the case-mix classification system to account for relative resource use, we proposed adjustments to the PT, OT, and NTA components in the proposed PDPM to account for changes in resource utilization over a stay. These adjustments were referred to as the variable per diem adjustments. We did not propose such adjustments to the SLP and nursing components based on findings and stakeholder feedback, as discussed above, that resource use tends to remain relatively constant over the course of a SNF stay.

As noted above and in the proposed rule (83 FR 21061), and discussed more thoroughly in section 3.9. of the SNF PMR technical report that accompanied the ANPRM (available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html), PT and OT costs decline at a slower rate than the decline in NTA costs. Therefore, in addition to proposing a variable per diem adjustment, we further proposed separate adjustment schedules and indexes for the PT and OT components and the NTA component to reflect more closely the rate of decline in resource utilization for each component. Table 30 of the proposed rule provided the adjustment factors and schedule that we proposed for the PT and OT components, while Table 31 of the proposed rule provided the adjustment factors and schedule that we proposed for the NTA component.

In Table 30 of the proposed rule, the adjustment factor for the PT and OT components was 1.00 for days 1 to 20. We explained that this was because the analyses described above indicated that PT and OT costs remain relatively high for the first 20 days and then decline. The estimated daily rates of decline for PT and OT costs relative to the initial 20 days were both 0.3 percent. Thus, we stated that a convenient and appropriate way to reflect this was to bin days in the PT and OT variable per diem adjustment schedules such that payment declines at less frequent intervals, while still reflecting a 0.3 percent daily rate of decline in PT and OT costs. Therefore, we proposed to set the adjustment factors such that payment would decline 2 percent every 7 days after day 20 (0.3 * 7 = 2.1). We explained that the 0.3 percent rate of decline was derived from a regression model that estimates the level of resource use for each day in the stay relative to the beginning of the stay. The regression methodology and results are presented in section 3.9. of the SNF PDPM technical report, available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html.

As described previously in this section and in the proposed rule (83 FR 21061), NTA resource utilization exhibits a somewhat different pattern. The analyses described above indicate that NTA costs are very high at the beginning of the stay, drop rapidly after the first 3 days, and remain relatively stable from the fourth day of the stay. We stated that starting on day 4 of a stay, the per diem costs drop to roughly one-third of the per diem costs in the initial 3 days. We explained that this suggests that many NTA services are provided in the first few days of a SNF stay. Therefore, we proposed setting the NTA adjustment factor to 3.00 for days 1 to 3 to reflect the extremely high initial costs, then setting it at 1.00 (two-thirds lower than the initial level) for subsequent days. We explained that the value of the adjustment factor was set at 3.00 for the first 3 days and 1.00 after (rather than, for example, 1.00 and 0.33, respectively) for simplicity. The results are presented in section 3.9. of the SNF PDPM technical report, available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html.

As we described in the proposed rule (83 FR 21061), case-mix adjusted federal per diem payment for a given component and a given day would be equal to the base rate for the relevant component (either urban or rural), multiplied by the CMI for that resident, multiplied by the variable per diem adjustment factor for that specific day, as applicable. Additionally, as described in further detail in section V.D.3.d. of the proposed rule, we stated that an additional 18 percent would be added to the nursing per-diem payment to account for the additional nursing costs associated with residents who have HIV/AIDS. We further explained that these payments would then be added together along with the non-case-mix component payment rate to create a resident's total SNF PPS per diem rate under the proposed PDPM. We invited comments on the proposed variable per diem adjustment factors and payment schedules discussed in this section.

Commenters submitted the following comments related to the proposed rule's discussion of the variable per diem adjustment factors and payment schedules. A discussion of these comments, along with our responses, appears below.

Comment: Some commenters supported the use of the variable per diem adjustment under PDPM. Several commenters stated that PDPM, specifically the variable per diem payment adjustments included in the PT, OT, and NTA components, may negatively affect access for beneficiaries with long stays and complex medical needs. These commenters stated that the variable per diem payment adjustments will encourage early discharges and the provision of fewer services. One commenter stated that residents with chronic conditions may not exhibit a trend of declining NTA utilization over a stay and that resource utilization for these patients is sustained at a relatively constant level throughout the stay. The commenter states that in these cases, variable per diem payment adjustments will incentivize facilities to provide less expensive medications later in the stay, which the commenter states may be harmful to the patient. Finally, one commenter seeks clarification on the rationale for beginning the decline in payment for the PT and OT components after day 20 of a stay.

Response: We note that we investigated the impact of PDPM on various resident subpopulations, including residents with many expensive comorbidities, multiple extensive services, severe cognitive impairment, long stays (utilization days = 100), stroke, IV medication, diabetes, wound infection, amputation/prosthesis care, Alzheimer's, or the presence of addictions, bleeding disorders, behavioral issues, chronic neurological conditions, or bariatric care. CMS investigated the potential impact of the proposed payment model on these subpopulations based on comments received in response to the 2017 ANPRM. For almost all of these subpopulations with complex medical needs, we found that PDPM is estimated to increase payment associated with these residents compared to RUG-IV, as discussed in the proposed rule (83 FR 21075) and section 3.12 of the SNF PDPM technical report. Thus, we do not believe the variable per diem payment will negatively affect access for beneficiaries with expensive comorbidities or complex medical needs. We estimated that payment associated with very long stays (utilization = 100 days) would decline by 1.9 percent under PDPM, and we obtained similar results for stays longer than 90 days. However, this decline in payment is a reflection of the lower resource utilization per day associated with longer stays. We observed that stays longer than 90 days have lower therapy and NTA costs per day than their shorter counterparts. However, the majority of such long stays are categorized as ultra-high rehabilitation groups in the current case-mix classification suggesting potential overpayment. Nevertheless, given the potential payment reduction for long stays, we plan to monitor provider behavior closely to identify facilities whose beneficiaries experience inappropriate early discharge or provision of fewer services.

Regarding the concern about resource utilization patterns of residents with chronic conditions, we would note that as discussed above, we estimated that PDPM would actually increase overall per-stay payment for many resident subpopulations with chronic conditions. Further, while payment would be highest during the early part of a stay, facilities would have flexibility to allocate this payment to cover costs later in a stay, as they do now. Our research, discussed in the proposed rule (83 FR 21061) and section 3.9 of the SNF PDPM technical report, revealed that for the average SNF stay, NTA utilization declines dramatically after the first 3 days of a stay. Of course, we acknowledge that there are cases that may not match this resource utilization pattern exactly. However, we believe that PDPM, because it is based on the observed relationship between patient characteristics and resource utilization, represents an improvement over the current payment model in terms of payment accuracy. Further, as stated, our investigations show that for many of the specific cases cited by commenters as potential concerns, we expect PDPM actually to increase associated payment compared to RUG-IV. While the variable per diem schedule decreases pay throughout the stay, the overall increase in payment accounts for the treatment cost of chronic conditions, which is costly due to the sustained level of care needed to manage chronic conditions.

As discussed in the proposed rule (83 FR 21060 through 21061) and section 3.9 of the SNF PDPM technical report, we developed a methodology to estimate per-diem resource use over a stay for PT, OT, and NTA. Based on this methodology, we observed that estimated per-diem PT and OT costs remain high for the first 20 days of a stay and decline thereafter. Therefore, we established a variable per diem payment adjustment schedule for the PT and OT components that begins to adjust payment downward beginning on day 21.

Comment: Some commenters suggested that CMS consider creating a waiver from the variable per diem adjustment for NTAs to mitigate potential access issues for patients in SNF stays that exceed 90 days. Additionally, these commenters expressed concern that, for long stays, the variable per diem payment adjustment may erode payments to the point where payment for the stay is below the cost of providing the associated services. Some commenters believe that decreasing payment for PT and OT over the course of the stay without exceptions is not patient-centered and urged CMS to identify certain diagnoses associated with longer duration of high-intensity therapy services as exceptions to the variable per diem schedule. Several commenters requested clarification on if and how CMS intends to monitor the impact of the variable per diem adjustment on patient access and length of stay, expressing concerns that the variable per diem adjustment could have a disproportionate impact on patients with chronic conditions. Finally, one commenter believed that reducing payments over time through the variable per diem adjustment will reduce treatment options for stroke and trauma victims.

Response: With regard to the waiver for either the PT and OT variable per diem adjustment or the NTA variable per diem adjustment in cases of long stays, we do not believe that such a waiver is necessary. While payments do reduce over time, as discussed above, this reduction is to reflect the decrease in patient costs over time. Therefore, given the parallel reductions in costs and payments, over the course of the stay, providers should be adequately reimbursed for the provision of care, even in cases of long stays. With regard to the commenters' concern regarding the impact on stroke and trauma patients, as well as those with chronic conditions, we do plan to monitor closely these types of SNF patients under PDPM to identify any adverse trends which may result from application of the variable per diem adjustment. That being said, given that we proposed to implement PDPM in a budget neutral manner, this means that while the overall sum of monies paid out under the SNF benefit would not change under PDPM, the allocation and distribution of that money to individual SNFs could change. Given that PT, OT, and NTA costs at the beginning of a stay tend to be higher than those at the middle or end of a stay, most notably in the case of long stay patients, maintaining a constant per diem rate will allocate too few funds at the beginning of the stay, thereby increasing the chance that the early portions of a stay may not be adequately reimbursed. By aligning the payments with the cost trends, this produces the best chance to ensure that providers receive adequate and appropriate reimbursement at every point in the stay. Finally, as stated above, we do plan to monitor the impact of this policy and may consider revisions to the policy if there is evidence of adverse trends either systemically or within certain patient populations.

Comment: One commenter questioned if CMS would expect the variable per diem adjustment to continue until the payment reaches zero, for purposes of calculating the UPL for the PT and OT components.

Response: As the variable per diem adjustment was developed based on Medicare Part A data, we cannot speak to the ability of the adjustment factor to be drawn out past the point of the Medicare Part A stay. Moreover, as coverage for a Medicare Part A stay cannot be longer than 100 days, the variable per diem adjustment, for purposes of calculating the UPL, would go as far as Day 100 in Table 30.

Accordingly, after considering the comments received, for the reasons discussed in the proposed rule and in this final rule, we are finalizing our proposal without modification to apply a variable per diem adjustment as part of the PDPM effective October 1, 2019. Table 30 sets forth the final PDPM Variable Per Diem Payment Adjustment Factors and Schedule for the PT and OT components, and Table 31 sets for the final PDPM Variable Per Diem Payment Adjustment Factors and Schedule for the NTA component.

Table 30—Variable Per-Diem Adjustment Factors and Schedule—PT and OT

Medicare payment days	Adjustment factor
1-20	1.00
21-27	0.98
28-34	0.96
35-41	0.94
42-48	0.92
49-55	0.90
56-62	0.88
63-69	0.86
70-76	0.84
77-83	0.82
84-90	0.80
91-97	0.78
98-100	0.76

Table 31—Variable Per-Diem Adjustment Factors and Schedule—NTA

Medicare payment days	Adjustment factor
1-3	3.0
4-100	1.0

D. Use of the Resident Assessment Instrument—Minimum Data Set, Version 3

1. Revisions to Minimum Data Set (MDS) Completion Schedule

Consistent with section 1888(e)(6)(B) of the Act, to classify residents under the SNF PPS, we use the MDS 3.0 Resident Assessment Instrument. Within the SNF PPS, there are two categories of assessments, scheduled and unscheduled. In terms of scheduled assessments, SNFs are currently required to complete assessments on or around days 5, 14, 30, 60, and 90 of a resident's Part A SNF stay, including certain grace days. Payments based on these assessments depend upon standard Medicare payment windows associated with each scheduled assessment. More specifically, each of the Medicare-required scheduled assessments has defined days within which the Assessment Reference Date (ARD) must be set. The ARD is the last day of the observation (or “look-back”) period that the assessment covers for the resident. The facility is required to set the ARD on the MDS form itself or in the facility software within the appropriate timeframe of the assessment type being completed. The clinical data collected from the look-back period is used to determine the payment associated with each assessment. For example, the ARD for the 5-day PPS Assessment is any day between days 1 to 8 (including Grace Days). The clinical data collected during the look-back period for that assessment is used to determine the SNF payment for days 1 to 14. Unscheduled assessments, such as the Start of Therapy (SOT) Other Medicare Required Assessment (OMRA), the End of Therapy OMRA (EOT OMRA), the Change of Therapy (COT) OMRA, and the Significant Change in Status Assessment (SCSA or Significant Change), may be required during the resident's Part A SNF stay when triggered by certain defined events.

For example, if a resident is being discharged from therapy services, but remaining within the facility to continue the Part A stay, then the facility may be required to complete an EOT OMRA. Each of the unscheduled assessments affects payment in different and defined manners. A description of the SNF PPS scheduled and unscheduled assessments, including the criteria for using each assessment, the assessment schedule, payment days covered by each assessment, and other related policies, are set forth in the MDS 3.0 RAI manual on the CMS website (available at https://downloads.cms.gov/​files/​MDS-30-RAI-Manual-v115-October-2017.pdf).

Table 32 outlines when each of the current SNF PPS assessments is required to be completed and its effect on SNF PPS payment.

Table 32—Current PPS Assessment Schedule

Medicare MDS assessment schedule type	Assessment reference date	Assessment reference date grace days	Applicable standard medicare payment days
Scheduled PPS assessments
5-day	Days 1-5	6-8	1 through 14.
14-day	Days 13-14	15-18	15 through 30.
30-day	Days 27-29	30-33	31 through 60.
60-day	Days 57-59	60-63	61 through 90.
90-day	Days 87-89	90-93	91 through 100.

Unscheduled PPS assessments
Start of Therapy OMRA	5-7 days after the start of therapy	Date of the first day of therapy through the end of the standard payment period.
End of Therapy OMRA	1-3 days after all therapy has ended	First non-therapy day through the end of the standard payment period.
Change of Therapy OMRA	Day 7 (last day) of the COT observation period	The first day of the COT observation period until end of standard payment period, or until interrupted by the next COT-OMRA assessment or scheduled or unscheduled PPS Assessment.
Significant Change in Status Assessment	No later than 14 days after significant change identified	ARD of Assessment through the end of the standard payment period.

As we explained in the proposed rule (83 FR 21062), an issue which has been raised in the past with regard to the existing SNF PPS assessment schedule is that the sheer number of assessments, as well as the complex interplay of the assessment rules, significantly increases the administrative burden associated with the SNF PPS. We stated that case-mix classification under the proposed SNF PDPM relies to a much lesser extent on characteristics that may change very frequently over the course of a resident's stay (for example, therapy minutes may change due to resident refusal or unexpected changes in resident status), but instead relies on more stable predictors of resource utilization by tying case-mix classification, to a much greater extent, to resident characteristics such as diagnosis information. We explained that in view of the greater reliance of the proposed SNF PDPM (as compared to the RUG-IV model) on resident characteristics that are relatively stable over a stay and our general focus on reducing administrative burden for providers across the Medicare program, we are making an effort to reduce the administrative burden on providers by concurrently proposing to revise the assessments that would be required under the proposed SNF PDPM. Specifically, we proposed to use the 5-day SNF PPS scheduled assessment to classify a resident under the proposed SNF PDPM for the entirety of his or her Part A SNF stay effective beginning FY 2020 in conjunction with the implementation of the proposed PDPM, except as described below. We stated that if we were to finalize this proposal, we would propose revisions to the regulations at § 413.343(b) during the FY 2020 rulemaking cycle so that such regulations would no longer reflect the RUG-IV SNF PPS assessment schedule as of the proposed conversion to the PDPM on October 1, 2019.

We also stated in the proposed rule (83 FR 21062) that we understand Medicare beneficiaries are each unique and can experience clinical changes which may require a SNF to reassess the resident to capture changes in the resident's condition. Therefore, to allow SNFs to capture these types of changes, effective October 1, 2019 in conjunction with the proposed implementation of the PDPM, we proposed to require providers to reclassify residents as appropriate from the initial 5-day classification using a new assessment called an Interim Payment Assessment (IPA), which would be comprised of the 5-day SNF PPS MDS Item Set (Item Set NP). We stated that providers would be required to complete an IPA in cases where the following two criteria are met:

(1) There is a change in the resident's classification in at least one of the first tier classification criteria for any of the components under the proposed PDPM (which are those clinical or nursing payment criteria identified in the first column in Tables 21, 23, 26, and 27 of the proposed rule), such that the resident would be classified into a classification group for that component that differs from that provided by the 5-day scheduled PPS assessment, and the change in classification group results in a change in payment either in one particular payment component or in the overall payment for the resident; and

(2) The change(s) are such that the resident would not be expected to return to his or her original clinical status within a 14-day period.

In addition, we proposed that the Assessment Reference Date (ARD) for the IPA would be no later than 14 days after a change in a resident's first tier classification criteria is identified. We stated that the IPA is meant to capture substantial changes to a resident's clinical condition and not everyday, frequent changes. We believe 14 days gives the facility an adequate amount of time to determine whether the changes identified are in fact routine or substantial. To clarify, we explained that the change in classification group described above refers not only to a change in one of the first tier classification criteria in any of the proposed payment components, but also to one that would be sufficient to change payment in either one component or in the overall payment for the resident. For example, we stated that given the collapsed categories under the PT and OT components, this would mean that a change from the medical management group to the cancer group would not necessitate an IPA, as they are both collapsed under the medical management group for purposes of the PT and OT components. However, we stated a change from the major joint replacement group to the medical management group would necessitate an IPA, as this would change the resident's clinical category group for purposes of categorization under the PT and OT components and would result in a change in payment.

We stated that we believe the proposed requirement to complete an IPA balances the need to ensure accurate payment and monitor for changes in the resident's condition with the importance of ensuring a more streamlined assessment approach under the proposed PDPM.

In cases where the IPA is required and a facility fails to complete one, we proposed that the facility would follow the guidelines for late and missed unscheduled MDS assessments which are explained in Chapters 2.13 and 6.8 of the MDS RAI Manual (https://downloads.cms.gov/​files/​MDS-30-RAI-Manual-v115-October-2017.pdf). Specifically, we stated in the proposed rule (83 FR 21063) that if the SNF fails to set the ARD within the defined ARD window for an IPA, and the resident is still in a Part A stay, the SNF would be required to complete a late assessment. The ARD can be no earlier than the day the error was identified. We explained that if the ARD on the late assessment is set for a date that is prior to the end of the time period during which the assessment would have controlled the payment, had the ARD been set timely, the SNF would bill the default rate for the number of days that the assessment is out of compliance. This is equal to the number of days between the day following the last day of the available ARD window and the late ARD (including the late ARD). We provided an example where a SNF Part A resident, who is in the major joint replacement payment category for the PT and OT components, develops a skin ulcer of such a nature that, in terms of developing a care and treatment plan for this resident, the skin ulcer takes precedence as the resident's primary diagnosis. As a result, the resident's primary diagnosis, as coded in item I8000, is for this skin ulcer, which would cause him to be classified into the medical management category for these components. The facility notes this clinical change on November 10, 2018. However, they do not complete the IPA until November 26, 2018 which is 16 days after the change in criteria was identified and two days after the ARD window. The facility would bill the default rate for the two days that it was out of compliance. We stated that if the SNF fails to set the ARD for an IPA within the defined ARD window for that assessment, and the resident has been discharged from Part A, the assessment is missed and cannot be completed. We noted that all days that would have been paid by the missed assessment (had it been completed timely) are considered provider-liable. Taking the example above, we stated that if the facility recognized the IPA needed to be completed after the resident has left the building, the facility would be liable for all days from November 10, 2018 until the date of the resident's Part A Discharge.

In addition to proposing to require completion of the IPA as described above, we also considered the implications of a SNF completing an IPA on the variable per diem adjustment schedule described in section V.D.4. the proposed rule. More specifically, in the proposed rule, we considered whether an SNF completing an IPA should cause a reset in the variable per diem adjustment schedule for the associated resident. In examining costs over a stay, we found that for certain categories of SNF services, notably PT, OT, and NTA services, costs declined over the course of a stay. Our analyses showed that, on average, the number of therapy minutes is lower for assessments conducted later in the stay. Additionally, we stated that we were concerned that by providing for the variable per diem adjustment schedule to be reset after an IPA is completed, providers may be incentivized to conduct multiple IPAs during the course of a resident's stay to reset the variable per diem adjustment schedule each time the adjustment is reduced. Therefore, in cases where an IPA is completed, we proposed that this assessment would reclassify the resident for payment purposes as outlined in Table 33 of the proposed rule, but that the resident's variable per diem adjustment schedule would continue rather than being reset on the basis of completing the IPA.

Finally, we stated that believe, regardless of the payment system or case-mix classification model used, residents should continue to receive therapy that is appropriate to their care needs, and this includes both the intensity and modes of therapy utilized. However, we recognized that because the initial 5-day PPS assessment would classify a resident for the entirety of his or her Part A SNF stay (except in cases where an IPA is completed) as outlined above, there would be no mechanism by which SNFs are required to report the amount of therapy provided to a resident over the course of the stay or by which we may monitor that they are in compliance with the proposed 25 percent group and concurrent therapy limit as described in section V.F. of the proposed rule. Therefore, for these reasons, under the proposed PDPM, we proposed to require that SNFs continue to complete the PPS Discharge Assessment, as appropriate (including the proposed therapy items discussed in section V.E.3. of the proposed rule), for each SNF Part A resident at the time of Part A or facility discharge (see section V.E. of this proposed rule for a discussion of our proposed revisions to this assessment to include therapy items). Under the current instructions in the MDS 3.0 RAI manual, the Part A PPS Discharge assessment is completed when a resident's Medicare Part A stay ends, but the resident remains in the facility (MDS 3.0 RAI Manual Chapter 2.5). However, we proposed to require this assessment to be completed at the time of facility discharge for Part A residents as well. Thus, we would continue to collect data on therapy provision as proposed in section V.F. of the proposed rule, to assure that residents are receiving therapy that is reasonable, necessary, and specifically tailored to meet their unique needs. We stated that we believe the combination of the 5-day Scheduled PPS Assessment, the IPA Assessment, and PPS Discharge Assessment would provide flexibility for providers to capture and report accurately the resident's condition, as well as accurately reflect resource utilization associated with that resident, while minimizing the administrative burden on providers under the proposed SNF PDPM.

In addition to these proposed changes, we also examined in the proposed rule (83 FR 21064) the current use of grace days in the MDS assessment schedule. Grace days have been a longstanding part of the SNF PPS. They were created in order to allow clinical flexibility when setting ARD dates of scheduled PPS assessments. In the FY 2012 final rule (76 FR 48519), we discussed that in practice, there is no difference between regular ARD windows and grace days and we encouraged the use of grace days if their use would allow a facility more clinical flexibility or would more accurately capture therapy and other treatments. Thus, we do not intend to penalize any facility that chooses to use the grace days for assessment scheduling or to audit facilities based solely on their regular use of grace days. We may explore the option of incorporating the grace days into the regular ARD window in the future; nevertheless, we will retain them as part of the assessment schedule at the present time consistent with the current policy and the new assessment schedule proposed in the proposed rule.

We proposed, effective beginning October 1, 2019, in conjunction with the proposed implementation of the PDPM, to incorporate the grace days into the existing assessment window. We explained that this proposal would eliminate grace days as such from the SNF PPS assessment calendar and provide for only a standard assessment window. We stated that, as discussed, there is no practical difference between the regular assessment window and grace days and there is no penalty for using grace days. Accordingly, we stated that we believe it would be appropriate to eliminate the use of grace days in PPS assessments.

Table 33 of the proposed rule, set forth at Table 33 of this final rule, sets forth the proposed SNF PPS assessment schedule, incorporating the proposed revisions discussed above, which we stated would be effective October 1, 2019 concurrently with the proposed PDPM.

Table 33—PPS Assessment Schedule Under PDPM

Medicare MDS assessment schedule type	Assessment reference date	Applicable standard Medicare payment days
5-day Scheduled PPS Assessment	Days 1-8	All covered Part A days until Part A discharge (unless an IPA is completed).
Interim Payment Assessment (IPA)	No later than 14 days after change in resident's first tier classification criteria is identified	ARD of the assessment through Part A discharge (unless another IPA assessment is completed).
PPS Discharge Assessment	PPS Discharge: Equal to the End Date of the Most Recent Medicare Stay (A2400C) or End Date	N/A.

We noted in the proposed rule (83 FR 21064) that, as in previous years, we intend to continue to work with providers and software developers to assist them in understanding changes we proposed to the MDS. Further, we noted that none of the proposals related to changes to the MDS assessment schedule should be understood to change any assessment requirements which derive from the Omnibus Budget Reconciliation Act of 1987 (OBRA 87), which establishes assessment requirements for all nursing home residents, regardless of payer. We invited comments on our proposals to revise the SNF PPS assessment schedule and related policies as discussed above.

We also solicited comment on the extent to which implementing these proposals would reduce provider burden.

Commenters submitted the following comments related to the proposed changes to the MDS assessment schedule and related assessment policies as discussed above. A discussion of these comments, along with our responses, appears below.

Comment: One commenter expressed approval of the proposal to incorporate grace days into the existing assessment window. This commenter agrees that this will simplify things and reduce burden, cost and time for providers. Many commenters agreed with using the 5-day assessment to establish per diem payment for the stay. However, several commenters were concerned that the timing of 5-day assessments may still be difficult for SNFs. These commenters stated that securing clinician sign off and all needed information, such as lab results, will be challenging for SNFs. Several commenters requested an allowance for 5-day assessments to be submitted up until Day 14 of a SNF stay.

Response: We appreciate the support for incorporating grace days into the existing assessment window and for using the 5-day assessment to establish per diem payment for the entirety of the stay, assuming that an IPA is not completed. Regarding the timing of the 5-day assessment under the current RUG-IV system, the 5-day assessment window (which goes until Day 8 of a SNF stay) is no different than that proposed under PDPM. FY 2017 MDS data show that almost 98 percent of 5-day assessments were completed timely. This demonstrates that facilities have been able to complete this assessment with minimal difficulty until now and we do not foresee the new system adding an amount of complexity that would prevent them from completing it going forward. Regarding the suggestion to allow providers to have until Day 14 to submit the initial assessment, we do not believe this is necessary or appropriate, given that, as the data above indicate, there is sufficient time for coding the 5-day assessment and because the 5-day assessment provides a snapshot of the resident closer to the point of admission.

Comment: One commenter questioned if on the 5-day Assessment a facility were to establish a RUG in the Ultra High category for a patient, would that RUG be maintained throughout the entire stay regardless of whether there is a drop in the amount of minutes of therapy provided in an assessment window.

Response: We would note that the proposed changes to the assessment schedule would take place upon implementation of PDPM, and under PDPM, patients would no longer be classified into RUG-IV categories. They would instead be classified into case mix groups (CMGs) based on PDPM classification as described in the proposed rule (83 FR 21034-21061). Once a patient is classified into a CMG, that payment group would be maintained through the entire stay unless an Interim Payment Assessment (IPA), as discussed below, is completed and reclassifies the patient into a different CMG.

Comment: Several commenters were concerned with the proposed reduction in payment assessments. They believe that the reduction in assessments could limit the ability of CMS and surveyors to track changes in status and progress of patients and reduce the amount of data CMS has available to use as a basis for future payment adjustments on. These commenters urged CMS to keep the existing PPS assessments as they are. Several commenters recommended that CMS revise the assessment period and ARD to align more closely with other PAC providers in order to implement standardized patient data elements as required by the IMPACT Act.

Response: We appreciate commenters' concern that a reduction in assessments could limit the ability of CMS and surveyors to track status changes and could reduce the amount of data available for use in future payment policy development. However, PDPM relies on stable characteristics that we do not expect to change significantly over the course of the stay. Therefore, additional SNF PPS payment assessments would not necessarily capture different data throughout the stay. Additionally, the OBRA assessment schedule will remain the same and those assessments would provide needed information and data for surveyors and research purposes. Moreover, if clinical characteristics do change, we would expect facilities to elect the option (as discussed further below) to complete the IPA to track these changes.

We appreciate the recommendation to revise the assessment period and ARD to align more closely with other PAC providers in order to implement standardized patient data elements required by the IMPACT Act. We believe that many of the policies being finalized as part of PDPM serve to improve alignment with other PAC settings such as the utilization of functional measures similar to those in IRFs, and the interrupted stay policy which is similar to the IRF and IPF policies, and we hope to continue to improve this alignment in future refinements. As such, we may consider these recommendations in the future.

Comment: Most commenters supported CMS reducing the number of assessments that are required for SNF payment. These commenters expressed that their support for the reduction of the number of payment assessments is due to burden relief and a desire to align with other PAC settings such as IRFs and Home Health that require far fewer patient assessments than SNFs require. One commenter was concerned that while the number of assessments have been reduced, the MDS itself has become more complex with new reporting requirements and items, leaving administrative burden unchanged. Additionally, most commenters conveyed confusion about the proposed IPA. The first area of confusion arose from which criteria CMS wants SNFs to use to determine whether an IPA needs to be completed. Commenters noted that in the proposed rule (83 FR 21063) we stated that there must be a change in the resident's classification in at least one of the first tier classification criteria for any of the components under the proposed PDPM (which are those clinical or nursing payment criteria identified in the first column in Tables 21, 23, 26, and 27), such that the resident would be classified into a classification group for that component that differs from that provided by the 5-day scheduled PPS assessment, and the change in classification group must result in a change in payment either in one particular payment component or in the overall payment for the resident. Additionally, the commenter stated that later in the proposed rule, we clarified that the change in classification group described above refers to not only a change in one of the first tier classification criteria in any of the proposed payment components, but also to one that would be sufficient to change payment in either one component or in the overall payment for the resident (83 FR 21063). Commenters questioned whether an IPA would be required when there is any clinical change that would cause a payment change for a SNF patient. Many commenters requested a general simplification and more guidance surrounding the IPA criterion. Additionally, several commenters believed that there should be guidance about whether an IPA is needed when a patient's functional status and need for specific services changes and whether the IPA should include section GG in order to capture function change.

Most commenters were concerned about the complexity of the proposed IPA. They believed it would create more burden for providers to have to monitor the clinical changes and subsequent payment changes that would trigger the IPA on a daily basis. Several commenters doubted whether the proposed changes would support CMS' Patients over Paperwork initiative and related Medicare Simplifying Document Requirements. One commenter stated that monitoring the first tier changes in each of the case-mix adjusted components would be just as burdensome as the current assessment schedule and is too high a bar, particularly for NTAs. Furthermore, some commenters communicated that the complexities and uncertainties of the IPA would cause providers not to do them and the aim of CMS to provide SNFs with satisfactory reimbursement would not come to fruition. Similarly, some commenters expressed that because of the confusion and burden related to the IPA, this would unnecessarily increase the risk of provider error and potential medical review. This, in turn, would cause facilities to complete fewer IPAs and consequently this could lead to less quality care provided to patients who otherwise would have needed it had it been identified appropriately using the IPA. Some commenters are concerned that the IPA will likely require MDS coordinators to take on more of a care coordination role which would require additional operating costs for SNFs.

Response: We are pleased that so many commenters support the proposal to reduce the number of payment assessments in SNFs. We agree that alignment across PAC settings is very important and anticipate that the reduction of assessments will further this alignment. We also agree that the reduction of assessments will significantly decrease the burden for providers.

We disagree with the commenter that stated that even though the number of assessments have been reduced, the MDS itself has become more complex with new reporting requirements and items, leaving administrative burden unchanged. Section VII. of this final rule discusses burden associated with the changes we are making and our calculations show that there is a significant reduction in administrative burden to providers under PDPM.

We thank the commenters for calling our attention to their questions and confusion about the IPA. We continue to believe that it is necessary for SNFs to continually monitor the clinical status of each and every patient in the facility regularly regardless of payment or assessment requirements and we believe that there should be a mechanism in place that would allow facilities to do this. However, we also believe that providers may be best situated, as in the case of the Significant Change in Status Assessment, to determine when a change has occurred that should be reported through the IPA. Therefore, to further ease the administrative burden associated with PDPM and improve clarity on when an IPA should be completed, we have decided to make the IPA an optional assessment. Facilities will be able to determine when IPAs will be completed for their patients to address potential changes is clinical status and what criteria should be used to decide when an IPA would be necessary. We are not finalizing the proposed criteria for the triggering of the IPA, but rather we will seek additional stakeholder input on this issue. We note that we are finalizing the proposal surrounding IPA completions and the variable per diem adjustment schedule (including the NTA variable per diem, that is, the completion of an IPA will not reset the variable per diem adjustment schedule)) even though the IPA will now be optional. However, because the IPA will be optional and providers can determine their own criteria for when an IPA is completed, we are revising the ARD criteria we proposed. The ARD for the IPA will be the date the facility chooses to complete the assessment relative to the triggering event that causes a facility to choose to complete the IPA. Payment based on the IPA will begin the same day as the ARD. The IPA will not be susceptible to assessment penalties, given the optional nature of the assessment. We reiterate that we expect facilities to complete IPAs as they deem necessary to address clinical changes throughout a SNF stay and that the removal of the requirement to complete these assessments does not in any way negate the need to provide excellent skilled nursing and rehabilitative care and continually monitor and document patient status.

Comment: Many comments addressed the IPA criteria that “. . . the resident would not be expected to return to his or her original clinical status within a 14-day period.” Commenters stated that this is a very subjective determination and that it is difficult for providers to predict the course of recovery for patients who have an acute clinical change and providers would not necessarily know if this episode would or would not resolve in a 2-week period. On the other hand, several other commenters expressed that the 14-day period seemed excessive since the average of most SNF stays is currently around 19 or 20 days and CMS estimates the majority of the stays under PDPM will be between 1-15 days. Some commenters recommend that CMS shorten the timeframe to 3 days consistent with the proposed interrupted stay policy. Other commenters suggested that this time period should be reduced to 7 days. One commenter recommended that CMS should use an approach similar to the change in status policy in the home health setting (https://www.cms.gov/​medicare/​quality-initiatives-patient-assessment-instruments/​oasis/​downloads/​qandadocument0909.pdf). Many comments requested more examples that show various scenarios in which an IPA would be required. Additionally, several commenters requested CMS to describe when an IPA would be used versus a Significant Change in Status Assessment.

Response: Given that the IPA will now be an optional assessment under PDPM and we are not finalizing criteria for when an IPA should be completed, but rather, will seek additional stakeholder input on this issue, we will take all of the comments regarding these criteria under consideration for future policy making.

Comment: One commenter requested clarification on the nursing classification change that would trigger an IPA. This commenter questioned what constitutes a substantial change versus a titration of services. Another commenter requested clarification as to what would constitute a first tier change in the nursing component. Another commenter requested clarification on what a first tier change in the SLP Case-Mix classification would look like.

Response: We appreciate the requests for clarification on the IPA triggers. However, because the IPA will now be an optional assessment, we will allow facilities to determine on their own when IPAs are necessary. As such, we will seek additional stakeholder feedback on this issue in the future.

Comment: Many commenters supported the addition of the IPA; with several commenters supporting the variable per diem adjustment policy relating to the IPA, stating that this would reduce the incentive for providers to complete multiple IPAs over the course of a SNF stay each time payment was reduced based on the adjustment. Some commenters disagreed and stated that the variable per diem for the NTA component should be reset following the completion of an IPA, while other commenters supported the variable per diem adjustment but had concerns about the NTA per diem rate following the completion of an IPA. These commenters suggested the variable per diem rate be reset for NTA services when an IPA is completed. Some commenters stated they recognize CMS' concern that providers might be incentivized to complete multiple IPAs in order to reset the NTA rate during one SNF stay. However, these commenters were concerned that in cases where IPAs are legitimately completed and the result is a change in NTA use, the potential financial loss could be significant or could result in re-hospitalization if facilities do not end up providing NTAs that patients need because of financial considerations. Commenters offered several solutions to this concern. One commenter suggested that the NTA variable per diem adjustment schedule be reset for patients who experience adverse changes in status resulting in the completion of an IPA. Another commenter suggested that CMS use the points associated with NTAs to develop a threshold of additional NTA points that would allow facilities to reset the NTA variable per diem rate to Day 1. One commenter suggested a physician verified post-stay process for patients to dispute the variable per diem adjustment when their need for PT, OT, or NTAs would substantially increase from what was originally anticipated.

Response: We are pleased that so many commenters supported the addition of the IPA, and appreciate the support for not resetting the variable per diem adjustment when an IPA is completed. We disagree with those commenters who suggested that the variable per diem be reset every time an IPA is completed. As stated in the proposed rule (83 FR 21060), in examining costs over a stay, we found that for certain categories of SNF services, notably PT, OT and NTA services, costs declined over the course of a stay. Our analyses showed that, on average, the number of therapy minutes is lower for assessments conducted later in the stay.

We believe that the ability to reset the variable per diem would incentivize providers to complete IPAs every time the variable per diem was reduced. We also believe it is possible that providers may refrain from coding certain conditions on an initial assessment and then code other conditions on later assessments to justify the variable per diem adjustment reset.

With regard to the ideas presented by commenters for when the variable per diem should be reset, we do not believe that the variable per diem should be reset except in cases of an entirely new SNF stay (we also refer readers to section V.F. of this final rule for a discussion of our interrupted stay policy). We understand that some commenters are concerned that unless the variable per diem adjustment schedule is reset, a patient's per diem rate may not reflect changes in NTA use identified in an IPA that is completed during a patient's stay. However, we note that if a new condition is coded on an IPA during a SNF stay, the SNF PPS per diem payment for the patient may in fact increase to reflect changes in the patient's clinical condition if the new condition results in a change to the patient's case-mix group. Thus, a patient's case-mix group and associated payment could change within a stay to reflect a change in NTA use on the IPA. However, we do not think that resetting the variable per diem adjustment would be appropriate each time such a change occurs. As we explained above, we found that for PT, OT, and NTA services, costs generally decline over the course of a stay and we believe the variable per diem adjustment appropriately accounts for this decline in costs. Furthermore, as the SNF PPS is a prospective payment system, it is not intended to reimburse for each additional condition or service separately, but rather provides a predictive payment based on a snapshot of the patient's condition. Resetting the variable per diem adjustment in each case of a change in the patient's condition would be more akin to a traditional fee-for-service model, providing additional payment for each additional service or condition, which is precisely the opposite of the goals of implementing PDPM.

Commenters were also concerned that there might be financial implications to not re-setting the variable per diem for NTAs and that this might result in facilities not providing the drugs that patients require because of financial reasons. However, we do not believe that the variable per diem adjustment creates new financial implications that would affect patient care, as this incentive also exists under the current payment system that utilizes a constant per diem rate and we have no evidence that SNF patients are being denied necessary medications or services. Further, we would note that there are quality safeguards in place such as readmission penalties and quality metrics such as the SNF QRP quality measures that should provide a disincentive against providers engaging in this type of stinting behavior.

Comment: Several commenters requested that CMS consider adding additional assessments to capture changes in patient need during the SNF stay. These commenters explained their concern that PDPM does not differentiate between processes designed to adjust payment and the continuous need to assess patient care needs. Additionally, these commenters believe that status changes-especially of the functional nature-that may not rise to the level of a required IPA might be missed, especially in longer stay patients. These commenters stated that therapy assessments may not be documented frequently enough to capture serious status changes of patients under PDPM. Specifically, they noted that patient care needs must be documented through an additional assessment after day 20 and they are apprehensive that the change in the variable per diem payment after Day 20 of a SNF stay may directly affect patient care if these assessments are not completed. These commenters suggested that CMS add an additional assessment after Day 20 of the stay that would specifically capture therapy needs.

Response: We appreciate the commenters' concerns regarding how assessments relate to functional change, the ongoing need to assess patient care needs, and the necessity to capture therapy needs throughout the stay, especially during long stays. It is our expectation that the optional nature of the IPA will allow facilities to capture all of these changes as they occur during a SNF stay. Facilities will determine when IPAs should be completed, and we expect them to pay special attention to clinical and functional changes. It should be noted that, even absent an IPA requirement, we expect SNFs to constantly evaluate, capture, document and treat clinical and functional changes that occur in patients throughout a SNF stay. We defer to the judgment of clinicians and expect that the care they are providing is always evaluative in nature, meaning that therapists are continually assessing the needs of the patient and changing interventions as needed throughout the course of the therapy regimen, and we note that the absence or presence of a required assessment tool should not change this.

Accordingly, after considering the comments received, for the reasons discussed in the proposed rule and in this final rule, we are finalizing our proposed changes to the MDS assessment schedule and related assessment policies as discussed in the proposed rule, with the following modifications. As discussed above, rather than making the IPA a required assessment as we proposed, this assessment will be optional, and providers may determine whether and when an IPA is completed. In addition, because the IPA is an optional assessment and providers can determine their own criteria for when an IPA is completed, we are revising the ARD criteria such that the ARD will be the date the facility chooses to complete the IPA relative to the triggering event that causes the facility to choose to complete the IPA. Payment based on the IPA would begin the same day as the ARD. These changes will be effective October 1, 2019 in conjunction with the implementation of the PDPM.

2. Item Additions to the Swing Bed PPS Assessment

As noted previously in section IV.C of this final rule, section 1883 of the Act permits certain small, rural hospitals to enter into a Medicare swing-bed agreement, under which the hospital can use its beds to provide either acute or SNF care, as needed. For critical access hospitals (CAHs), Part A pays on a reasonable cost basis for SNF services furnished under a swing-bed agreement. However, in accordance with section 1888(e)(7) of the Act, such services furnished by non-CAH rural hospitals are paid under the SNF PPS, effective with cost reporting periods beginning on or after July 1, 2002. A more detailed discussion of this provision appears in section IV.C of the proposed rule.

For purposes of the proposed PDPM, we proposed to add three items to the Swing Bed PPS Assessment. Until now, these additional items have not been part of the Swing Bed PPS Assessment form because they have not been used for payment. However, we stated in the proposed rule (83 FR 21064) that presence of each of these items would be used to classify swing bed residents under the proposed SNF PDPM as explained in section V.D. of the proposed rule. Thus, we stated that believed it was necessary and appropriate to include these items in the Swing Bed PPS Assessment beginning October 1, 2019, in conjunction with the proposed implementation of the PDPM. The items we proposed to add to the Swing Bed PPS assessment are provided in Table 34 of the proposed rule (also set forth in Table 34).

Commenters submitted the following comments related to the proposed addition of three items to the Swing Bed PPS assessment. A discussion of these comments, along with our responses, appears below.

Comment: Commenters supported the addition of the three proposed items to the Swing Bed PPS assessment and stated that these items will be important to establish the SLP and NTA component case-mix rates.

Response: We are pleased that commenters support the addition of these items to the Swing Bed PPS Assessment. We agree that these items are necessary to determine the SLP and NTA case-mix rates. We will continue to consider additions to the Swing Bed PPS Assessment as it becomes necessary to ensure consistency between swing bed and non-swing bed providers.

Accordingly, after considering the comments received, for the reasons discussed in the proposed rule and in this final rule, we are finalizing the addition of the items in Table 34 to the Swing Bed PPS Assessment as proposed without modification, effective October 1, 2019 in conjunction with the implementation of the PDPM.

Table 34—Items To Add to Swing Bed PPS Assessment

MDS Item No.	Item name	Related PDPM payment component
K0100	Swallowing Disorder	SLP
I4300	Active Diagnoses: Aphasia	SLP
O0100D2	Special Treatments, Procedures and Programs: Suctioning, While a Resident	NTA

3. Items To Be Added to the PPS Discharge Assessment

Under the MDS 3.0, the Part A PPS Discharge assessment is completed when a resident's Medicare Part A stay ends, but the resident remains in the facility (MDS 3.0 RAI Manual Chapter 2.5). The PPS Discharge Assessment uses the Item Set NPE and does not currently contain section O of the MDS 3.0. The therapy items in section O of the MDS allow CMS to collect data from providers on the volume, type (physical therapy, occupational therapy and speech-language pathology), and mode (individual, concurrent, or group therapy) of the therapy provided to SNF residents. As noted in comments received on the ANPRM in relation to therapy provision, this data would be particularly important to monitor. Specifically, a significant number of commenters expressed concerns that the amount of therapy provided to SNF residents, were RCS-I to have been implemented, would drop considerably as compared to the amount currently delivered under RUG-IV. Commenters noted that this is because the incentive to provide a high volume of therapy services to SNF residents (to achieve the highest resident therapy group classification) would no longer exist under RCS-I, potentially leading providers to reduce significantly the amount of therapy provided to SNF residents.

We stated in the proposed rule (83 FR 21065) that, given that the RCS-I model and PDPM both present the potential for providers to reduce significantly the amount of therapy provided to SNF residents as compared to RUG-IV, we believe that the same potential result may occur under the proposed PDPM as commenters identified with RCS-I. To better track therapy utilization under PDPM, and to better ensure that residents continue to receive an appropriate amount of therapy commensurate with their needs, given the reduction in the frequency of resident assessments required under the proposed PDPM, we proposed to add therapy collection items to the PPS Discharge assessment and to require providers to complete these items beginning October 1, 2019, in conjunction with the proposed implementation of the PDPM.

Specifically, we proposed to add the items listed in Table 35 of the proposed rule (as set forth in Table 35 of this final rule) to the PPS Discharge Assessment.

Table 35—Items To Add to SNF PPS Discharge Assessment

MDS Item No.	Item name
O0400A5	Special Treatments, Procedures and Programs: Speech-Language Pathology and Audiology Services: Therapy Start Date.
O0400A6	Special Treatments, Procedures and Programs: Speech-Language Pathology and Audiology Services: Therapy End Date.
O0400A7	Special Treatments, Procedures and Programs: Speech-Language Pathology and Audiology Services: Total Individual Minutes.
O0400A8	Special Treatments, Procedures and Programs: Speech-Language Pathology and Audiology Services: Total Concurrent Minutes.
O0400A9	Special Treatments, Procedures and Programs: Speech-Language Pathology and Audiology Services: Total Group Minutes.
O0400A10	Special Treatments, Procedures and Programs: Speech-Language Pathology and Audiology Services: Total Days.
O0400B5	Special Treatments, Procedures and Programs: Occupational Therapy: Therapy Start Date.
O0400B6	Special Treatments, Procedures and Programs: Occupational Therapy: Therapy End Date.
O0400B7	Special Treatments, Procedures and Programs: Occupational Therapy: Total Individual Minutes.
O0400B8	Special Treatments, Procedures and Programs: Occupational Therapy: Total Concurrent Minutes.
O0400B9	Special Treatments, Procedures and Programs: Occupational Therapy: Total Group Minutes.
O0400B10	Special Treatments, Procedures and Programs: Occupational Therapy: Total Days.
O0400C5	Special Treatments, Procedures and Programs: Physical Therapy: Therapy Start Date.
O0400C6	Special Treatments, Procedures and Programs: Physical Therapy: Therapy End Date.
O0400C7	Special Treatments, Procedures and Programs: Physical Therapy: Total Individual Minutes.
O0400C8	Special Treatments, Procedures and Programs: Physical Therapy: Total Concurrent Minutes.
O0400C9	Special Treatments, Procedures and Programs: Physical Therapy: Total Group Minutes.
O0400C10	Special Treatments, Procedures and Programs: Physical Therapy: Total Days.

We stated that for the proposed items, which refer to the total number of minutes for each therapy discipline and each therapy mode, this would allow CMS both to conduct reviews of changes in the volume and intensity of therapy services provided to SNF residents under the proposed PDPM compared to that provided under RUG-IV, as well as to assess compliance with the proposed group and concurrent therapy limit discussed in section V.F of the FY 2019 SNF PPS proposed rule. We further stated that the proposed “total days” items for each discipline and mode of therapy would further support our monitoring efforts for therapy, as requested by commenters on the ANPRM, by allowing us to monitor not just the total minutes of therapy provided to SNF residents under the proposed PDPM, but also assess the daily intensity of therapy provided to SNF residents under the proposed PDPM, as compared to that provided under RUG-IV. As we explained in the proposed rule, ultimately, these proposed items would allow facilities to easily report therapy minutes provided to SNF residents and allow us to monitor the volume and intensity of therapy services provided to SNF residents under the proposed PDPM, as suggested by commenters on the ANPRM. We stated that if we discovered that the amount of therapy provided to SNF residents did change significantly under the proposed PDPM, if implemented, then we would assess the need for additional policies to ensure that SNF residents continued to receive sufficient and appropriate therapy services consistent with their unique needs and goals.

Commenters submitted the following comments related to the proposed rule's discussion of the SNF PPS Discharge Assessment. A discussion of these comments, along with our responses, appears below.

Comment: Several commenters opposed the addition of items and the reporting of therapy services in section O of the SNF PPS Discharge Assessment. These commenters acknowledged that the fundamental design of PDPM (which will no longer tie payment to the amount of therapy a patient receives, as occurs under the current RUG-IV payment system) could perhaps cause some patients appropriately to receive less therapy. The commenters stated that this would be a positive anticipated outcome for many residents considering that the recurring concern of RUG-IV has been that the model may incentivize SNFs to provide therapy services beyond what patients need. These commenters noted that while they recognize the importance of monitoring the impacts of policy changes especially in the initial stages of the implementation, they were disappointed that CMS appears to be interested in collecting this data merely in order to monitor changes in volume of services and that CMS did not discuss evaluating this aspect of PDPM in relation to quality and outcomes measures (such as through the SNF Quality Reporting Program) that are normally associated with effective therapy provision. These commenters noted that the MDS should be used for care-planning and case-mix payment determination and that since therapy time is not relevant to the case-mix methodology under PDPM, this proposed addition of therapy collection of items serves no purpose on the MDS. These commenters suggested that instead of collecting therapy provision information on the MDS, facilities should gather and report therapy provision information on claims on a line-item, date-of-service basis that would be in line with Medicare Part B and other payers and limit provider burden.

Response: We agree with commenters that it is possible that, in some cases, less therapy will be provided under PDPM than under RUG-IV and that this would be a positive development in those cases where therapy was provided regardless of patient need and simply because of higher payments for higher volumes of therapy. However, we continue to be concerned that under PDPM, providers may reduce the amount of therapy provided to SNF patients because of financial considerations. We agree with commenters that quality and outcomes measures (like those in the SNF Quality Reporting Program) would be a positive way to evaluate the efficacy of therapy provision, and we will take this into consideration for future policy development. However, we disagree that the collection of these items is not relevant to case-mix determination. While the days and minutes of therapy provided will not be a determining factor in the therapy case-mix classification under PDPM, the need to ensure beneficiary protection under this payment system is very relevant to the therapy case-mix classification, and the ability to collect this data will safeguard the integrity of the case-mix classification and help ensure that patients receive an appropriate amount of therapy services. Should we discover that the amount of therapy under PDPM is distinctly different from the amount of therapy under RUG-IV, we will evaluate the potential reasons for this change and consider potential actions, either at the provider or systemic level, to address these issues.

We appreciate the commenters' suggestion of using claims information as the basis for therapy reporting, but would note that this mechanism would be more complicated and not provide the same level of detail in the data as is currently reported in section O of the MDS. Further, as providers are already familiar with the section O items, we believe that this method will provide the simplest transition for providers.

Comment: Many commenters supported the proposal to add therapy collection items to the SNF PPS Discharge Assessment in order to monitor compliance with the group and concurrent therapy limits. One commenter stated that they believed this proposal may protect against therapists being pressured to provide an unreasonable amount of group or concurrent therapy. Several commenters, however, were concerned that the monitoring effort proposed is not strong enough to enforce the aforementioned limits. One commenter suggested that based on CMS' assertion that “services furnished to SNF residents may be considered reasonable and necessary insomuch as services are consistent with `the individual's particular medical needs,' ” (83 FR 21068) they question whether excessive group and concurrent therapy serves as justification to deny SNF coverage. This commenter proposed that rather than a “warning edit” that would notify providers that they have exceeded the group and concurrent threshold, CMS should decide whether these occurrences violate coverage requirements and if it is determined that they do, payment should be denied for the claim. Many commenters suggested that in addition to monitoring the therapy provision, CMS should monitor resident outcomes. One commenter recommended that CMS utilize the four new SNF QRP section GG outcome measures, and current readmission measures and qualitatively measure the current the effectiveness of therapy provided in the SNF.

Response: We appreciate the comments we received in support of the proposal to add therapy collection items to the SNF PPS Discharge Assessment. We agree that this proposal would enable us to monitor group and concurrent therapy compliance and will hopefully help prevent therapists from feeling pressured to provide an unreasonable amount of group and/or concurrent therapy. We appreciate the concern that the monitoring effort proposed is not strict enough to enforce the concurrent and group therapy limits. We would note that the monitoring plan is intended for this exact reason. As stated in the proposed rule (83 CFR 21067), as part of our regular monitoring efforts on SNF Part A services, we would monitor group and concurrent therapy utilization under the proposed PDPM and consider making future proposals to address abuses of this proposed policy or flag providers for additional review should an individual provider be found to consistently exceed the proposed threshold after the implementation of the proposed PDPM.

We appreciate the suggestion to deny claims if the threshold is exceeded and we may consider this option further in the future. As stated in the FY 2019 SNF PPS proposed rule (83 FR 21068), services furnished to SNF residents may be considered reasonable and necessary insomuch as the services are consistent with the individual's particular medical needs and that excessive levels of group and/or concurrent therapy could constitute a reason to deny SNF coverage for such stays. We appreciate the suggestion to monitor patient outcomes in addition to collecting therapy provision data, as well as the recommendation to specifically use the four new SNF QRP section GG outcome measures and current readmission measures to measure the effectiveness of therapy provided in SNFs. We may consider these suggestions in future policy making decisions.

Accordingly, after considering the comments received, for the reasons discussed in the proposed rule and in this final rule, we are finalizing the addition of the items in Table 35 to the PPS Discharge Assessment as proposed, without modification, effective October 1, 2019 in conjunction with the implementation of the PDPM.

E. Revisions to Therapy Provision Policies Under the SNF PPS

Currently, almost 90 percent of residents in a Medicare Part A SNF stay receive therapy services. Under the current RUG-IV model, therapy services are case mix-adjusted primarily based on the therapy minutes reported on the MDS. As discussed in the proposed rule (83 FR 21065), when the original SNF PPS model was developed, most therapy services were furnished on an individual basis, and the minutes reported on the MDS served as a proxy for the staff resource time needed to provide the therapy care. Over the years, we have monitored provider behavior and have made policy changes as it became apparent that, absent safeguards like quality measurement to ensure that the amount of therapy provided did not exceed the resident's actual needs, there were certain inherent incentives for providers to furnish as much therapy as possible. Thus, for example, in the SNF PPS FY 2010 final rule (74 FR 40315 through 40319), we decided to allocate concurrent therapy minutes for purposes of establishing the RUG-IV group to which the patient belongs, and to limit concurrent therapy to two patients at a time who were performing different activities.

As we explained in the proposed rule (83 FR 21066), following the decision to allocate concurrent therapy, using STRIVE data as a baseline, we found two significant provider behavior changes with regard to therapy provision under the RUG-IV payment system. First, there was a significant decrease in the amount of concurrent therapy that was provided in SNFs. Simultaneously, we observed a significant increase in the provision of group therapy, which was not subject to allocation at that time. We concluded that the manner in which group therapy minutes were counted in determining a patient's RUG-IV group created a payment incentive to provide group therapy rather than individual therapy or concurrent therapy, even in cases where individual therapy (or concurrent therapy) was more appropriate for the resident. Thus, we stated that we made two policy changes regarding group therapy in the FY 2012 SNF PPS final rule (76 FR 48511 through 48517). We defined group therapy as exactly four residents who are performing the same or similar therapy activities. Additionally, we allocated group therapy among the four patients participating in group therapy—meaning that the total amount of time that a therapist spent with a group will be divided by 4 (the number of patients that comprise a group) to establish the RUG-IV group to which the patient belongs.

We stated in the proposed rule (83 FR 21066) that since we began allocating group therapy and concurrent therapy, these modes of therapy (group and concurrent) represent less than one percent of total therapy provided to SNF residents. Table 36, which appeared in the FY 2014 SNF PPS Proposed Rule (78 FR 26464) (and was also presented in the FY 2019 SNF PPS proposed rule) and sets forth our findings with respect to the effect of policies finalized in the FY 2012 SNF PPS Final Rule, demonstrates the change in therapy provision between the STRIVE study and the implementation of the therapy policy changes in FY 2012. As we noted in the proposed rule, the distribution of therapy modes presented in Table 36 reflecting therapy provision in FY 2012 is also an accurate reflection of current therapy provision based on resident data collected in the QIES Database and continued monitoring of therapy utilization.

Table 36—Mode of Therapy Provision

	STRIVE (%)	FY 2011 (%)	FY 2012 (%)
Individual	74	91.8	99.5
Concurrent	25	0.8	0.4
Group	<1	7.4	0.1

As we explained in the proposed rule (83 FR 21066), based on our prior experience with the provision of concurrent and group therapy in SNFs, we again were concerned that if we were to implement the proposed SNF PDPM, providers may base decisions regarding the particular mode of therapy to use for a given resident on financial considerations rather than on the clinical needs of SNF residents. We stated that because the proposed SNF PDPM would not use the minutes of therapy provided to a resident to classify the resident for payment purposes, we were concerned that SNFs may once again become incentivized to emphasize group and concurrent therapy, over the kind of individualized therapy which is tailored to address each beneficiary's specific care needs which we believe is generally the most appropriate mode of therapy for SNF residents. As we stated in the FY 2012 proposed rule (76 FR 26387), while group therapy can play an important role in SNF patient care, group therapy is not appropriate for either all patients or for all conditions, and is primarily effective as a supplement to individual therapy, which we maintain should be considered the primary therapy mode and standard of care in therapy services provided to SNF residents. We stated in the FY 2012 proposed rule that, as evidenced by the application of a cap on the amount of group therapy services that may be provided to SNF residents, we do not believe that a SNF providing the preponderance of therapy in the form of group therapy would be demonstrating the intensity of therapy appropriate to this most frail and vulnerable nursing home population.

We stated in the FY 2019 SNF PPS proposed rule (83 FR 21066) that since the inception of the SNF PPS, we have limited the amount of group therapy provided to each SNF Part A resident to 25 percent of the therapy provided to them by discipline. We referred to the FY 2000 final rule (64 FR 41662), where we stated that although we recognize that receiving PT, OT, or ST as part of a group has clinical merit in select situations, we do not believe that services received within a group setting should account for more than 25 percent of the Medicare resident's therapy regimen during the SNF stay.

We explained that although we recognize that group and concurrent therapy may have clinical merit in specific situations, we also continue to believe that individual therapy is generally the best way of providing therapy to a resident because it is most tailored to that specific resident's care needs. As such, we stated that individual therapy should represent the majority of the therapy services received by SNF residents both from a clinical and payment perspective. As we stated in the FY 2012 proposed rule (76 CFR 26372), even under the previous RUG-53 model, it was clear that the predominant mode of therapy that the payment rates were designed to address was individual therapy rather than concurrent or group therapy.

We stated in the proposed rule (83 FR 21066) that to help ensure that SNF residents would receive the majority of therapy services on an individual basis, if we were to implement the proposed PDPM, we believed concurrent and group therapy combined should be limited to no more than 25 percent of a SNF resident's therapy minutes by discipline. In combination, this limit would ensure that at least 75 percent of a resident's therapy minutes are provided on an individual basis. We stated that because the change in how therapy services would be used to classify residents under the proposed PDPM gives rise to the concern that providers may begin to utilize more group and concurrent therapy due to financial considerations, we proposed to set a combined 25 percent limit on concurrent therapy and group therapy for each discipline of therapy provided. For example, if a resident received 800 minutes of physical therapy, no more than 200 minutes of this therapy could be provided on a concurrent or group basis. Finally, we noted that under RUG-IV, we currently allocate minutes of therapy because we pay for therapy based on therapy minutes and not resident characteristics. We stated that given that therapy minutes would no longer be a factor in determining payment classifications for residents under the proposed PDPM, we would utilize the total, unallocated number of minutes by therapy mode reported on the MDS, to determine compliance with the proposed limit. We explained that utilizing unallocated therapy minutes also serves to underscore the patient-driven nature of the PDPM, as it focuses the proposed limit on concurrent and group therapy on the way in which the therapy is received by the beneficiary, rather than furnished by the therapist, and would better ensure that individual therapy represents at least a vast majority of the therapy services received by a resident.

In the proposed rule (83 FR 21067), we considered other possible limits, and even no limit, on group and concurrent therapy. For example, we considered placing no limit on group or concurrent therapy, in order to afford providers the greatest degree of flexibility in designing a therapy program for each SNF resident. However, even in response to this option to have no limit on concurrent and group therapy, many commenters on the ANPRM expressed concerns regarding the lack of appropriate safeguards for ensuring that SNF residents continue to receive an appropriate level of therapy under the revised case-mix model. We stated in the proposed rule that we agree with these commenters and believe that there should be some limit on the amount of group and concurrent therapy that is provided to residents in order to ensure that residents receive an appropriate amount of individual therapy that is tailored to their specific needs. Also, in the ANPRM, we discussed the possibility of proposing a 25 percent limit on each of concurrent and group therapy, allowing for up to 50 percent of therapy services provided in the SNF to be provided in a non-individual modality. We stated in the proposed rule that this option sought to balance the flexibility afforded to therapists in designing an appropriate therapy plan that meets the needs and goals of the specific resident with the importance of ensuring that SNF residents receive an appropriate level of individual therapy. However, we were concerned that a separate 25 percent limit for group and concurrent therapy would not provide sufficient assurance that at least a majority of a resident's therapy would be provided on an individual basis. Therefore, we stated that we believe the separate 25 percent limits on concurrent and group therapy discussed in the ANPRM, or any option which would impose a higher limit on group and concurrent therapy, would not provide the necessary protection for SNF residents. By contrast, we stated that we believe a combined 25 percent limit on group and concurrent therapy would provide sufficient assurance that at least a majority of each resident's therapy would be provided on an individual basis, consistent with our position that individual therapy is generally the best way of providing therapy to SNF residents because it is most tailored to their care needs. We noted that, assuming that existing therapy delivery patterns (as set forth in Table 36) are accurate and they reflect the individually-tailored needs of SNF residents currently being treated under the SNF benefit, the number of group and concurrent minutes that have been reported by SNFs thus far are significantly lower than the limit described in our proposal. In other words, we stated that, based on the data presented in Table 36, the proposed limit on group and concurrent therapy affords a significantly greater degree of flexibility on therapy modality than appears to be required to meet the needs of SNF residents, given that less than one percent of therapy currently being delivered is either group or concurrent therapy. Therefore, we concluded that a combined limit of 25 percent for group and concurrent therapy should provide SNFs with more than enough flexibility with respect to therapy mode to meet the care needs of their residents.

As discussed in the proposed rule (83 FR 21067), we believe that individual therapy is usually the best mode of therapy provision as it permits the greatest degree of interaction between the resident and therapist, and should therefore represent, at a minimum, the majority of therapy provided to an SNF resident. However, we recognized that, in very specific clinical situations, group or concurrent therapy may be the more appropriate mode of therapy provision, and therefore, we stated we would want to allow providers the flexibility to be able to utilize these modes. We continued to stress that group and concurrent therapy should not be utilized to satisfy therapist or resident schedules, and that all group and concurrent therapy should be well documented in a specific way to demonstrate why they are the most appropriate mode for the resident and reasonable and necessary for his or her individual condition.

Currently the RUG-IV grouper calculates the percentage of group therapy each resident receives in the SNF based on the algorithms described in section 6.6 of the MDS RAI Manual (found at https://downloads.cms.gov/​files/​MDS-30-RAI-Manual-v115-October-2017.pdf). When a resident is found to have exceeded the 25 percent group therapy limit, the minutes of therapy received in excess are not counted towards the calculation of the RUG-IV therapy classification. We explained that because the proposed PDPM would not use the minutes of therapy provided to a resident to classify the resident for payment purposes, we would need to determine a way under the proposed PDPM to address situations in which facilities exceed the combined 25 percent group and concurrent therapy limit.

Therefore, we proposed that at a component level (PT, OT, SLP), when the amount of group and concurrent therapy exceeds 25 percent within a given therapy discipline, that providers would receive a non-fatal warning edit on the validation report that the provider receives when submitting an assessment which would alert the provider that the therapy provided to that resident exceeded the threshold. To explain, a fatal error in the QIES ASAP system occurs when one or more items in the submitted record fail to pass the requirements identified in the MDS data submission specifications. A warning error occurs when an item or combination of items in the submitted record trigger a non-fatal edit in the QIES ASAP system. We stated that the non-fatal warning would serve as a reminder to the facility that they are out of compliance with the proposed limit for group and concurrent therapy. We also stated that, as part of our regular monitoring efforts on SNF Part A services, we would monitor group and concurrent therapy utilization under the proposed PDPM and consider making future proposals to address abuses of this proposed policy or flag providers for additional review should an individual provider consistently be found to exceed the proposed threshold after the implementation of the proposed PDPM. We noted that as the proportion of group and/or concurrent therapy (which are, by definition, non-individual modes of therapy provision) increases, the chances that the provider is still meeting the individualized needs of each resident would diminish. We stated that given that meeting the individualized needs of the resident is a component of meeting the coverage requirements for SNF Part A services, as described in section 1814(a)(2)(B) of the Act and further described in section 30 of Chapter 8 of the Medicare Benefit Policy Manual (accessible at https://www.cms.gov/​Regulations-and-Guidance/​Guidance/​Manuals/​downloads/​bp102c08.pdf), where it states that services furnished to SNF residents may be considered reasonable and necessary insomuch as the services are consistent with “the individual's particular medical needs”, excessive levels of group and/or concurrent therapy could constitute a reason to deny SNF coverage for such stays. We invited comments on this proposed compliance mechanism.

Commenters submitted the following comments related to the proposed revisions to the therapy policies under the PDPM. A discussion of these comments, along with our responses, appears below.

Comment: The vast majority of commenters supported the proposal to limit concurrent and group therapy to 25 percent. Several stated that the combined limit is not restrictive enough and recommended that CMS implement further real-time efforts (beyond the warning edit outlined in the proposed rule) to ensure that patients are receiving the therapy they need, monitor compliance, and have stricter enforcement outcomes. Conversely, several commenters supported the notion that CMS could possibly raise the group and concurrent therapy limit following close monitoring of utilization and determining that patients are indeed receiving the individualized therapy they need even in group and concurrent sessions, and that SNFs are not taking advantage of the financial incentives that providing group and concurrent therapy offer. These commenters stated that they were in favor of the idea that providers would be reporting and counting the patients' time in therapy rather than the therapists' allocated time to determine compliance with the proposed group and concurrent therapy limit under PDPM since this is more consistent with the concept of patient-centered care and best clinical practice.

Response: We are pleased that the vast majority of commenters supported the proposal to limit concurrent and group therapy to 25 percent. We appreciate both the concern that 25 percent may not be restrictive enough and the concern that it is too restrictive, and we will continue to track the amount of therapy provided via the different modes in conjunction with our monitoring efforts described throughout section V. of this final rule. We will determine whether group and concurrent therapy are being over or underutilized and we will consider revising the policy and enforcement efforts as necessary. Because therapy minutes would no longer be a factor in determining case-mix classification under the PDPM, as it is under RUG-IV, we agree with the commenters that using the total, unallocated number of minutes by therapy mode reported on the MDS versus therapists' allocated time makes the most sense in determining compliance with the group and concurrent therapy limit, and we appreciate that the commenters recognized the patient-centered nature of the proposal.

Comment: Several commenters stated that the current policy regarding group and concurrent therapy allocation has increased provider costs. Specifically, these comments stated that concurrent and group therapy are more cost-effective modes than individual therapy and that the 25 percent drop in the delivery of concurrent and group therapy from FY 2011 until now demonstrates a significant increase in provider costs. These commenters believe that restoring flexibility in therapy service under PDPM will permit SNFs to develop more cost-effective innovative approaches to care.

Response: We disagree with the assertion that the current policy to allocate group and concurrent therapy increases cost. As we stated in the FY 2012 final rule (76 FR 48515), to fulfill our responsibilities to ensure appropriate payment based on resource utilization and cost, we proposed the allocation of group therapy minutes, which equalizes the reimbursement incentives across modes of therapy. Although case-mix classification under PDPM is not based primarily on volume of services provided, as is the case with the RUG-IV payment system, it is still important that there are equal financial incentives to provide the different modes of therapy. Further, given that the payment incentives are equal among the various therapy modes because of the allocation of minutes and that over 99 percent of therapy minutes are reported as individual therapy, this provides evidence that the mode of therapy that providers believe is most effective in addressing a beneficiary's needs is individual therapy. Regarding the need to restore flexibility in therapy service under PDPM, we think that the 25 percent cap will allow for flexibility in therapy services. As mentioned above, since currently, over 99 percent of therapy minutes are delivered individually, SNFs should continue to have adequate leeway to provide the mode of therapy which is most appropriate for the patients even with the revised cap. Nevertheless, to the extent that provider costs have increased, these cost increases have been captured as part of the data analysis used to set the case-mix weights under PDPM. To the extent that these costs change as a result of PDPM, more specifically changes in the mode of therapy service delivery, we can consider revising the case-mix weights to reflect these changes in provider costs.

Comment: Several commenters opposed the proposed limitations on group and concurrent therapy and expressed concern that even though there is a lack of data demonstrating what the most appropriate threshold is for each individual patient, the combined 25 percent group and concurrent therapy limit is an arbitrary amount and would restrict therapists' ability to make appropriate treatment decisions. These commenters also stated that setting a limit on group and concurrent therapy may also restrict some patients from receiving the most appropriate mode of therapy for their individual need and that group or concurrent therapy might indeed be the most appropriate mode of therapy for a patient. These commenters stressed the importance of trusting the professional judgment of therapists in deciding which combination of each mode of therapy is appropriate for each patient in conjunction with Medicare guidelines for skilled therapy and medical necessity.

Response: We agree that therapists are the most appropriate professionals to determine the mode of therapy a patient should receive and that professional judgment must be trusted and used in SNFs. However we do not agree that 25 percent is an arbitrary amount. As stated in the proposed rule, (83 FR 21066), since we began allocating group therapy and concurrent therapy, these modes of therapy (group and concurrent) represent less than one percent of total therapy provided to SNF residents. Further, we do not agree that data do not exist with regard to the appropriate threshold for each individual patient, as over 99 percent of therapy services are currently reported as individual. This would suggest that a much lower threshold for concurrent and group therapy would likely be acceptable and appropriate, though we also believe that added flexibility is important under a new payment system. Therefore, we believe it is appropriate to use the 25 percent combined therapy limit for concurrent and group therapy.

We also do not agree that setting a limit on group and concurrent therapy may restrict some patients from receiving the most appropriate mode of therapy for their individual needs. We currently have a 25 percent limit in place for group therapy and, based on our data, this limit has not restricted patients from receiving what we assume is the most appropriate amount of therapy for their individual needs. Given the stakeholders' comments that individual therapy is the most costly form of therapy along with the evidence of therapy being furnished to SNF patients on the basis of financial considerations rather than patient need, the extremely high prevalence of individual therapy would indicate that the amount of individual therapy, despite being the most costly, is the most effective for beneficiaries, which would comport with our reasons for supporting either the limit we proposed or a lower such limit. To hold otherwise would indicate that the minutes currently being reported are an inaccurate representation of the way in which therapy is currently being delivered, which could potentially constitute fraud on the part of some SNF providers. Based on the MDS assessment data mentioned above that demonstrate that almost no group or concurrent therapy is being reported on the MDS currently, the commenters' characterization of the proposed limit (which is far above the current level of furnishing such services) as insufficiently flexible would actually beg the question of why commenters would appear to believe that group and concurrent therapy would be better suited to address patient needs under PDPM rather than under RUG-IV.

Given the historical precedent of 25 percent as a therapy threshold and the very limited amount that group and concurrent therapy that has actually been reported in SNFs, we believe it is an appropriate threshold. That being said, using the new items in section O of the PPS Discharge Assessment, we will monitor therapy provision as discussed in section V.D of this final rule and we will consider policy changes as we receive data and see how therapy is being furnished under PDPM.

Comment: Some commenters suggested that CMS revise the group therapy definition to include two to four participants while many commenters suggested that CMS revise the definition to include two to six participants doing the same or similar activities. In addition to better aligning with other settings such as Inpatient Rehabilitation Facilities (IRFs), commenters stated that this revision would allow increased flexibility so that patients in smaller SNFs could utilize and benefit from group therapy. One commenter stated that the CMS definition of concurrent therapy is arbitrary and does not reflect therapists' preferred practice. This commenter urged us to redefine concurrent therapy. Several commenters requested that CMS reconsider the “rigid” documentation requirements that accompany group therapy provision, stating a preference as a practitioner to use group therapy when patients can benefit from it. One commenter requested that we provide additional guidance to providers and MACs related to the level of appropriate documentation required for participation in group or concurrent therapy.

Response: We recognize the importance of alignment across settings. We may consider changing the definition of group therapy and/or concurrent therapy to align with other PAC settings in future rulemaking efforts.

With regard to the “rigid” documentation requirements, we would like to remind the commenter that we did not impose new documentation requirements on SNFs with regard to concurrent and group therapy. Rather, in the FY 2012 proposed rule, we simply clarified certain already-established documentation standards (76 FR 26387 through 26388). As we wrote in the FY 2012 final rule in response to comments, since we simply clarified existing expectations, we did not agree that these documentation guidelines would increase or create undue burden on therapists, or that these guidelines create a disincentive for clinicians to perform group therapy due to increased paperwork. We stated that there should be no additional burden to provide this documentation, as it should be a standard part of any documentation. We agreed with those commenters who stated that rehabilitation professionals need to support the work they do through documentation, and that the documentation should reflect the need for skilled care and the mode of therapy provided, as well as demonstrate how the therapy provision will support patients' needs and goals. (76 FR 48516).

We continue to believe that it is vital for SNFs to document services appropriately in order to demonstrate the skilled nature and the fact that the services are reasonable and necessary. This will be especially important when the 25 percent cap on concurrent and group therapy is in place after the implementation of PDPM. We will monitor the mode of therapy given and we will be interested to see how facilities document the therapy used so we can determine whether we will increase, decrease or maintain the limit following extensive monitoring.

Regarding the request to provide additional guidance related to documentation of group and concurrent therapy, we remind commenters of the guidance provided in the FY2012 proposed rule (76 FR 26388) regarding group therapy: Because group therapy is not appropriate for either all patients or all conditions, and in order to verify that group therapy is medically necessary and appropriate to the needs of each beneficiary, SNFs should include in the patient's plan of care an explicit justification for the use of group, rather than individual or concurrent, therapy. This description should include, but need not be limited to, the specific benefits to that particular patient of including the documented type and amount of group therapy; that is, how the prescribed type and amount of group therapy will meet the patient's needs and assist the patient in reaching the documented goals. In addition, we believe that the above documentation is necessary to demonstrate that the SNF is providing services to attain or maintain the highest practicable physical, mental, and psychosocial well-being of each resident in accordance with section 1819(b)(2) of the Act.

While the above guidance was provided in relation to group therapy, we believe that it applies to concurrent therapy as well.

Comment: Some commenters disagreed with the proposal of a combined limit of 25 percent for concurrent and group therapy, with one commenter stating that this contradicted the discussion in the ANPRM that considered a 25 percent limit on concurrent therapy and a separate 25 percent limit on group therapy. This commenter pointed out that we stated, we believe that individual therapy is usually the best mode of therapy provision as it permits the greatest degree of interaction between the resident and therapist, and should therefore represent, at a minimum, the majority of therapy provided to an SNF resident (82 FR 21004). This commenter and several others requested that CMS return to the separate 25 percent caps for concurrent therapy and for group therapy, as discussed in the ANPRM. According to these commenters, prior to CMS allocating concurrent and group therapy in FY 2011 and FY 2012, respectively, the average amount of concurrent and group therapy that was furnished to all residents combined was about 26 percent. These commenters believe this means that there were many residents who received higher amounts than an average of 25 percent group and concurrent therapy and others who received lower amounts based on their clinical status and need. According to these commenters, CMS has not produced any evidence the quality of care changed dramatically since FY 2011 and FY 2012, which would suggest the quality of care furnished in FY 2010 and earlier was meeting individual resident needs of patients. One commenter suggested that we implement a 25 percent combined cap for group and concurrent therapy at a facility level rather than at a per-patient level. One commenter requested that CMS consider having providers report “individual” and “non-individual” therapy, rather than separately reporting group and concurrent therapy.

Response: We do not agree that there is a contradiction between the ANPRM and our current proposal. We continue to believe that individual therapy should represent a majority of therapy provided in a SNF. We continue to contend that although group and concurrent therapy may have clinical merit in specific situations, we believe that individual therapy is generally the best way of providing therapy to a resident because it is most tailored to that specific resident's care needs. As such, we believe that individual therapy should represent at least the majority of the therapy services received by SNF residents. (82 FR 21004).

Our latest (FY 2017) data indicate that individual therapy was provided 99.77 percent of the time, meaning that group and concurrent therapy combined was reported as having been provided 0.23 percent of the time. If therapy continues to be provided in the same way, there is no reason to believe that a combined 25 percent limit on group and concurrent therapy is not a generous limit given the amount of group and concurrent therapy that has been provided under RUG IV. Therefore, we do not agree with the request to implement the separate 25 percent caps for group and concurrent therapy discussed in the ANPRM. We further disagree that CMS put restrictions on the “ability to furnish concurrent and group therapy.” We did not change any restrictions in FY 2011 and FY 2012 on the amount or type of therapy provided. The 25 percent cap on group therapy was in place since the inception of the SNF PPS. Rather, we allocated first concurrent therapy in FY 2011 (74 FR 40315-40319) and then group therapy in FY 2012 (76 FR 48511-48517) as a way to equalize payment across therapy modes and remove any financial incentives for providing a certain therapy mode, which appeared to drive at least some portion of the approximately 1,000 percentage increase in the amount of group therapy provided under the SNF Part A benefit in FY 2011. This was not an effort to restrict any mode of therapy. As we wrote in the FY 2012 final rule (76 FR 48513, 48514), by allocating group therapy among the four group therapy participants, we are also equalizing the reimbursement incentive across the modes of therapy. We stated we believe this would once again encourage clinicians to choose the mode of therapy based on clinical rather than financial reasons. We stated in the FY 2012 final rule that the purpose of our allocation policy is to provide payment that better reflects resource utilization and cost, and that we do not believe this policy should affect clinical determinations regarding the appropriate mode of therapy provided to a patient.

We appreciate the suggestion to implement a combined 25 percent group and concurrent therapy limit at the facility level rather than the patient level; however, given that a significant part of the reason we proposed a limit on group and concurrent therapy is so that patients receive therapy that reflects their individualized needs, we believe that implementing a facility based limit on concurrent and group therapy would defeat the purpose. With regard to a facility level limit, as opposed to the patient-level limit, we believe that therapy decisions should be driven by clinical standards and judgment related to an individual patient and not in relation to all patients within a facility. Utilizing a facility-level cap may allow for certain patients to receive excessive levels of group or concurrent therapy, which we do not believe would be advisable for any patient.

With regard to the comment that providers not be required to report group and concurrent therapy separately, while we have a combined cap, we believe that it is important to understand which of the two modes of therapy, concurrent or group therapy, is actually occurring in relation to this cap. Given that some commenters requested separate caps on group and concurrent therapy, we would not be in a position to assess the need for this separation in the future if group and concurrent therapy were reported under a single heading.

Comment: Several commenters expressed concern with how the combined group and concurrent therapy limit would interplay with student supervision in SNFs. One commenter stated the following, “Students' minutes are often counted as concurrent therapy when the clinical instructor is also treating a patient and we anticipate residents being treated by students will quickly exceed the 25 percent threshold.” The commenters went on to explain that the 25 percent limitation on group and concurrent therapy minutes could make it inefficient for the treating therapist or assistant and could deter facilities from taking students. One commenter was concerned that “CMS currently requires that student treatment must be labeled as “concurrent,” and therefore, this would fall under the 25 percent limitation on group and concurrent therapy. They stated that positive clinical education experiences in post-acute settings often translate into quality therapists and assistants getting jobs in those settings upon graduation. One commenter explained that if a SNF accepts more students, “the average of 1 percent for group and concurrent therapy represented in CMS data may not prove accurate.” They described a scenario where SNFs that prefer to have higher than average volumes of students may deliver concurrent therapy in excess of 25 percent and that the combined 25 percent limit of group and concurrent therapy could be a deterrent to SNFs taking therapy students One commenter recommended that CMS create a reporting requirement that would delineate between student and therapist/assistant minutes so that those minutes could be separated from the total of group and concurrent therapy minutes.

Response: We appreciate the concern that these commenters raised. We agree that our policies should not deter SNFs from taking students, and we agree that the therapy student internship is crucial to ensuring that students gain valuable SNF experience that would cause quality therapist and assistant graduates to pursue employment at SNFs when they eventually graduate. We appreciate the candor with which the commenters have described how they provide concurrent therapy at the same time as their therapy students consistent with current policy allowances. We would like to clarify that CMS does not require that student therapy be labeled as concurrent. The following is written in the MDS 3.0 RAI Manual (Chapter 3, section O):

When a therapy student is involved with the treatment, and one of the following occurs, the minutes may be coded as concurrent therapy: The therapy student is treating one resident and the supervising therapist/assistant is treating another resident, and both residents are in line of sight of the therapist/assistant or student providing their therapy.

This instruction is describing one possible scenario. We would like to reiterate that CMS does not require students to do concurrent therapy. As stated in the FY 2012 final rule (76 FR 48511), as the therapy student is under the direction of the supervising therapist (even if no longer required to be under line-of-sight supervision), the time the student spends with a patient will continue to be billed as if it were the supervising therapist alone providing the therapy. In other words, the therapy student, for the purpose of billing, is treated as simply an extension of the supervising therapist rather than being counted as an additional practitioner.

We suspect that, as noted in the FY 2012 final rule referenced above, because we do not allow facilities to count therapy students' independent time on the MDS, many facilities rely on the MDS instructions above (allowing a therapist or assistant and a student to treat one patient each while both residents are in line of sight of the therapist/assistant or student providing their therapy) to permit them to count student concurrent therapy time. However, this should in no way be considered mandatory practice and like all concurrent therapy, should be used sparingly.

Further, as mentioned above, our most recent (FY 2017) data show that individual therapy was provided 99.77 percent of the time, meaning that group and concurrent therapy combined was reported as having been provided 0.23 percent of the time. It concerns us that commenters have stated that they are providing so much concurrent therapy with students that the 25 percent cap would be too low for them, because this would suggest that either the comments were provided mistakenly or that facilities are falsely reporting concurrent therapy as individual therapy. While we agree with commenters that the opportunity to supervise student therapists in SNFs is valuable to the education of future therapists and assistants, our data indicate that a 25 percent combined cap on group and concurrent therapy should not deter facilities from taking more therapy students. We believe the recommendation to monitor student therapy minutes along with just therapist/assistant minutes has merit and it is something we will consider for future policy making.

Comment: Some commenters expressed concern with CMS' implication that clinical decisions about therapy are principally driven by “. . . financial considerations rather than the clinical needs of the SNF residents”.

Response: The available data support our assertion that at least some SNFs principally utilize financial considerations, rather than relying on clinical judgment, when making decisions regarding the manner and amount of care to provide to SNF residents. In 2016, CMS released the Skilled Nursing Facility Utilization and Payment Public Use File (Skilled Nursing Facility PUF) (https://www.cms.gov/​Research-Statistics-Data-and-Systems/​Statistics-Trends-and-Reports/​Medicare-Provider-Charge-Data/​SNF.html). The Skilled Nursing Facility PUF contained information on utilization, payment (allowed amount, Medicare payment and standard payment), submitted charges, and beneficiary demographic and chronic condition indicators organized by CMS Certification Number (6-digit provider identification number), Resource Utilization Group (RUG), and state of service. The SNF PUF included information on the number of provider assessments where residents were classified into an Ultra-High Rehabilitation RUG or a Very-High Rehabilitation RUG. It also included the percentage of those assessments that were within ten minutes of the minimum threshold used to classify a resident into that Rehabilitation RUG category (that is, between 500-510 minutes for RV RUGs and 720-730 minutes for RU RUGs). Based on this information, we found the following:

• 51 percent of all RV assessments showed therapy provided between 500 and 510 minutes.
• 65 percent of all RU assessments showed therapy provided between 720 and 730 minutes.
• For 88 providers, all of their RV assessments showed therapy provided between 500 and 510 minutes.
• For 215 providers, all of their RU assessments showed therapy provided between 720 and 730 minutes.
• More than one in five providers had more than 75 percent of both RU and RV assessments that showed therapy provided within 10 minutes of the minimum threshold.

This clear evidence of thresholding behavior supports our assertion regarding SNFs that are driven by payment considerations rather than therapy needs of patients. Furthermore, we received a significant number of comments from stakeholders on the proposed rule who believe that the quality and volume of therapy services are likely to diminish under PDPM. This belief is, itself, predicated on the notion that SNFs will continue to utilize financial considerations as the basis for care planning decisions. However, with better and more reliable patient diagnosis and characteristic data and given the removal of therapy service volume as a component of the payment system, we expect that we will be better positioned under PDPM to exercise our authority to make case-mix creep adjustments under section 1888(e)(4)(F) of the Act, as may be appropriate, to address any changes in payment which are merely the result of changes in the coding or classification of SNF patients that do not reflect actual changes in case mix. This type of analysis will also be a part of CMS monitoring efforts under PDPM.

Comment: One commenter recommended that, in the future, CMS consider whether it would be reasonable to track rehabilitative versus maintenance therapy, similar to how it is done in the home health setting.

Response: We appreciate this suggestion and may take it into consideration for future policy making decisions.

Accordingly, after considering the comments received, for the reasons discussed in the proposed rule and in this final rule, we are finalizing our proposal, without modification, to set a combined 25 percent limit on group and concurrent therapy per discipline. Additionally, we are finalizing our proposal, without modification, to implement a non-fatal warning edit on the validation report upon submission when the amount of group and concurrent therapy exceeds 25 percent within a given therapy discipline, which would alert the provider to the fact that the therapy provided to that resident exceeded the threshold.

F. Interrupted Stay Policy

Under section 1812(a)(2)(A) of the Act, Medicare Part A covers a maximum of 100 days of SNF services per spell of illness, or “benefit period.” A benefit period starts on the day the beneficiary begins receiving inpatient hospital or SNF benefits under Medicare Part A. (See section 1861(a) of the Act; § 409.60). SNF coverage also requires a prior qualifying, inpatient hospital stay of at least 3 consecutive days' duration (counting the day of inpatient admission but not the day of discharge). (See section 1861(i) of the Act; § 409.30(a)(1)). Once the 100 available days of SNF benefits are used, the current benefit period must end before a beneficiary can renew SNF benefits under a new benefit period. For the current benefit period to end so a new benefit period can begin, a period of 60 consecutive days must elapse throughout which the beneficiary is neither an inpatient of a hospital nor receiving skilled care in a SNF. (See section 1861(a) of the Act; § 409.60). Once a benefit period ends, the beneficiary must have another qualifying 3-day inpatient hospital stay and meet the other applicable requirements before Medicare Part A coverage of SNF care can resume. (See section 1861(i); § 409.30)

While the majority of SNF benefit periods, approximately 77 percent, involve a single SNF stay, it is possible for a beneficiary to be readmitted multiple times to a SNF within a single benefit period, and such cases represent the remaining 23 percent of SNF benefit periods. For instance, a resident can be readmitted to a SNF within 30 days after a SNF discharge without requiring a new qualifying 3-day inpatient hospital stay or beginning a new benefit period. SNF admissions that occur between 31 and 60 days after a SNF discharge require a new qualifying 3-day inpatient hospital stay, but fall within the same benefit period. (See sections 1861(a) and (i) of the Act; §§ 409.30, 409.60)

Other Medicare post-acute care (PAC) benefits have “interrupted stay” policies that provide for a payment adjustment when the beneficiary temporarily goes to another setting, such as an acute care hospital, and then returns within a specific timeframe. In the inpatient rehabilitation facility (IRF) and inpatient psychiatric facility (IPF) settings, for instance, an interrupted stay occurs when a patient returns to the same facility (or in the case of an IPF, the same or another IPF) within 3 days of discharge. The interrupted stay policy for long-term care hospitals (LTCHs) is more complex, consisting of several policies depending on the length of the interruption and, at times, the discharge destination: An interruption of 3 or fewer days is always treated as an interrupted stay, which is similar to the IRF PPS and IPF PPS policies; if there is an interruption of more than 3 days, the length of the gap required to trigger a new stay varies depending on the discharge setting. In these three settings, when a beneficiary is discharged and returns to the facility within the interrupted stay window, Medicare treats the two segments as a single stay.

As we explained in the proposed rule (83 FR 21068), while other Medicare PAC benefit categories have interrupted stay policies, the SNF benefit under the RUG-IV case-mix model had no need for such a policy because, given a resident's case-mix group, payment did not change over the course of a stay. In other words, assuming no change in a patient's condition or treatment, the payment rate was the same on Day 1 of a covered SNF stay as it is at Day 7. Accordingly, a beneficiary's readmission to the SNF—even if only a few days may have elapsed since a previous discharge—could essentially be treated as a new and different stay without affecting the payment rates.

However, as described in section V.D of the proposed rule (83 FR 21068) and section V.C.4 of this final rule, we stated that the PDPM would adjust the per diem rate across the length of a stay (the variable per diem adjustment) to better reflect how and when costs are incurred and resources used over the course of the stay, such that earlier days in a given stay receive higher payments, with payments trending lower as the stay continues. In other words, the adjusted payment rate on Day 1 and Day 7 of a SNF stay may not be the same. Although we stated that we believe this variable per diem adjustment schedule more accurately reflects the increased resource utilization in the early portion of a stay for single-stay benefit periods (which represent the majority of cases), we considered whether and how such an adjustment should be applied to payment rates for cases involving multiple stays per benefit period. In other words, in the proposed rule, we considered instances in which a resident has a Part A stay in a SNF, leaves the facility for some reason, and then is readmitted to the same SNF or a different SNF; and how this readmission should be viewed in terms of both resident classification and the variable per diem adjustment schedule under the proposed PDPM. We explained that application of the variable per diem adjustment is of particular concern because providers may consider discharging a resident and then readmitting the resident shortly thereafter to reset the resident's variable per diem adjustment schedule and maximize the payment rates for that resident.

We stated in the proposed rule (83 FR 21068) that, given the potential harm which may be caused to the resident if discharged inappropriately, and other concerns outlined previously in this section and in the proposed rule, we discussed in last year's FY 2018 ANPRM the possibility of adopting an interrupted stay policy under the SNF PPS in conjunction with the implementation of the RCS-I case-mix model. Several commenters expressed support for this interrupted stay policy in responding to the ANPRM, saying that the interrupted stay policy is in alignment with similar policies in other post-acute settings, and that a similar policy would likely be implemented under any cross-setting PAC payment system.

Thus, we proposed to implement an interrupted stay policy as part of the SNF PPS, effective beginning FY 2020 in conjunction with the proposed implementation of the SNF PDPM. Specifically, in cases where a resident is discharged from a SNF and returns to the same SNF by 12:00 a.m. at the end of the third day of the interruption window (as defined below), we proposed treating the resident's stay as a continuation of the previous stay for purposes of both resident classification and the variable per diem adjustment schedule. In cases where the resident's absence from the SNF exceeds this 3-day interruption window (as defined below), or in any case where the resident is readmitted to a different SNF, we proposed treating the readmission as a new stay, in which the resident would receive a new 5-day assessment upon admission and the variable per diem adjustment schedule for that resident would reset to Day 1. We stated in the proposed rule (83 FR 21068 through 21069) that, consistent with the existing interrupted stay policies for the IRF and IPF settings, we would define the interruption window as the 3-day period starting with the calendar day of discharge and additionally including the 2 immediately following calendar days. We stated that for the purposes of the interrupted stay policy, the source of the readmission would not be relevant. That is, the beneficiary may be readmitted from the community, from an intervening hospital stay, or from a different kind of facility, and the interrupted stay policy would operate in the same manner. We explained that the only relevant factors in determining if the interrupted stay policy would apply are the number of days between the resident's discharge from a SNF and subsequent readmission to a SNF, and whether the resident is readmitted to the same or a different SNF.

In the proposed rule (83 FR 21069), we presented the following examples, which we believed aided in clarifying how this policy would be implemented:

Example A: A beneficiary is discharged from a SNF on Day 3 of the stay. Four days after the date of discharge, the beneficiary is then readmitted (as explained above, this readmission would be in the same benefit period) to the same SNF. The SNF would conduct a new 5-day assessment at the start of the second admission and reclassify the beneficiary accordingly. In addition, for purposes of the variable per diem adjustment schedule, the payment schedule for the second admission would reset to Day 1 payment rates for the beneficiary's new case-mix classification.

Example B: A beneficiary is discharged from a SNF stay on Day 7 and is readmitted to the same SNF within the 3-day interruption window. For the purposes of classification and payment, this would be considered a continuation of the previous stay (an interrupted stay). The SNF would not conduct a new 5-day assessment to reclassify the patient and for purposes of the variable per diem adjustment schedule, the payment schedule would continue where it left off at the rate for the day of discharge; we stated in the proposed rule that, in this case, the first day of the second stay would be paid at the Day 8 per diem rates under that schedule.

Example C: A beneficiary is discharged from a SNF stay on Day 7 and is readmitted to a different SNF within the 3-day interruption window. The SNF would conduct a new 5-day assessment at the start of the second admission and classify the beneficiary accordingly. In addition, for purposes of the variable per diem adjustment schedule, the payment schedule for the second admission would reset to Day 1 payment rates for the beneficiary's new case-mix classification.

We note two clarifications to the preceding examples. In each of the above examples, when the beneficiary is discharged from the SNF stay, the SNF would complete the required PPS Discharge Assessment (see Table 33: PPS Assessment Schedule under PDPM). Additionally, in Example B, we inadvertently indicated in the proposed rule that the first day of the second stay would be paid at the Day 8 per diem rates. However, the first day of the second stay would actually be paid at the rate for the day of discharge, Day 7. These points are further addressed in our responses to comments below.

We also stated in the proposed rule (83 FR 21069) that we considered alternative ways of structuring the interrupted stay policy. For example, we considered possible ranges for the interrupted stay window other than the 3 calendar day window proposed. For example, we considered windows of fewer than 3 days (for example, 1 or 2 day windows for readmission), as well as windows of more than 3 days (for example, 4 or 5 day windows for readmission). However, we stated we believe that 3 days represents a reasonable window after which it is more likely that a resident's condition and resource needs will have changed. We also stated that we believe consistency with other payment systems, like that of IRF and IPF, is helpful in providing clarity and consistency to providers in understanding Medicare payment systems, as well as making progress toward standardization among PAC payment systems.

In addition, we explained that, to determine how best to operationalize an interrupted stay policy within the SNF setting, we considered three broad categories of benefit periods consisting of multiple stays. The first type of scenario, SNF-to-SNF transfers, is one in which a resident is transferred directly from one SNF to a different SNF. The second case we considered, and the most common of all three multiple-stay benefit period scenarios, is a benefit period that includes a readmission following a new hospitalization between the two stays—for instance, a resident who was discharged from a SNF back to the community, re-hospitalized at a later date, and readmitted to a SNF (the same SNF or a different SNF) following the new hospital stay. The last case we considered was a readmission to the same SNF or a different SNF following a discharge to the community, with no intervening re-hospitalization.

We further explained that, to simplify the analysis, we primarily examined benefit periods with two stays. We stated that benefit periods with exactly two stays account for a large majority (70 percent) of all benefit periods with multiple stays, and benefit periods with more than two stays represent a very small portion (less than 7 percent) of all benefit periods overall. We therefore assume the data for cases where there are exactly two stays in a benefit period are representative of all benefit periods with multiple stays. We noted that, of cases where there are exactly two stays in a benefit period, over three quarters (76.4 percent) consist of re-hospitalization and readmission (to the same SNF or a different SNF). Discharge to the community and readmission without re-hospitalization cases represent approximately 14 percent of cases, while direct SNF-to-SNF transfers represent approximately 10 percent.

For each of these case types, in which a resident was readmitted to a SNF after discharge, we explained that we examined whether (1) the variable per diem adjustment schedule should be “reset” back to the Day 1 rates at the outset of the second stay versus “continuing” the variable per diem adjustment schedule at the point at which the previous stay ended, and (2) a new 5-day assessment and resident classification should be required at the start of the subsequent SNF stay.

With regard to the first question above, specifically whether or not a readmission to a SNF within the proposed 3-day interruption window would reset the resident's variable per diem adjustment schedule, we stated that in each of the cases described above, we were concerned generally that an interrupted stay policy that “restarts” the variable per diem adjustment schedule to Day 1 after readmissions could incentivize unnecessary discharges with quick readmissions. We explained that this concern is particularly notable in the second and third cases described above, as the beneficiary may return to the same facility. As we discussed in the proposed rule (83 FR 21069), to investigate this question, we conducted linear regression analyses to examine changes in costs in terms of both PT/OT and NTA costs per day from the first to second admission for the three scenarios described above (SNF-to-SNF direct transfers, readmissions following re-hospitalization, and readmissions following community discharge). As discussed in section V.D.4. of the proposed rule (83 FR 21060 through 21061) and in section V.C.4 of this final rule, investigations revealed that utilization of PT, OT, and NTA services changes over the course of a stay. Based on both empirical analysis and feedback from multiple technical expert panels, we determined that SLP and nursing utilization remained fairly constant over a stay. Therefore, we proposed variable per diem adjustment schedules for the PT, OT, and NTA components but not for the SLP or nursing components. We stated in the proposed rule that, because the analysis of changes in costs across two stays in a single benefit period is relevant to determining how the variable per diem payment adjustments should apply to benefit periods with multiple stays, we restricted our analysis to the three payment components for which we are proposing variable per diem adjustments (PT, OT, and NTA). For this analysis, both the re-hospitalization and community discharge cases were separated into two sub-cases: When the resident returns to the same SNF, and when the resident is admitted to a different SNF. By definition, SNF-to-SNF transfer cases always have different providers for the first and second stays. We stated in the proposed rule that the regression results showed that PT/OT costs from the first to second admission were very similar for SNF-to-SNF transfers and for readmissions to a different provider following re-hospitalization or discharge to community, suggesting that the second admission is comparable to a new stay. NTA costs from the first to second admission also were very similar for SNF-to-SNF transfers. We stated that, for readmissions following re-hospitalization or discharge to community, NTA costs for readmissions to the same provider were notably less than NTA costs for readmissions to a different provider. We explained that, overall, these results suggest that a readmission to a different SNF, regardless of whether it was a direct SNF-to-SNF transfer, or whether the beneficiary was re-hospitalized or discharged to the community before the second admission, are more comparable to a new stay than an interrupted stay. Thus, we proposed to always reset the variable per diem adjustment schedule to Day 1 whenever residents are discharged and readmitted to a different SNF. We acknowledged that this could lead to patterns of inappropriate discharges and readmissions that could be inconsistent with the intent of this policy; for example, we stated we would be concerned about patients in SNF A consistently being admitted to SNF B to the exclusion of other SNFs in the area. We explained that should we discover such behavior, we would flag these facilities for additional scrutiny and review and consider potential policy changes in future rulemaking. However, based on the results of our regression analyses, and because of the concern that a SNF provider could discharge and promptly readmit a resident to reset the variable per diem adjustment schedule to Day 1, we stated that in cases where a resident returns to the same provider we were proposing to allow the payment schedule to reset only when the resident has been out of the facility for at least 3 days. As previously mentioned, we stated that believe 3 days represents a reasonable window after which it is more likely that a resident's condition and resource needs will have changed, and this 3-day requirement is also consistent with the interrupted stay policies of similar Medicare PAC benefits. Moreover, we stated that while we found that PT and OT costs for cases where the gap is longer than 3 days are similar to PT and OT costs for cases where the gap is shorter than 3 days, NTA costs are notably higher for cases where the gap is longer than 3 days. We explained that this provides further support for resetting the variable per diem schedule for cases where the gap is longer than 3 days (as costs tend to be higher, similar to a new stay). More information on these analyses can be found in section 3.10.3. of the SNF PMR technical report available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html.

We explained in the proposed rule (83 FR 21070) that with regard to the question of whether or not SNFs would be required to complete a new 5-day assessment and reclassify the resident after returning to the SNF within the proposed 3-day interruption window, we investigated changes in resident characteristics from the first to the second stay within a benefit period. First, we looked at changes in clinical categories from the first to second stay for residents with an intervening re-hospitalization. We explained that this analysis could only be conducted for residents with a re-hospitalization because, as described in section 3.10.2. of the SNF PMR technical report, for research purposes, classification into clinical categories was based on the diagnosis from the prior inpatient stay. We stated that for those residents who had a re-hospitalization and were readmitted to a SNF (either the same or a different SNF), and therefore, could be reclassified into a new clinical category (because of new diagnostic information as a result of the intervening re-hospitalization), we found that a majority had the same clinical category for both the first and second admission. We further explained that because we could not conduct this investigation for SNF-to-SNF transfers or community discharge cases (as they lack a new hospitalization), we separately investigated changes in function from the first to second stay for SNF-to-SNF transfers and for readmissions following community discharge. We found that in a large majority of cases, there was no change in function from the first to second stay, regardless of whether the second provider was the same or different as the first provider. Thus, we stated we believe it would be appropriate to maintain the classification from the first stay for those residents returning to the same SNF no more than 3 calendar days after discharge from the same facility. However, we stated that because we proposed to exclude from the interrupted stay policy readmissions to a different SNF (regardless of the number of days between admissions) and readmissions to the same SNF when the gap between admissions is longer than 3 days, and to treat these readmissions as new stays for purpose of the variable per diem adjustment schedule, we believe it would be appropriate and consistent to treat these cases as new stays for purposes of clinical classification and to require a new 5-day PPS assessment. More information on these analyses can be found in section 3.10.2. of the SNF PMR technical report available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html. Additionally, we noted that under the approach discussed in section V.E.1. of the proposed rule, providers would be afforded the flexibility to use the IPA, which would allow for resident reclassification under certain circumstances.

We also noted that we believe that frequent SNF readmissions may be indicative of poor quality care being provided by the SNF. Given this belief, we stated we plan to monitor the use of this policy closely to identify those facilities whose beneficiaries experience frequent readmission, particularly facilities where the readmissions occur just outside the 3-day window used as part of the proposed interrupted stay policy. We stated that should we discover such behavior, we would flag these facilities for additional scrutiny and review and consider potential policy changes in future rulemaking.

We invited comments on the proposals outlined above. Commenters submitted the following comments related to the proposed rule's discussion of the proposed interrupted stay policy under the PDPM. A discussion of these comments, along with our responses, appears below.

Comment: Several commenters pointed out a potential adverse incentive associated with the interaction between the interrupted stay policy, the proposed Interim Payment Assessment (IPA), and the variable per diem adjustment. Specifically, these comments were concerned with issues that could arise because an IPA does not return the NTA component to day 1 payment rates under the variable per diem adjustment schedule. Commenters stated that if a patient requires a new high cost medication or piece of equipment, the inability to return to day 1 of the variable per diem adjustment schedule could result in an array of unintended issues. Commenters noted that these unintended issues include incentivizing unnecessary discharges to a hospital followed by quick readmissions (which, the commenter pointed out, was a risk CMS had specifically considered and attempted to avoid in crafting the proposed interrupted stay policy) and reluctance to admit patients who are at high risk of changes in care needs. One commenter stated that CMS has not aligned the planned monitoring of unnecessary discharges with existing quality measures, and instead has created an incentive for unnecessary discharges and readmissions just outside the 3-day interruption window by prohibiting providers from returning patients to days one through three of the variable per diem adjustment schedule for typically high cost NTAs when an IPA is conducted or in the instance of interrupted stays of 3 or less days.

Response: While we appreciate the commenters' concerns regarding the potential for an adverse incentive, we believe that frequent SNF readmissions may be indicative of poor quality care being provided by the SNF. CMS plans to monitor the use of this policy closely to identify those facilities whose beneficiaries experience frequent readmission, particularly facilities where the readmissions occur just outside the 3-day window used as part of the proposed interrupted stay policy. Should we discover such behavior, we will flag these facilities for additional scrutiny and review and consider potential policy changes in future rulemaking.

We do not believe that facilities have cause for reluctance to admit patients who are at high risk of changes in care needs. The optional IPA allows for patients to be reclassified in cases of significant changes in care needs.

With regard to the question of the IPA resetting the variable per diem adjustment, this issue is addressed in our responses to comments in section V.D. of this final rule.

With regard to the question of interruptions of 3 or less days resetting the variable per diem adjustment, as we stated in the proposed rule, our analyses found that some costs, specifically NTA costs, are notably higher for cases where the gap is longer than 3 days, compared to cases where the interruption is 3 or less days, where costs are more similar to uninterrupted stays. We believe this supports the use of a 3-day gap as the metric for when to reset the variable per diem adjustment.

Regarding any current alignment of quality measures and the monitoring of unnecessary discharges, we interpret the commenter to be suggesting that CMS does not currently have in place quality measures that address unnecessary discharges to the hospital during the SNF Stay. We disagree with this assertion in that CMS has developed and implemented a hospital readmission measure for SNF.

Comment: Commenters requested clarification as to whether the readmission of a patient under the interrupted stay policy (for example, within the 3-day window) would trigger an IPA.

Response: As discussed in section V.D. of this final rule, the IPA under PDPM would be an optional assessment. Therefore, readmission after an interrupted stay would not trigger an IPA. If the provider believes, even in the case of a short absence from the facility, that an IPA is warranted, then we would encourage the provider to complete an IPA in that instance.

Comment: Several commenters requested clarification about completing initial evaluations for therapy upon readmission of a patient in an interrupted stay under the interrupted stay policy. Commenters questioned whether CMS has an expectation that therapists will always complete a new evaluation upon the resident's return to a SNF as currently instructed in the MDS RAI Manual, or whether CMS would defer to the clinical judgment of the therapist in a way that is more like the EOT/EOT-R practice. Commenters also questioned whether CMS would require SNFs to indicate on the claim form when a resident has been readmitted and/or when an evaluation was complete after the resident was readmitted. Commenters pointed out that, per the current instruction in section O of the MDS RAI Manual, “If a resident returns from a hospital stay, an initial evaluation must be performed after entry to the facility, and only those therapies that occurred since admission/reentry to the facility and after the initial evaluation shall be counted.” (MDS 3.0 Chapter 3, section O, V1.15, page O-19). On the other hand, commenters pointed out that the premise for the interrupted stay policy is similar to the policy for the End-of Therapy (EOT) Other Medicare Required Assessment (OMRA), which leaves it to the clinician's judgment whether or not a new therapy evaluation should be completed. Commenters stated that when therapy is the primary skill, and the patient misses 3 consecutive calendar days of therapy, the provider must complete an EOT OMRA, which effectively changes the payment resource utilization group (RUG). Commenters pointed out that in cases where therapy resumes after the EOT-OMRA is performed and the resumption of therapy date is no more than 5 consecutive calendar days after the last day of therapy provided, and the therapy services have resumed at the same RUG-IV classification level, and with the same therapy plan of care that had been in effect prior to the EOT OMRA, an EOT OMRA with Resumption (EOT-R) may be completed. Commenters noted that in these cases, it is left to the clinician's judgment whether or not a new therapy evaluation should be completed.

Response: Given that an interrupted stay does not prompt the need for a new 5-day PPS assessment and continues the stay from the point when the interruption occurred, providers should not be required to always complete an evaluation upon the resident's readmission after an interrupted stay. Per the proposed interrupted stay policy, a new 5-day assessment must be completed only if the interruption lasts longer than 3 days (or if the beneficiary is readmitted to a different SNF). If the interruption was less than 3 days but patient care needs have changed significantly, clinicians may complete an IPA at their discretion. The instructions in the MDS RAI Manual will be updated accordingly as part of the implementation of PDPM.

With regard to whether providers would be required to report on the claim form when a patient is readmitted or an evaluation is completed for such a patient, we do not anticipate such changes in claims reporting, though we would have providers report on the claim when an interrupted stay occurred.

Comment: Many commenters had questions and concerns related to discharge practices under the interrupted stay policy, and requested clarification of the requirements surrounding the PPS Part A Discharge (NPE) when beneficiaries meet the criteria of an interrupted stay. One commenter stated that it is unclear in the proposed rule whether the NPE would be completed in example B in the FY 2019 SNF PPS proposed rule (83 FR 21069). Assuming that an NPE would be required once the resident has been out of the facility for 24 hours, whether the resident returns within 1 day or 3 days, commenters questioned how the facility would manage the assessment schedule versus the payment schedule. Other commenters questioned whether CMS expects SNFs to wait to see whether the beneficiary returns before completing the discharge assessment. Commenters questioned what the implications would be for setting the Assessment Reference Date (ARD) approximately 4 to 5 days after discharge in cases when the beneficiary does not return within the 3-day window. Commenters stated that as currently defined, doing this would be considered a late assessment, and could subject the SNF to penalties. Commenters also stated that if this discharge assessment is required, then this adds to the administrative burden, which is contradictory to CMS' stated goals.

Response: As is the current policy, SNFs would be expected to complete the PPS discharge assessment and/or OBRA discharge assessment upon any discharge and within currently established timeframes, regardless of any expectation as to whether or not a patient might be readmitted and/or whether the readmission would be considered an interrupted stay. This does not add administrative burden beyond what SNFs are currently expected to do. This information is also important in our ability to assess instances in which facilities may abuse the interrupted stay policy.

With regard to managing the assessment schedule and payment schedule, we would refer commenters to the assessment schedule discussed in section V.D of this final rule, which outlines both the assessment calendar and payment timeline for each assessment under PDPM.

Comment: Some commenters sought clarification as to how the SNF should count the total volume, mode, and type of therapy to report in section O of the MDS for purposes of the discharge assessment when a resident's stay included one or more interrupted stays. Would they count it from Day 1, the original admission date, even though there was an interrupted stay, or would this discharge assessment only include the volume, mode, and type of therapy delivered since the time of return to discharge?

Response: In cases where a resident is discharged and then readmitted to a SNF in a manner that triggers an interrupted stay under the interrupted stay policy, only those therapies that occurred since the readmission would be included in section O of the MDS for each discharge assessment.

Comment: A commenter expressed concerns related to the use of the length of an interruption in days (for example, less than or equal to 3 days) as the trigger for a 5-day assessment. The commenter stated appreciation for CMS efforts to reduce the number of 5-day assessments, but stated that no reduction in burden is achieved by not requiring a 5-day assessment for patients returning following 3 or fewer days, assuming that SNFs must still conduct a patient assessment upon readmission for all patients. Also, the commenter believes not performing a 5-day assessment for all returning patients creates unneeded risk for patients and SNFs. The commenter recommended performing the 5-day assessment after every readmission, the result of which—not the number of days in the interruption—should determine whether the patient's condition has changed and new care needs are present that would warrant resetting the variable per diem rate. Commenters stated that the number of days in an interruption is irrelevant to costs of treatment and it is the patient's condition upon return from the interruption that should determine whether the payment resets to day 1 per diem rates or not.

Response: Contrary to the commenter's assertion, we believe that a reduction in burden is, in fact, achieved by not requiring a 5-day assessment for patients returning following 3 or fewer days. While SNFs may be required to complete OBRA assessments and other statutorily required assessments beyond the scope of SNF PPS payment, it will no longer be the case that SNFs must conduct a patient assessment upon readmission for all patients for the purposes of PPS payment. As discussed above, in conjunction with the implementation of the PDPM, CMS will reduce the assessment schedule significantly to ease provider burden (see section V. E. and Table 33 of the proposed rule). The Start of Therapy OMRA, the assessment that would have previously been required for PPS payment upon a readmission, is no longer required. The new schedule utilizes the 5-day Assessment and PPS Discharge Assessments as the only required assessments, with IPAs being optional at clinician discretion.

We disagree that not performing an assessment for all returning patients creates unneeded risk. We believe that the new assessment schedule we proposed achieves efficiencies in terms of provider burden while still providing enough data to accurately monitor provider behavior, changes in patient condition, and outcomes via the 5-day assessment, IPA assessments, and discharge assessments. While a 5-day assessment would not be required upon readmission in the case of an interrupted stay, the provider has the option of completing an IPA as it determines appropriate to assess whether the patient's condition and care needs have changed.

While we appreciate the commenter's concern, we believe the use of the number of days between discharge and readmission to determine whether there is an interrupted stay is appropriate. As described in the proposed rule, our analyses found that some types of costs, notably NTA costs, tend to be higher for cases where the gap is longer than 3 days, suggesting that such stays are more like new stays than continuing stays and thus supporting the 3-day metric for resetting the variable per diem schedule. The length of the interruption is also used in determining whether there is an interrupted stay in other Medicare post-acute payment systems and we expect that its use here will be just as effective.

With regard to the commenters' recommendation that a 5-day assessment be completed upon readmission after an interrupted stay, we believe that this would constitute an unnecessary burden on providers, particularly given the provider's option to complete an IPA upon readmission to the SNF. We also do not believe a 5-day assessment is necessary upon readmission after an interrupted stay of 3 days or less. While we found that PT and OT costs for cases where the gap is longer than 3 days are similar to PT and OT costs for cases where the gap is shorter than 3 days, NTA costs are notably higher for cases where the gap is longer than 3 days. We explained that this provides further support for resetting the variable per diem schedule for cases where the gap is longer than 3 days (as costs tend to be higher, similar to a new stay). As discussed in section 3.10 of the SNF PMR technical report, our analyses also showed that clinical category (in cases with an intervening re-hospitalization) and functional status (in cases involving SNF-to-SNF transfers and readmissions following community discharge) tended not to change between the first stay and the second stay in an interrupted stay of 3 days or less. Thus, we believe our research suggests that stays with interruptions of 3 days or less are more similar in cost to uninterrupted stays and are less likely to involve significant changes in patient condition or function. Therefore, we do not agree that a 5-day assessment should be required upon readmission after an interrupted stay, or that it is appropriate to reset the variable per diem adjustment schedule to day 1 after an interrupted stay.

We agree with the commenter that the patient's condition should be the most relevant factor in determining the need for a new assessment, and CMS has given providers the option of performing an IPA at their discretion based on changing conditions. As we explained previously, if a new condition is coded on an IPA, the SNF PPS per diem payment for the patient could increase to reflect changes in the patient's clinical condition if there is a change in the patient's case-mix group.

Comment: A commenter stated that CMS does not explicitly discuss discharge to the community and the interrupted stay policy, and requested clarification.

Response: In the FY 2019 SNF PPS proposed rule (83 FR 21068 through 21069), we discussed discharge to the community and the interrupted stay policy. The beneficiary may be readmitted from the community, from an intervening hospital stay, or from a different kind of facility, and the interrupted stay policy would operate in the same manner. The interrupted stay policy would operate in the same manner for discharges to the community.

Comment: One commenter commented that the RAI User's Manual instructions for A2400A, on page A-32, are to code 1, yes, if the resident has had a Medicare Part A covered SNF stay since the most recent admission/entry or reentry. The commenter stated that providers also use the Medicare Stay End Date Algorithm on page A-37 of the RAI User's Manual to correctly code A2400C, the end of the Medicare SNF stay. A2400C is also used to determine whether the PPS Part A Discharge assessment is required. The commenter referenced Example B on page 21069 of the proposed rule, which describes a beneficiary who is discharged on day 7 and is readmitted to the same SNF within the 3-day interruption window. The example states a SNF would not conduct a new 5-day assessment, and for the purposes of payment, this would be considered a continuation of the previous stay. The commenter expressed concern that, even though the Example B beneficiary is considered a continuation of the previous stay for payment purposes, A2400 on the MDS would still be coded as two separate Medicare stays. The commenter stated that when the resident is discharged on day 7, this date would be considered the end of the Medicare stay at A2400C. The entry record completed when the resident returned would have a new Medicare start date (A2400B) that would equal the reentry date. The commenter stated that this could lead to unmatched stays and inaccurate SNF QRP measures.

Response: We appreciate the comments on the potential revisions needed to the MDS manual or any technical specifications associated with SNF programs to implement the interrupted stay policy, and will consider these issues when making revisions to these materials as part of implementing the PDPM and related policies. With regard to the commenter's concern about the alignment of individual stays in the SNF QRP and the PDPM, we are aware of the issue and will revise the codes so that a hospital admission and return to the SNF does not trigger a new Medicare stay for purposes of the SNF QRP.

Comment: A commenter expressed concern regarding how the interrupted stay policy will operate in situations where the SNF provided the resident with the Notice of Medicare Non-Coverage (NOMNC), which is required to be provided prior to a discharge to the community. The commenter requested clarification on how or if issuance of the NOMNC or SNFABN would have any effect on the interrupted stay policy. Their concern was that if a resident meets the criteria of an interrupted stay following a discharge where denial notices were issued, the resident would be considered a new admission to the SNF. The commenter stated the cost of an admission in this situation is more like that of a new admission than a readmission. They recommended that the interrupted stay policy not be applied following a discharge with issuance of denial notices.

Response: The basic purpose of the interrupted stay policy is to ensure that when two segments of a resident's stay in the facility are separated by only a brief absence, the variable per diem payment adjustment is not inappropriately reset to Day 1 upon the resident's return. We do not believe that the mere issuance of a denial notice such as a NOMNC or SNFABN prior to the resident's departure would, in itself, have any effect on the nature of the care needed by the resident upon subsequent resumption of SNF care, the costs of readmission, or the way in which providers would be paid under the PDPM, and, accordingly, we are not adopting the commenter's suggestion.

Comment: A commenter expressed concern about the impact an OBRA Discharge Return Not Anticipated assessment would have on the interrupted stay policy. The commenter stated that currently, when a resident discharges to the community with the intent not to return, the SNF is required to complete the OBRA Discharge Return Not Anticipated assessment and would combine this assessment with the PPS Part A Discharge. The commenter stated that the OBRA Discharge Return Not Anticipated ends the resident's “episode of care.” The commenter stated that if this resident were to be readmitted to the SNF within the interruption window, this would be considered a new admission, require an admission type of entry record, and start a new “episode of care.” Furthermore, the commenter stated that this discharge would end all of the resident's orders, meaning that a new admission order is required, along with new physician certification of skilled care and new therapy evaluations. The commenter was highly concerned that the interrupted stay policy would apply following an OBRA Discharge Return Not Anticipated assessment, when the resident is considered a “new admission” for all other regulations. The commenter stated that the cost of an admission in this situation is more like that of a new admission than a readmission. The commenter recommended that the interrupted stay policy not be applied following a Discharge Return Not Anticipated.

Response: We appreciate this concern though we do not agree that the interrupted stay policy should not apply in cases where the resident is discharged return not anticipated. While the provider may have prepared a discharge plan for this patient based on the notion that the patient would not return, the patient's return to the SNF within that 3-day window would suggest that either the patient was not adequately prepared for discharge or may have been discharged too early from the facility. Further, providers should consider the possibility that a patient may return before finalizing the precise discharge type coded on the MDS. Finally, we believe that exempting such discharges from the interrupted stay policy could incentivize providers to merely code discharges in this manner only for this purpose and without sufficient basis.

Comment: One commenter stated that currently a Medicare Part A stay in the SNF will end if the resident has been discharged to the community, has been admitted to the hospital, or is on a hospital observation stay or emergency room visit that spans midnight and exceeds 24 hours. The commenter stated that the interrupted stay policy would consider any readmission within the 3-day interruption window as a continuation of the previous stay, therefore changing the number of Medicare stays the facility would have had prior to this proposal. One commenter expressed concern that the reduction in Medicare stays has the potential to affect the SNF QRP measures adversely by resulting in a higher number of unmatched stays and potential errors with SNF QRP measure calculation. The commenter referenced the Skilled Nursing Facility Quality Reporting Program Measure Calculation and Reporting User's Manual 1.0 (https://www.cms.gov/​Medicare/​Quality-Initiatives-Patient-Assessment-Instruments/​NursingHomeQualityInits/​Downloads/​SNF-QM-Users-Manual-V10-FINAL-5-22-17.pdf) and the instructions on how to identify a Medicare Part A stay for SNF QRP: Start by sorting assessments in reverse order during the 12-month target period. If the most recent assessment is a PPS Part A Discharge assessment, look for the next qualifying assessment; if the assessment is a 5-day, this is a matched assessment, and if not a 5-day, the stay is unmatched. The commenter expressed concern for potential negative impact to the SNF QRP measures, regardless of whether the Discharge Assessment NPE is required with the discharge prior to the interrupted stay, with the following reasoning.

The commenter described a sequence of assessments and events that the commenter stated would occur under the current payment system if Example B on page 21069 were to occur: 5-day assessment, NPE, discharge of less than 3 days, 5-day assessment, and final NPE. This would be counted as two Medicare stays for SNF QRP.

The commenter then described how this sequence might differ under the new system, depending on whether the NPE is required or not. In Example B, if the NPE was required on day 7 when the resident was discharged, but a new 5-day assessment was not required when the resident returned within the interruption window, then the sequence of assessments and events would be: 5-day assessment, NPE, interrupted stay, NPE. This would result in one unmatched stay (between the return from the interrupted stay to the final NPE) and one matched stay.

In Example B, if the NPE is not required on day 7 when the resident discharges for less than 3 days, the sequence would be: 5-day assessment, interrupted stay, NPE. This would result in only one Medicare stay.

The commenter requested clarification on how the Medicare stays will be calculated with the interrupted stay policy, presumably for the purposes of the QRP, and recommended evaluation by the SNF QRP CMS team to evaluate any further risks, errors, or concerns that may arise from this proposed policy.

Response: We agree with the commenter's description of how the current matching occurs for assessments. As previously discussed, we are aware that admissions and discharges are currently coded for purposes of the SNF QRP in a way that might conflict with how stays will be captured under the new PDPM. We intend to revise the codes so that a Medicare stay is captured the same way for purposes of the SNF QRP and the PDPM.

Comment: One commenter stated concerns with the suggestion that CMS would monitor this interrupted stay policy for frequent readmission, particularly facilities where the readmissions occur just outside the 3-day window used as part of the proposed interrupted stay policy. The commenter stated that SNFs already are the most highly regulated and monitored profession in health care. They stated a new policy with additional scrutiny and risk increases provider burden. They pointed out that CMS has programs in place to monitor and penalize SNFs for rehospitalization. The commenter stated that the SNF Rehospitalization VPB Program reduces all SNF rates by 2 percent. The commenter further stated that SNFs may earn a portion of these funds back by keeping rehospitalization rates low. Also, the commenter pointed out that SNF performance on return to community and related quality measures under the SNF QRP are publicly reported. The commenter stated that SNFs that perform poorly on QRP measures are less likely to be included in Medicare Advantage Plan or Accountable Care Organization provider networks. Thus, the commenter concluded that heightened scrutiny for poor performance already is in place. They recommended that SNF re-admissions to hospitals under the existing program—presumably meaning the SNF Rehospitalization VPB Program—should serve as the monitoring tool. They stated that, as with the SNF VBP Program, QRP performance also will serve a monitoring tool. They added that poorly performing SNFs will be penalized by the market, so that no additional government action is needed.

Response: We acknowledge that these monitoring tools exist and will utilize these existing tools to the fullest extent possible, but will also monitor specifically for inappropriate behavior in the context of the interrupted stay policy and decide the appropriate form of administrative action for whatever behavior is identified.

Comment: One commenter stated that CMS should develop a policy specific to the interrupted stay and the calculation of group/concurrent minutes. An interrupted stay could prevent the individual therapy minutes from being provided, and therefore, result in exceeding the 25 percent threshold. For example, if a resident is admitted to a facility and receives 100 percent group therapy on Day 1 of their SNF stay, with the full intent to move the resident to individual therapy in the days that follow, and then an interrupted stay occurs on Day 2 of the resident's stay; what would be the resulting impact to the facility from the resident receiving over the allowed 25 percent group therapy?

Response: As noted in section V.E of this final rule, there currently is no penalty associated with the group and concurrent therapy limits; instead, providers will receive a non-fatal warning edit on the validation report. We stated that we would monitor and evaluate how group and concurrent therapy are used under PDPM and consider making future proposals to address abuses of this policy or flag providers for additional review should a provider be found to consistently exceed the threshold. That being said, in terms of calculating adherence with the concurrent and group therapy limit, such a calculation is, as described in section V.E. of this final rule, completed at the stay level. Therefore, in cases of an interrupted stay, the therapy minutes over the course of the entire stay, both before and after the interruption, would be used to calculate the proportion of therapy time furnished within a concurrent or group setting. We believe this is the fairest option, as to calculate the proportion of such minutes based on only one portion of the stay may unduly identify a given provider as having failed to adhere to the established limit only because that particular portion of the stay had a larger amount of a given therapy mode.

Comment: Several commenters pointed out a discrepancy in the Medicare days count in Example B in the FY 2019 SNF PPS proposed rule (83 FR 21069). Specifically, commenters highlighted that Example B states that the resident is discharged on day 7 and that “the first day of the second stay would be paid at the Day 8 per diem rates under that schedule.” This implies that if a SNF resident has an interrupted stay, for the purposes of determining day in the stay for the per diem payment, when the patient returns to the SNF after the interruption, the stay resumes on the next day of the stay. For example, if a SNF resident is on day 7 of a stay which is then interrupted, when the resident returns within a certain time frame the day in the stay would be day 8. If the resident is discharged on day 7 of the stay, the SNF would be unable to bill for this day, resulting in the beneficiary using only 6 of the Medicare days. This would be unfair for both the resident and the SNF. Commenters recommended that CMS clarify the policy so that providers are paid for the day when a resident leaves a SNF in the case of an interrupted stay. Commenters said that under the policy as proposed, providers would not be paid for the day the resident leaves the SNF and so would lose one day of reimbursement.

Response: We agree with commenters regarding this typographical error and that payments should resume at the rate of the day of discharge, rather than the day after discharge. In other words, if a SNF resident is on day 7 of a stay which is interrupted, when the resident is readmitted, the payment rate would resume at day 7, not day 8, as Example B incorrectly stated.

The day of discharge in an interrupted stay would not be counted against the beneficiary's count of 100 days of covered Part A care in a benefit period. SNFs are not currently paid for the day of discharge, even with an anticipated leave of absence, unless the patient returns to the SNF before midnight of the same day. We do not believe there is anything about the interrupted stay policy that warrants changing this.

Comment: Multiple commenters expressed general support for the interrupted stay policy as proposed. Commenters supported the implementation of a SNF interrupted stay policy that is consistent with the policies in other post-acute care settings. Commenters recognized that with the proposed changes under the PDPM, which include variable per diem payment adjustments that provide higher payments at the beginning of the stay, implementing an interrupted stay policy will be appropriate for SNFs. As a further point of support, commenters noted that under the current system, rates of discharge to institutions (such as acute hospital or emergency department) are monitored very closely. Commenters expected that the proposed interrupted stay policy would allow for short term discharges where medically necessary while allowing for appropriate payment across a patient's stay.

Response: We agree with the commenters that the PDPM will benefit from the interrupted stay policy proposed.

Accordingly, after considering the comments received, for the reasons discussed in the proposed rule and in this final rule, we are finalizing our proposed interrupted stay policy without modification, to be effective October 1, 2019 in conjunction with the implementation of the PDPM.

G. Relationship of the PDPM to Existing Skilled Nursing Facility Level of Care Criteria

As discussed in the proposed rule (83 FR 21070), the establishment of the SNF PPS did not change Medicare's fundamental requirements for SNF coverage. However, because the case-mix adjustment aspect of the SNF PPS has been based, in part, on the beneficiary's need for skilled nursing care and therapy, we have coordinated claims review procedures with the existing resident assessment process and case-mix classification system. This approach includes an administrative presumption that utilizes a beneficiary's initial classification in one of the upper 52 RUGs of the 66-group RUG-IV system to assist in making certain SNF level of care determinations.

As further discussed below, in the proposed rule (83 FR 21070-72), we proposed to adopt a similar approach under the PDPM effective October 1, 2019, by retaining an administrative presumption mechanism that would utilize the initial assignment of one of the case-mix classifiers that we designate for this purpose to assist in making certain SNF level of care determinations. This designation would reflect an administrative presumption under the PDPM that beneficiaries who are correctly assigned one of the designated case-mix classifiers on the initial 5-day, Medicare-required assessment are automatically classified as meeting the SNF level of care definition up to and including the assessment reference date on the 5-day Medicare required assessment.

We stated that, as under the existing RUG-IV administrative presumption, a beneficiary who is not assigned one of the designated classifiers would not automatically be classified as either meeting or not meeting the level of care definition, but instead would receive an individual level of care determination using the existing administrative criteria. We stated that the use of the administrative presumption reflects the strong likelihood that those beneficiaries who are assigned one of the designated classifiers during the immediate post-hospital period require a covered level of care, which would be less likely for other beneficiaries.

In the ANPRM (82 FR 21007), we discussed some potential adaptations of the RUG-IV model's administrative presumption to accommodate specific features of the RCS-I model, including the possible designation of the following case-mix classifiers for purposes of the administrative presumption:

• Continued designation of the same nursing (non-rehabilitation) groups that currently comprise the Extensive Services, Special Care High, Special Care Low, and Clinically Complex categories under RUG-IV, as those groups would crosswalk directly from RUG-IV to the RCS-I model we were considering;
• In addition, designation of the most intensive functional score (14 to 18) under the RCS-I model's combined PT/OT component, as well as the uppermost comorbidity score (11+) under its NTA component.

In response, a number of comments expressed concern that the possible adaptations of the presumption could adversely affect access to care for some beneficiaries. Others questioned whether using the PT/OT component's highest functional score bin (14 to 18) as a trigger for the presumption would be appropriate, inasmuch as the residents that typically require the most therapy are those with only moderate functional impairments. In addition, commenters questioned the discussion's inclusion of the RCS-I model's NTA component as a possible classifier under the presumption, as well as its omission of RCS-I's SLP component.

Regarding the commenters' concerns about access to care, we noted in the proposed rule that we have indicated in the ANPRM and in previous rulemaking that the actual purpose of the level of care presumption has always been to afford a streamlined and simplified administrative procedure for readily identifying those beneficiaries with the greatest likelihood of meeting the level of care criteria; however, we have also emphasized that in focusing on such beneficiaries, this approach in no way serves to disadvantage other beneficiaries who may also meet the level of care criteria. As we noted in the ANPRM, an individual beneficiary's inability to qualify for the administrative presumption would not in itself serve to disqualify that resident from receiving SNF coverage. While such residents are not automatically presumed to require a skilled level of care, neither are they automatically classified as requiring nonskilled care; rather, any resident who does not qualify for the presumption would instead receive an individual level of care determination using the existing administrative criteria (82 FR 21007). As we further explained in the FY 2016 SNF PPS final rule (80 FR 46406, August 4, 2015), structuring the presumption in this manner serves specifically to ensure that the presumption does not disadvantage such residents, by providing them with an individualized level of care determination that fully considers all pertinent factors.

As for concerns about the appropriateness of certain classifiers, including the possible use of the PT/OT component's highest functional score bin (14 to 18) for this purpose under RCS-I, we noted in the proposed rule that the case-mix classification model for PT and OT that we were proposing in connection with the PDPM would essentially reconfigure the PT/OT component from the RCS-I model. As discussed in section V.D.3.b. of the proposed rule, the proposed PDPM would divide the RCS-I model's combined PT/OT component into two separate case-mix adjusted components, under which each resident would be assigned separate case-mix groups for PT and OT payment. Those groups would classify residents based on clinical category and function score, the two resident characteristics shown to be most predictive of PT and OT utilization.

The proposed rule's discussion also cited section III.B.4. of the ANPRM (“Variable Per Diem Adjustment Factors and Payment Schedule”), as well as section V.D.4. of the proposed rule itself, which indicated that our initial analyses revealed that in contrast to the SLP component—where per diem costs remain relatively constant over time—costs for the PT, OT, and NTA components typically are highest at the outset and then decline over the course of the stay. The proposed rule noted that our research to date continues to show a strong correlation between the dependent variables used for the proposed separate PT and OT components and a similarity in predictors, in that the associated costs for both therapy disciplines remain highest in the initial (and typically most intensive) portion of the SNF stay. We stated that this heightened resource intensity during the initial part of the SNF stay under the PT, OT, and NTA components, in turn, more closely reflects the distinctive utilization patterns that served as the original foundation for the level of care presumption itself—that is, the tendency as noted in the FY 2000 SNF PPS final rule for SNF stays to be at their most intensive and unstable immediately following admission as justifying a presumption of coverage at the very outset of the SNF stay (64 FR 41667, July 30, 1999). We also stated that we believe this would make the most intensive classifiers within each of these three proposed components well-suited to serve as clinical proxies for identifying those beneficiaries with the most intensive care needs and greatest likelihood of requiring an SNF level of care.

Accordingly, for purposes of the administrative presumption under the proposed PDPM, we proposed to continue utilizing the same designated nursing (non-rehabilitation) categories under the PDPM as had been used to date under RUG-IV. We noted that the most direct crosswalk between the existing RUG-IV model and the proposed PDPM would involve nursing services, for which, under the proposed PDPM, each resident would continue to be classified into one of the groups that fall within the existing non-rehabilitation RUG-IV categories. (As explained in section V.D.3.d. of the proposed rule, while the PDPM would streamline the total number of nursing case-mix groups from the current 43 under RUG-IV down to 25 through the consolidation of similar groups within individual categories, the overall number and structure of the nursing categories themselves would remain the same.) Under our proposal, effective in conjunction with the proposed implementation of the PDPM (that is, as of October 1, 2019), we stated that the administrative presumption would apply to those groups encompassed by the same nursing categories as have been designated for this purpose under the existing RUG-IV model:

• Extensive Services;
• Special Care High;
• Special Care Low; and,
• Clinically Complex.

In addition, along with the continued use of the RUG-IV nursing categories above, we also proposed to apply the administrative presumption using those other classifiers under the proposed PDPM that we identified as relating the most directly to identifying a patient's need for skilled care at the outset of the SNF stay. We proposed to designate such classifiers for this purpose based on their ability to fulfill the administrative presumption's role as described in the FY 2000 SNF PPS final rule (64 FR 41668 through 41669, July 30, 1999)—that is, to identify those situations that involve a high probability of the need for skilled care when taken in combination with the characteristic tendency for an SNF resident's condition to be at its most unstable and intensive state at the outset of the SNF stay.

Specifically, we additionally proposed to designate for this purpose proposed PT and OT case-mix groups TB, TC, TD, TF, and TG, the groups displayed in Table 21 of the proposed rule that collectively accounted for the five highest case-mix indexes for PT, as well as for OT and, thus, would consistently be associated with the most resource-intensive care across both of these therapy disciplines. We also proposed to designate the uppermost comorbidity group under the NTA component, in the belief that this particular classifier would serve to identify those cases that are the most likely to involve the kind of complex medication regimen (for example, a highly intensive drug requiring specialized expertise to administer, or an exceptionally large and diverse assortment of medications posing an increased risk of adverse drug interactions) that would require skilled oversight to manage safely and effectively. As discussed in section V.D.3.e of this final rule, the specific value assigned to the NTA component's uppermost comorbidity score (which was 11+ under the RCS-I model and is 12+ under PDPM) might change once again in the future if the NTA score bins are reconfigured to reflect changes in the resident population and care practices over time.

We further explained that under this proposed approach, those residents not classifying into a case-mix group in one of the designated nursing RUG categories under the proposed PDPM on the initial, 5-day Medicare-required assessment could nonetheless still qualify for the administrative presumption on that assessment by being placed in one of the designated case-mix groups for either the PT or OT components, or by receiving the uppermost comorbidity score under the NTA component. We indicated that these particular case-mix classifiers would appropriately serve to fulfill the administrative presumption's role of identifying those cases with the highest probability of requiring an SNF level of care throughout the initial portion of the SNF stay. We additionally noted that in order to help improve the accuracy of these newly-designated groups in serving this function, we would continue to review the new designations going forward and may make further adjustments to the proposed designations over time as we gain actual operating experience under the new classification model. As discussed above, this proposed administrative presumption mechanism would take effect October 1, 2019 in conjunction with the proposed PDPM itself. We invited comments on our proposed administrative presumption mechanism under the proposed PDPM.

Commenters submitted the following comments related to the proposed rule's discussion on our proposed administrative presumption mechanism under the proposed PDPM. A discussion of these comments, along with our responses, appears below.

Comment: One commenter mistakenly assumed that under the PDPM, the administrative presumption would change from its current use of the initial, 5-day Medicare-required assessment to using the initial MDS assessment (that is, the OBRA-required Admission assessment) instead, and expressed concern that the timeframes associated with the latter would be inappropriate for this purpose.

Response: We note that consistent with the discussion in the proposed rule (83 FR 21070-21072), the presumption's current use of the initial, 5-day Medicare-required assessment will, in fact, continue under the PDPM.

Comment: Several commenters urged us to designate other therapy groups, in addition to those set forth in the proposed rule, as appropriately serving to identify a level of acuity that would qualify for the presumption. They equated the omission of a given case-mix classifier from the presumption with a restriction on access and coverage, and characterized the individual level of care determinations that SNFs would routinely conduct absent the presumption as an added administrative burden. The commenters specifically cited as a concern the proposed rule's omission of any PT and OT groups for non-orthopedic conditions, as well as of any groups at all from the SLP component. One commenter took issue with the proposed rule's stated rationale for the omission of SLP (that is, that such services, unlike PT and OT, remain relatively constant over time and are not concentrated in the initial portion of the stay), noting that nursing services similarly do not taper off over the course of the stay and yet have been utilized under the presumption ever since its inception. The commenter pointed out that as with the other components, it is possible to identify individual groups within the SLP component that have relatively high service intensity. Along with the groups from the PT and OT components that were already proposed for designation under the presumption, the commenter recommended the designation of several additional PT and OT groups (that is, TA, TE, TJ, TK, TN, and TO), as well as a number of groups (that is, SC, SE, SF, SH, SI, SJ, SK, and SL) from the SLP component, and presented these particular groups as reflecting the most intensive therapy needs within their respective clinical categories. The commenter also suggested that the proposed designation of the NTA's uppermost comorbidity group might not actually be necessary, as anyone assigned to that group would likely qualify for the presumption already, based on their classification under the nursing component. Another commenter recommended that all of the PT and OT groups in the Other Orthopedic category should be designated for use under the presumption, and pointed out that under PDPM, the NTA component's uppermost comorbidity score is actually 12+ rather than 11+ as indicated in the proposed rule.

Response: We agree with the commenters that the administrative presumption should encompass all of the groups that serve to fulfill the basic purpose of this provision—that is, readily identifying those beneficiaries with the greatest likelihood of meeting the level of care criteria. With one exception, we also concur with the commenters' analysis that the additional therapy groups recommended for designation under the presumption would appropriately serve to reflect the most intensive therapy needs within their respective clinical categories, as evidenced by the relatively high CMI that is associated with each of the recommended groups. However, regarding the recommendation to designate all PT and OT groups in the Other Orthopedic category, we note that one such group, TH, has a significantly lower CMI than all of the other recommended groups and, thus, is not being selected for designation under the presumption. Accordingly, we are adopting the remainder of the commenters' recommendations regarding the designation of additional groups from the PT and OT components, as well as all of the recommended groups from the SLP component. In addition, we are finalizing as proposed the use of the designated classifiers from the nursing component along with the uppermost comorbidity score of the NTA component. Regarding the latter, we appreciate the comment pointing out that the specific value assigned to the NTA component's uppermost comorbidity score under the PDPM is, in fact, 12+ and not 11+ as incorrectly indicated in the proposed rule's discussion of the presumption. We also appreciate another commenter's concern that the proposed NTA classifier might in some instances prove redundant in relation to the nursing groups; however, because we believe, as stated above, that the presumption should encompass all appropriate classifiers, we are finalizing the use of this particular classifier as we believe this particular classifier would serve to identify those cases that are the most likely to involve the kind of complex medication regimen that would require skilled oversight to manage safely and effectively. We also will evaluate the use of this classifier in actual operation and confirm whether there are instances in which it appropriately serves this function independently of the nursing groups. As we indicated in the proposed rule (83 FR 21072) regarding the NTA and other components, we will continue to review the new designations going forward and make further adjustments over time as we gain actual operating experience under the new classification model.

However, we would also note in this context that we do not share and cannot support the view that would essentially equate a given case-mix classifier's non-designation under the administrative presumption with a restriction on access or a denial of SNF coverage, or an increase in administrative burden. SNF coverage ultimately is based not on whether a beneficiary is assigned one of the designated classifiers, but on whether the SNF level of care criteria are met. As further explained in the proposed rule (83 FR 21071), the purpose of the administrative presumption is solely to afford a streamlined and simplified administrative procedure for readily identifying those beneficiaries with the greatest likelihood of meeting the level of care criteria, which in no way serves to disadvantage other beneficiaries who may also meet the level of care criteria. In fact, far from creating an overall increase in administrative burden from the non-designated classifiers, we expect that the presumption's framework of streamlined and simplified initial determinations for the designated classifiers will actually serve to free up staff resources, which can then be used for assessing coverage in the other cases.

Accordingly, for the reasons set forth in the proposed rule and in this final rule, we are finalizing our proposed classifiers for purposes of applying the administrative presumption under the PDPM with the following modifications. As discussed above, we are adding the following PT and OT classifiers to those we proposed: TA, TE, TJ, TK, TN and TO. We are also adding the following 8 SLP classifiers: SC, SE, SF, SH, SI, SJ, SK, and SL. Thus, effective October 1, 2019, we are designating the classifiers shown below for purposes of the administrative presumption under the PDPM:

• The case-mix classifiers in the following nursing categories: Extensive Services, Special Care High, Special Care Low, and Clinically Complex;
• The following PT and OT groups: TA, TB, TC, TD, TE, TF, TG, TJ, TK, TN, and TO;
• The following SLP groups: SC, SE, SF, SH, SI, SJ, SK, and SL; and
• The NTA component's uppermost comorbidity group (which, as finalized in this final rule, is 12+).

H. Effect of PDPM on Temporary AIDS Add-On Payment

As discussed in section V.I. of the proposed rule (83 FR 21072) and also in section III.E. of the ANPRM (82 FR 21007), section 511(a) of the MMA amended section 1888(e)(12) of the Act to provide for a temporary increase of 128 percent in the PPS per diem payment for any SNF residents with Acquired Immune Deficiency Syndrome (AIDS), effective with services furnished on or after October 1, 2004. This special add-on for SNF residents with AIDS was intended to be of limited duration, as the MMA legislation specified that it was to remain in effect only until the Secretary certifies that there is an appropriate adjustment in the case mix to compensate for the increased costs associated with such residents.

The temporary add-on for SNF residents with AIDS is also discussed in Program Transmittal #160 (Change Request #3291), issued on April 30, 2004, which is available online at www.cms.gov/​transmittals/​downloads/​r160cp.pdf. In the SNF PPS final rule for FY 2010 (74 FR 40288, August 11, 2009), we did not address this certification in that final rule's implementation of the case-mix refinements for RUG-IV, thus allowing the add-on payment required by section 511 of the MMA to remain in effect for the time being.

In the House Ways and Means Committee Report that accompanied the MMA, the explanation of the MMA's temporary AIDS adjustment notes the following under Reason for Change: “According to prior work by the Urban Institute, AIDS patients have much higher costs than other patients in the same resource utilization groups in skilled nursing facilities. The adjustment is based on that data analysis” (H. Rep. No. 108-178, Part 2 at 221). The data analysis from that February 2001 Urban Institute study (entitled “Medicare Payments for Patients with HIV/AIDS in Skilled Nursing Facilities”), in turn, had been conducted under a Report to Congress mandated under a predecessor provision, section 105 of the BBRA. This earlier BBRA provision, which ultimately was superseded by the temporary AIDS add-on provision required by the MMA, had amended section 1888(e)(12) of the Act to provide for special consideration for facilities serving specialized patient populations (that is, those who are “immuno-compromised secondary to an infectious disease, with specific diagnoses as specified by the Secretary”).

As we noted in the ANPRM and in the proposed rule, at that point over a decade and a half had elapsed since the Urban Institute conducted its study on AIDS patients in SNFs, a period that has seen major advances in the state of medical practice in treating this condition. We stated that these advances have notably included the introduction of powerful new drugs and innovative prescription regimens that have dramatically improved the ability to manage the viral load (the amount of human immunodeficiency virus (HIV) in the blood). We noted that the decrease in viral load secondary to medications has contributed to a shift from intensive nursing services for AIDS-related illnesses to an increase in antiretroviral therapy. We further stated that this phenomenon, in turn, is reflected in our recent analysis of differences in SNF resource utilization, which indicates that while the overall historical disparity in costs between AIDS and non-AIDS patients has not entirely disappeared, that disparity is now far greater with regard to drugs than it is for nursing. Specifically, as explained in the proposed rule, NTA costs per day for residents with AIDS were 151 percent higher than those for other residents while the difference in wage-weighted nursing staff time between the two groups was only 19 percent, as discussed in section 3.8.3. of the SNF PMR technical report (available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html), which the ANPRM referenced for further information on the underlying data analysis (82 FR 21007 through 21008). In the ANPRM, we also described how the RCS-I model would account for those NTA costs, including drugs, which specifically relate to residents with AIDS (82 FR 20997 through 20999). We additionally discussed in the ANPRM the possibility of making a specific 19 percent AIDS adjustment as part of the case-mix adjustment of the nursing component (82 FR 20995 through 20997). We further expressed our belief in the ANPRM (82 FR 21008) that when taken collectively, these adjustments would appropriately serve to justify issuing the certification prescribed under section 511(a) of the MMA, which would permit the MMA's existing, temporary AIDS add-on to be replaced by a permanent adjustment in the case mix that appropriately compensates for the increased costs associated with these residents.

As discussed in the proposed rule, in response to the ANPRM, we received comments expressing concerns that a projected 40 percent drop in overall payments for SNF residents with AIDS under the RCS-I model could adversely affect access to care for this patient population. Regarding those concerns, we noted in the proposed rule that the special add-on for SNF residents with AIDS itself was never meant to be permanent, and does not serve as a specific benchmark for use in establishing either the appropriate methodology or level of payment for this patient population. Rather, we stated that, as discussed in the ANPRM, it was designed to be only a temporary measure, representing a general approximation that reflected the current state of research and clinical practice at the time (82 FR 21007 through 21008). As such, we stated that the special add-on would not account for the significant changes in the care and treatment of this condition that have occurred over the intervening years. We further noted that as a simple across-the-board multiplier, the MMA adjustment by its very nature is not accurately targeted at those particular rate components that actually account for the disparity in cost between AIDS patients and others.

As discussed in section V.D.3.e. of the proposed rule (83 FR 21058), our updated investigations into the adequacy of payments under the proposed PDPM for residents with HIV/AIDS indicated that the four proposed ancillary payment components (PT, OT, SLP, and NTA) would adequately reimburse ancillary costs associated with HIV/AIDS residents (see section 3.8.2. of the SNF PDPM technical report, available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html). Therefore, we stated that we believe it would be appropriate to issue the prescribed certification under section 511(a) of the MMA on the basis of the proposed PDPM's ancillary case-mix adjustment alone, as effectively providing the required appropriate adjustment in the case mix to compensate for the increased costs associated with such residents. However, to further ensure that the proposed PDPM would account as fully as possible for any remaining disparity with regard to nursing costs, as discussed in section V.D.3.d. of the proposed rule (83 FR 21055), we additionally proposed to include a specific AIDS adjustment as part of the case-mix adjustment of the nursing component. As discussed in section V.D.3.d. of the proposed rule, we used the STRIVE data to quantify the effects of HIV/AIDS diagnosis on nursing resource use. Regression analyses found that wage-weighted nursing staff time is 18 percent higher for residents with HIV/AIDS, controlling for the non-rehabilitation RUG of the resident. We noted that this figure is slightly lower than the 19 percent increase in wage-weighted nursing staff time reported in the ANPRM and the SNF PRM technical report because the updated investigation uses a FY 2017 study population and is based on the PDPM case-mix groups, while the earlier analysis was based on a FY 2014 study population and the RCS-I case-mix groups. More information on this analysis can be found in section 3.8.2. of the SNF PDPM technical report (available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html). Thus, we proposed an 18 percent increase in payment for the nursing component for residents with HIV/AIDS under the proposed PDPM to account for the increased nursing costs for such residents. We stated that similar to the proposed NTA adjustment for residents with HIV/AIDS, this adjustment would be identified by ICD-10-CM code B20 on the SNF claim and would be processed through the PRICER software used by CMS to set the appropriate payment rate for a resident's SNF stay. We also explained (83 FR 21073) that the 18 percent adjustment would be applied to the unadjusted base rate for the nursing component, and then this amount would be further case-mix adjusted per the resident's PDPM nursing classification.

In the proposed rule, we expressed the belief that when taken collectively, these adjustments under the proposed PDPM would appropriately serve to justify issuing the certification prescribed under section 511(a) of the MMA effective with the proposed conversion to the PDPM on October 1, 2019, thus permitting the MMA's existing, temporary AIDS add-on to be replaced by a permanent adjustment in the case mix (as proposed under the PDPM) that appropriately compensates for the increased costs associated with these residents, and we invited comments on this proposal. At the same time, we acknowledged that even with an accurately targeted model that compensates for the increased costs of SNF residents with AIDS, an abrupt conversion to an altogether different payment methodology might nevertheless be potentially disruptive for facilities, particularly those that serve a significant number of patients with AIDS and may have become accustomed to operating under the existing payment methodology for those patients. Accordingly, we also invited comments on possible ways to help mitigate any potential disruption stemming from the proposed replacement of the special add-on payment with the permanent case-mix adjustments for SNF residents with AIDS under the proposed PDPM.

Commenters submitted the following comments related to the proposed rule's discussion on the Effect of the Proposed PDPM on Temporary AIDS Add-on Payment. A discussion of these comments, along with our responses, appears below.

Comment: Several commenters expressed concern about the adequacy of payments under the PDPM for SNF residents with AIDS, once again citing the projected decrease in payments relative to the RUG-IV model (which currently includes the MMA's temporary 128 percent adjustment for such residents). One commenter specifically questioned the adequacy of the PDPM's NTA component in addressing the drug costs of AIDS patients, and cited a 2017 MedPAC report that characterized the SNF PPS's NTA payments as poorly targeted.

Response: We note that as with the previous comments on the corresponding aspect of the ANPRM, most of the commenters' concerns in this area stemmed from comparing the projected payment levels under the PDPM to those under the existing RUG-IV model's temporary 128 percent AIDS adjustment, and focused specifically on the contrast in payment levels between the two models. However, as noted above and explained in the proposed rule (83 FR 21073), it is not appropriate to use the MMA adjustment as a benchmark in assessing the accuracy of the PDPM's payment methodology, as the special add-on for SNF residents with AIDS itself was never meant to be permanent, and does not serve as a specific benchmark for use in establishing either the appropriate methodology or level of payment for this patient population. Rather, it was designed to be only a temporary measure, representing a general approximation that reflected the current state of research and clinical practice at the time. As such, the special add-on would not account for the significant changes in the care and treatment of this condition that have occurred over the intervening years. Moreover, as a simple across-the-board multiplier, the MMA adjustment by its very nature is not accurately targeted at those particular rate components that actually account for the disparity in cost between AIDS patients and others.

Regarding that final point about the imprecision of applying an across-the-board multiplier in this context, we further noted in the proposed rule (83 FR 20180) that our research found that HIV/AIDS was associated with a negative and statistically significant decrease in PT, OT and SLP costs per day. This means inherently that, to the extent that the existing add-on is applied against the full SNF PPS per diem payment, the magnitude of the add-on payment increases with increases in therapy payment, which conflicts with the data described above regarding the relationship between therapy costs and the presence of an AIDS diagnosis. As a result, maintaining the current add-on would create an inconsistency between how SNF payments would be made and the data regarding AIDS diagnoses and resident therapy costs.

Furthermore, to the extent that the RUG-IV model's case-mix classification system may have included inherent incentives toward the overprovision of therapy services, the MMA adjustment's operation as an across-the-board multiplier would actually serve to magnify the effects of any such incentives, by inflating the resulting payment levels even further beyond the patient's actual therapy care needs. In this context, we note that the specific standard prescribed for the Secretary's required certification under section 511(a) of the MMA is that “. . . there is an appropriate adjustment in the case mix . . . to compensate for the increased costs” associated with SNF residents with AIDS. As set forth in the proposed rule, we believe that the PDPM's payment methodology for patients with AIDS clearly meets this statutory standard of appropriately accounting for the actual costs incurred in caring for such patients. In fact, we believe it provides a far more accurate and current accounting of those costs than the temporary MMA adjustment that it would replace, which represents only a very broad approximation that was developed at a time when the treatment regimens for this condition differed dramatically from what they are currently. Finally, it is worth noting that the cited 2017 MedPAC report, which characterized the SNF PPS's NTA payments as poorly targeted, reflected that the SNF PPS has always included NTA costs within its nursing component rather than accounting for them separately, and the longstanding concerns about that approach were, in fact, the very impetus behind our development of a separate component for NTA costs under the PDPM.

Accordingly, for the reasons discussed in the proposed rule and in this final rule, the Secretary is certifying that there is an appropriate adjustment in the PDPM to compensate for the increased costs associated with residents with AIDS, and thus we are finalizing our proposal without modification to replace the temporary MMA add-on with the PDPM's permanent adjustment in the case mix that appropriately accounts for the increased costs of patients with AIDS, effective with the conversion to the PDPM on October 1, 2019.

I. Potential Impacts of Implementing the PDPM and Parity Adjustment

This section outlines the projected impacts of implementing the PDPM effective October 1, 2019 under the SNF PPS and the related policies finalized in sections V of this final rule that would be effective in conjunction with the PDPM. This impact analysis makes a series of assumptions, as described below (as were discussed in the proposed rule (83 FR 21073 through 21080)). First, the impacts presented here assume consistent provider behavior in terms of how care is provided under RUG-IV and how care might be provided under the PDPM, as we do not make any attempt to anticipate or predict provider reactions to the implementation of the PDPM. That being said, we acknowledge the possibility that implementing the PDPM could substantially affect resident care and coding behaviors. Most notably, based on the concerns raised during a number of TEPs, we acknowledge the possibility that, as therapy payments under the PDPM would not have the same connection to service provision as they do under RUG-IV, it is possible that some providers may choose to reduce their provision of therapy services to increase margins under the PDPM. However, we do not have any basis on which to assume the approximate nature or magnitude of these behavioral responses, nor have we received any sufficiently specific guidance on the likely nature or magnitude of behavioral responses from ANPRM commenters, TEP panelists, or other sources of feedback. As a result, lacking an appropriate basis to forecast behavioral responses, we do not adjust our analyses of resident and provider impacts discussed in this section for projected changes in provider behavior. However, we do intend to monitor behavior which may occur in response to the implementation of PDPM, and may consider proposing policies in the future to address such behaviors to the extent determined appropriate. Additionally, we acknowledge that a number of states utilize some form of the RUG-IV case-mix classification system as part of their Medicaid programs and that any change in Medicare policy can have an impact on state programs. Again, we do not have any basis on which to assume the approximate nature or magnitude of these responses, for the same reasons cited above. Additionally, we do not expect impacts on state Medicaid programs resulting from PDPM implementation to have a notable impact on payments for Medicare-covered SNF stays, which are the basis for the impact analyses discussed in this section. Therefore, we do not consider possible changes to state Medicaid programs when conducting these analyses. We invited comments on our assumptions that behavior would remain unchanged under the proposed PDPM and that changes in state Medicaid programs resulting from PDPM implementation would not have a notable impact on payments for Medicare-covered SNF stays. We also invited comment on the impact of these policy proposals on state Medicaid programs. These comments are addressed among the general comments in section V.A. of this final rule.

As with prior system transitions, we proposed to implement the PDPM case-mix system, along with the other policy changes discussed throughout this section, in a budget neutral manner through application of a parity adjustment to the case-mix weights under the proposed PDPM, as further discussed below. We proposed to implement the PDPM in a budget neutral manner because, as with prior system transitions, in proposing changes to the case-mix methodology, we do not intend to change the aggregate amount of Medicare payments to SNFs. Rather, we aim to utilize a case-mix methodology to classify residents in such a manner as to best ensure that payments made for specific residents are an accurate reflection of resource utilization without introducing potential incentives which could encourage inappropriate care delivery, as we believe may exist under the current case-mix methodology. Therefore, the impact analysis presented here assumes implementation of these proposed changes in a budget neutral manner. We invited comments on the proposal, as further discussed below, to implement the PDPM in a budget neutral manner. In addition, we solicited comment on whether it would be appropriate to implement the proposed PDPM in a manner that is not budget neutral.

As discussed above, the impact analysis presented here assumes implementation of these changes in a budget neutral manner without a behavioral change. The prior sections describe how case-mix weights are set to reflect relative resource use for each case-mix group. We stated in the proposed rule that the proposed PDPM payment before application of a parity adjustment would be calculated using the unadjusted CMI for each component, the variable per diem payment adjustment schedule, the unadjusted urban and rural federal per diem rates shown in Tables 12 and 13, the labor-related share, and the geographic wage indexes. In applying a parity adjustment to the case-mix weights, we stated in the proposed rule that we would maintain the relative value of each CMI but would multiply every CMI by a ratio to achieve parity in overall SNF PPS payments under the PDPM and under the RUG-IV case-mix model. The parity adjustment multiplier was calculated through the following steps, as described in the proposed rule (83 FR 21074). First, we calculated RUG-IV total payment. Total RUG-IV payments were calculated by adding total allowed amounts across all FY 2017 SNF claims. The total allowed amount in the study population was the summation of Medicare and non-Medicare payments for Medicare-covered days. More specifically, it was the sum of Medicare claim payment amount, National Claim History (NCH) primary payer claim paid amount, NCH beneficiary inpatient deductible amount, NCH beneficiary Part A coinsurance liability amount, and NCH beneficiary blood deductible liability amount. Second, we calculated what total payment would have been under the proposed PDPM in FY 2017 before application of the parity adjustment. Total estimated payments under PDPM were calculated by summing the predicted payment for each case-mix component together for all FY 2017 SNF stays. This represented the total allowed amount if PDPM had been in place in FY 2017. Total estimated FY 2017 payments under the PDPM were calculated using resident information from FY 2017 SNF claims, the MDS assessment, and other Medicare claims, as well as the unadjusted CMI for each component, the variable per diem payment adjustment schedule, the unadjusted urban and rural federal per diem rates shown in Tables 12 and 13, the labor-related share, and the geographic wage indexes. After calculating total actual RUG-IV payments and total estimated case-mix-related PDPM payments, we subtracted non-case-mix component payments from total RUG-IV payments, as this component does not change across systems. This subtraction did not include the temporary add-on for residents with HIV/AIDS in the RUG-IV system, which PDPM replaces with additional payments for residents with HIV/AIDS through the NTA and nursing components (as discussed in section V.I. of the proposed rule and section V.H. of this final rule). By retaining the portion of non-case-mix component payments associated with the temporary HIV/AIDS add-on in total RUG-IV payments, all payments associated with the add-on under RUG-IV were re-allocated to the case-mix-adjusted components in PDPM. This was appropriate because, as discussed, under the PDPM, additional payments for residents with HIV/AIDS are made exclusively through the case-mix-adjusted components (that is, the nursing and NTA components). Lastly, in calculating budget neutrality, we set total estimated case-mix-related payment under PDPM such that it equals total allowable Medicare payments under RUG-IV. To do this, we divided the remaining total RUG-IV payments over the remaining total estimated PDPM payments prior to the parity adjustment. This division yielded a ratio (parity adjustment) of 1.46 by which the PDPM CMIs were multiplied so that total estimated payments under the PDPM would be equal to total actual payments under RUG-IV, assuming no changes in the population, provider behavior, and coding. We stated in the proposed rule that, if this parity adjustment had not been applied, total estimated payments under the PDPM would be 46 percent lower than total actual payments under RUG-IV, therefore the implementation of the PDPM would not be budget neutral. More details regarding this calculation and analysis are described in section 3.11.2. of the SNF PDPM technical report (available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html). The impact analysis presented in this section (and in the proposed rule) focuses on how payments under the PDPM would be re-allocated across different resident groups and among different facility types, assuming implementation in a budget neutral manner.

The projected resident-level impacts are presented in Table 37. The first column identifies different resident subpopulations and the second column shows what percent of SNF stays in FY 2017 are represented by the given subpopulation. The third column shows the projected change in total payments for residents in a given subpopulation, represented as a percentage change in actual FY 2017 payments made for that subpopulation under RUG-IV versus estimated payments which would have been made to that subpopulation in FY 2017 had the PDPM been in place. Total RUG-IV payments are calculated by adding total allowed amounts across all FY 2017 SNF claims associated with a resident subpopulation. The total allowed amount in the study population is the summation of Medicare and non-Medicare payments for Medicare-covered days. More specifically, it is the summation of Medicare claim payment amount, NCH primary payer claim paid amount, NCH beneficiary inpatient deductible amount, NCH beneficiary Part A coinsurance liability amount, and NCH beneficiary blood deductible liability amount. Payments corresponding to the non-case-mix component are subtracted from the RUG-IV total payments, not including the portion of non-case-mix payments corresponding to the temporary add-on for residents with HIV/AIDS. Total estimated payments under PDPM are calculated by summing the predicted payment for each case-mix component together for all FY 2017 SNF stays associated with a resident subpopulation. Positive changes in this column represent a projected positive shift in payments for that subpopulation under the PDPM, while negative changes in this column represent projected negative shifts in payment for that subpopulation. More information on the construction of current payments under RUG-IV and payments under the PDPM for purposes of this impact analysis can be found in section 3.12. of the SNF PDPM technical report (available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html). Based on the data presented in Table 37, we observe that the most significant shift in payments created by implementation of the PDPM would be to redirect payments away from residents who are receiving very high amounts of therapy under the current SNF PPS, which strongly incentivizes the provision of therapy, to residents with more complex clinical needs. For example, we project that for residents whose most common therapy level is RU (ultra-high therapy)—the highest therapy level, there would be a reduction in associated payments of 8.4 percent, while payments for residents currently classified as non-rehabilitation would increase by 50.5 percent. Other resident types for which there may be higher relative payments under the PDPM are: Residents who have high NTA costs, receive extensive services, are dually enrolled in Medicare and Medicaid, use IV medication, have ESRD, diabetes, or a wound infection, receive amputation/prosthesis care, and/or have longer prior inpatient stays. Additionally, we received several comments in response to the 2017 ANPRM requesting that we estimate the impact of RCS-1 on the following potentially vulnerable subpopulations: Residents with addictions, bleeding disorders, behavioral issues, chronic neurological conditions, and bariatric care. In response to these comments, we added these subpopulations to our PDPM impact analysis. Table 37 shows that the PDPM is projected to increase the proportion of total payment associated with each of those subpopulations.

Table 37—PDPM Impact Analysis, Resident-Level

Resident characteristics	Percent of stays	Percent change
All Stays	100.0	0.0
Sex:
Female	60.3	−0.8
Male	39.7	1.2
Age:
Below 65 years	10.3	7.2
65-74 years	24.1	3.1
75-84 years	32.5	−0.4
85-89 years	17.6	−3.1
Over 90 years	15.6	−4.3
Race/Ethnicity:
White	83.8	−0.2
Black	11.2	0.8
Hispanic	1.7	0.9
Asian	1.3	−0.6
Native American	0.5	7.1
Other or Unknown	1.5	0.8
Medicare/Medicaid Dual Status:
Dually Enrolled	34.7	3.3
Not Dually Enrolled	65.3	−2.1
Original Reason for Medicare Enrollment:
Aged	74.6	−1.7
Disabled	24.5	4.8
ESRD	0.9	10.5
Utilization Days:
1-15 days	35.4	13.7
16-30 days	33.8	0.0
31+ days	30.9	−2.5
Utilization Days = 100:
No	98.4	0.1
Yes	1.6	−1.9
Length of Prior Inpatient Stay:
0-2 days	2.2	1.3
3 days	22.5	−3.3
4-30 days	73.6	0.7
31+ days	1.7	6.7
Most Common Therapy Level:
RU	58.4	−8.4
RV	22.4	11.4
RH	6.8	27.4
RM	3.3	41.1
RL	0.1	67.5
Non-Rehab	9.1	50.5
Number of Therapy Disciplines Used:
0	2.3	63.1
1	2.4	44.2
2	51.6	1.6
3	43.7	−3.1
Physical Therapy Utilization:
No	3.7	50.9
Yes	96.3	−0.7
Occupational Therapy Utilization:
No	4.5	47.7
Yes	95.5	−0.8
Speech Language Pathology Utilization:
No	55.0	2.8
Yes	45.0	−2.5
Therapy Utilization:
PT+OT+SLP	43.7	−3.1
PT+OT Only	50.8	1.3
PT+SLP Only	0.4	27.3
OT+SLP Only	0.4	30.1
PT Only	1.3	41.3
OT Only	0.6	47.9
SLP Only	0.5	46.8
Non-Therapy	2.3	63.1
NTA Costs ($):
0-10	13.7	−3.5
10-50	44.5	−3.2
50-150	32.2	4.2
150+	9.6	18.7
NTA Comorbidity Score:
0	23.5	−10.4
1-2	30.5	−4.7
3-5	31.0	4.0
6-8	9.9	15.0
9-11	3.6	24.4
12+	1.4	27.2
Extensive Services Level:
Tracheostomy and Ventilator/Respirator	0.3	22.2
Tracheostomy or Ventilator/Respirator	0.6	7.3
Infection Isolation	1.1	9.1
Neither	98.0	−0.3
CFS Level:
Cognitively Intact	58.5	−0.3
Mildly Impaired	20.7	−0.2
Moderately Impaired	16.8	−0.7
Severely Impaired	3.9	8.8
Clinical Category:
Acute Infections	6.5	3.4
Acute Neurologic	6.4	−3.7
Cancer	4.6	−3.2
Cardiovascular and Coagulations	9.8	0.5
Major Joint Replacement or Spinal Surgery	8.6	−2.1
Medical Management	30.4	0.0
Non-Orthopedic Surgery	10.8	5.7
Non-Surgical Orthopedic/Musculoskeletal	5.9	−6.1
Orthopedic Surgery (Except Major Joint Replacement or Spinal Surgery)	8.9	−2.4
Pulmonary	8.1	5.4
Level of Complications in MS-DRG of Prior Inpatient Stay:
No Complication	35.8	−3.1
CC/MCC	64.2	1.7
Stroke:
No	90.9	0.0
Yes	9.1	0.3
HIV/AIDS:
No	99.7	0.3
Yes	0.3	−40.5
IV Medication:
No	91.7	−2.1
Yes	8.3	23.5
Diabetes:
No	64.0	−3.0
Yes	36.0	5.4
Wound Infection:
No	98.9	−0.3
Yes	1.1	22.2
Amputation/Prosthesis Care:
No	100.0	0.0
Yes	0.0	6.4
Presence of Dementia:
No	70.9	0.5
Yes	29.1	−1.2
MDS Alzheimer's:
No	95.2	0.0
Yes	4.8	−0.3
Unknown	0.0	5.0
Presence of Addictions:
No	94.6	−0.1
Yes	5.4	1.8
Presence of Bleeding Disorders:
No	90.9	−0.1
Yes	9.1	1.5
Presence of Behavioral Issues:
No	53.1	−0.9
Yes	46.9	1.0
Presence of Chronic Neurological Conditions:
No	74.4	−0.2
Yes	25.6	0.6
Presence of Bariatric Care:
No	91.3	−0.6
Yes	8.7	6.5

The projected provider-level impacts are presented in Table 38. The first column identifies different facility subpopulations and the second column shows what percentage of SNFs in FY 2017 are represented by the given subpopulation. The third column shows the projected change in total payments for facilities in a given subpopulation, represented as a percentage change in actual FY 2017 payments made for that subpopulation under RUG-IV versus estimated payments which would have been made to that subpopulation in FY 2017 had the PDPM been in place. Total RUG-IV payments are calculated by adding total allowed amounts across all FY 2017 SNF claims associated with a facility subpopulation. The total allowed amount in the study population is the summation of Medicare and non-Medicare payments for Medicare-covered days. More specifically, it is the summation of Medicare claim payment amount, NCH primary payer claim paid amount, NCH beneficiary inpatient deductible amount, NCH beneficiary Part A coinsurance liability amount, and NCH beneficiary blood deductible liability amount. Payments corresponding to the non-case-mix component are subtracted from the RUG-IV total payments, not including the portion of non-case-mix payments corresponding to the temporary add-on for residents with HIV/AIDS. Total estimated payments under PDPM are calculated by summing the predicted payment for each case-mix component together for all FY 2017 SNF stays associated with a facility subpopulation. Positive changes in this column represent a projected positive shift in payments for that subpopulation under the PDPM, while negative changes in this column represent projected negative shifts in payment for that subpopulation. More information on the construction of current payments under RUG-IV and payments under the PDPM for purposes of this impact analysis can be found in section 3.12. of the SNF PDPM technical report (available at https://www.cms.gov/​Medicare/​Medicare-Fee-for-Service-Payment/​SNFPPS/​therapyresearch.html). Based on the data presented in Table 38, we observe that the most significant shift in Medicare payments created by implementation of the PDPM would be from facilities with a high proportion of rehabilitation residents (particularly facilities with high proportions of Ultra-High Rehabilitation residents) to facilities with high proportions of non-rehabilitation residents. We project that payments to facilities that bill 0 to 10 percent of utilization days as RU (ultra-high rehabilitation) would increase an estimated 27.6 percent under the PDPM while facilities that bill 90 to 100 percent of utilization days as RU would see an estimated decrease in payments of 9.8 percent. Other facility types that may see higher relative payments under the PDPM are small facilities, non-profit facilities, government-owned facilities, and hospital-based and swing-bed facilities.

Table 38—PDPM Impact Analysis, Facility-Level

Provider characteristics	Percent of providers	Percent change
All Stays	100.0	0.0
Ownership:
For profit	72.0	−0.7
Non-profit	22.6	1.9
Government	5.4	4.2
Number of Certified SNF Beds:
0-49	10.0	3.5
50-99	38.2	0.6
100-149	34.7	−0.2
150-199	11.1	−0.3
200+	5.9	−1.8
Location:
Urban	72.7	−0.7
Rural	27.3	3.8
Facility Type:
Freestanding	96.2	−0.3
Hospital-Based/Swing Bed	3.8	16.7
Location by Facility Type:
Urban | Freestanding:	70.6	−1.0
Urban | Hospital-Based/Swing Bed	2.2	15.3
Rural | Freestanding	25.6	3.2
Rural | Hospital-Based/Swing Bed	1.6	21.1
Census Division:
New England	5.9	2.0
Middle Atlantic	10.8	−2.6
East North Central	20.6	0.7
West North Central	12.5	6.7
South Atlantic	15.7	−0.4
East South Central	6.6	1.0
West South Central	13.1	−1.0
Mountain	4.7	1.1
Pacific	10.1	−0.8
Location by Region:
Urban | New England	5.1	1.8
Urban | Middle Atlantic	9.5	−2.9
Urban | East North Central	14.4	−0.1
Urban | West North Central	6.0	4.6
Urban | South Atlantic	12.6	−1.1
Urban | East South Central	3.6	0.3
Urban | West South Central	8.7	−1.2
Urban | Mountain	3.4	0.1
Urban | Pacific	9.5	−0.9
Rural | New England	0.8	4.0
Rural | Middle Atlantic	1.3	2.7
Rural | East North Central	6.2	3.6
Rural | West North Central	6.5	10.5
Rural | South Atlantic	3.1	4.2
Rural | East South Central	3.0	2.1
Rural | West South Central	4.4	−0.1
Rural | Mountain	1.3	6.2
Rural | Pacific	0.6	2.2
% Stays with Maximum Utilization Days = 100:
0-10	94.4	0.1
10-25	5.1	−2.8
25-100	0.4	−3.6
% Medicare/Medicaid Dual Enrollment:
0-10	8.6	−1.3
10-25	17.5	−1.3
25-50	36.0	0.3
50-75	26.5	1.3
75-90	8.2	0.4
90-100	3.1	1.6
% Utilization Days Billed as RU:
0-10	8.9	27.6
10-25	8.0	15.5
25-50	24.1	7.0
50-75	39.2	−0.4
75-90	17.2	−6.0
90-100	2.6	−9.8
% Utilization Days Billed as Non-Rehab:
0-10	79.8	−1.5
10-25	16.6	8.6
25-50	2.7	23.1
50-75	0.4	35.8
75-90	0.2	41.8
90-100	0.4	33.6

We proposed to implement the PDPM effective beginning in FY 2020 (that is, October 1, 2019). This effective date would incorporate a 1-year period to allow time for provider education and training, internal system transitions, and to allow states to make any Medicaid program changes which may be necessary based on the changes related to PDPM.

With regard to the changes finalized in this rule, we provide our reasons for each change throughout the subsections above. Below in this section, we discuss alternatives we considered which relate generally to implementation of the PDPM.

When making major system changes, CMS often considers possible transition options for providers and other stakeholders between the former system and the new system. For example, when we updated OMB delineations used to establish a provider's wage index under the SNF PPS in FY 2015, we utilized a blended rate in the first year of implementation, whereby 50 percent of the provider's payment was derived from their former OMB delineation and 50 percent from their new OMB delineation (79 FR 45644-45646).

However, due to the fundamental nature of the change from the current RUG-IV case-mix model to the PDPM, which includes differences in resident assessment, payment algorithms, and other policies, as we stated in the proposed rule (83 FR 21079), we believe that proposing a blended rate for the whole system (that would require two full case-mix systems—RUG-IV and the PDPM—to run concurrently) is not advisable as part of any transition strategy for implementing the PDPM, due to the significant administrative and logistical issues that would be associated with such a transition strategy. Specifically, CMS and providers would be required to manage both the RUG-IV payment model and PDPM simultaneously, creating significant burden and undue complexity for all involved parties. Furthermore, providers would be required to follow both sets of MDS assessment rules, each of which carries with it its own level of complexity. CMS would also be required to process assessments and claims under each system, which would entail a significant amount of resources and burden for CMS, MACs, and providers. Finally, a blended rate option would also mitigate some of the burden reduction associated with implementing PDPM, estimated to save SNFs close to $200 million per year as compared to estimated burden under RUG-IV, given that the current assessment schedule would need to continue until full implementation of PDPM was achieved. As we stated in the proposed rule, we believe these issues also would be implicated in any alternative transition strategy which would require both case-mix systems to exist concurrently, such as giving providers a choice in the first year of implementation of operating under either the RUG-IV or PDPM. Therefore, we did not pursue any alternatives which required concurrent operation of both the RUG-IV and PDPM.

As discussed in the proposed rule (83 FR 21079), we then considered alternative effective dates for implementing the PDPM, and other associated policy changes. We considered implementing the new case-mix model effective beginning in FY 2019, but we believe that this would not permit sufficient time for providers and other stakeholders, including CMS, to make the necessary preparations for a change of this magnitude in the SNF PPS. We also believe that such a quick transition would not be in keeping with how similar types of SNF PPS changes have been implemented in the past. We also considered implementing PDPM more than one year after being finalized, such as implementing the PDPM effective beginning October 1, 2020 (FY 2021). However, we believe that setting the effective date of PDPM this far out is not necessary, based on our prior experience with similar SNF PPS changes. As is customary, we plan to continue to provide free software to providers which can be used to group residents under the PDPM, as well as providing data specifications for this grouper software as soon as is practicable, thereby mitigating potential concerns around software vendors having sufficient time to develop products for PDPM. Moreover, given the issues identified throughout the proposed rule and this final rule with the current RUG-IV model, notably the issues surrounding the burden and complexity of the current SNF PPS assessment schedule and concerns around the incentives for therapy overprovision under the RUG-IV system, we believe it appropriate to implement the PDPM as soon as is practicable.

Finally, we considered alternatives related to the proposal discussed in section V.I. of this final rule, specifically the proposed certification that we have met the requirements set forth in section 511(a) of the MMA, which would permit us to use the PDPM's permanent case-mix adjustments for SNF residents with AIDS to replace the temporary special add-on in the PPS per diem payment for such residents. As noted in section V.I. of this final rule, this special add-on for SNF residents with AIDS was intended to be of limited duration, as the MMA legislation specified that it was to remain in effect only until the Secretary certifies that there is an appropriate adjustment in the case mix to compensate for the increased costs associated with such residents. We considered maintaining this adjustment under the PDPM. However, given the adjustment incorporated into the NTA and nursing components under the PDPM to account for the increased costs of treating residents with AIDS, this would result in a substantial increase in payment for such residents beyond even the current add-on payment. Moreover, as discussed in section V.I. of this final rule, we believe that the PDPM provides a tailored case-mix adjustment that more accurately accounts for the additional costs and resource use of residents with AIDS, as compared to an undifferentiated add-on which simply applies an across-the-board multiplier to the full SNF PPS per diem. Finally, as stated in section 3.8.2. of the SNF PDPM technical report (available at https://www.cms.gov/​Medicare/​Medicare-Feefor-Service-Payment/​SNFPPS/​therapyresearch.html), HIV/AIDS was associated with a negative and statistically significant decrease in PT, OT and SLP costs per day. This means inherently that, to the extent that the existing add-on is applied against the full SNF PPS per diem payment, the magnitude of the add-on payment increases with increases in therapy payment, which conflicts with the data described above regarding the relationship between therapy costs and the presence of an AIDS diagnosis. As a result, maintaining the current add-on would create an inconsistency between how SNF payments would be made and the data regarding AIDS diagnoses and resident therapy costs. Therefore, we proposed (and are finalizing in this rule) replacing this add-on payment with appropriate case-mix adjustments for the increased costs of care for this population of residents through the NTA and nursing components of the PDPM.

We invited comments on the projected impacts and on the proposals and alternatives discussed throughout this section.

Commenters submitted the following comments related to the proposed rule's discussion of the Potential Impacts of Implementing the Proposed PDPM and Proposed Parity Adjustment. A discussion of these comments, along with our responses, appears below.

Comment: Commenters agreed that PDPM should be implemented in a budget neutral manner. With regard to the impact analysis, several commenters suggest that CMS run the entire PDPM model on a second year of data (or partial 2018 data) to examine the impact on individual providers and beneficiaries. Commenters state that using only one year of data does not allow analysis of the impact of changing patient populations over time.

Response: We appreciate the support for our proposed budget neutral implementation. With regards to the comment that CMS should use more than one year of data for the impact analysis, we would note that while CMS did not specifically examine the impact of PDPM on individual providers and beneficiaries across multiple years, we did take several steps to ensure robustness of our results. First, to ensure that the classification would be relevant for the current SNF population, we used the latest complete year of data available, FY 2017, to construct the payment model. Second, based on comments received in response to the 2017 ANPRM, we used four years of data (FYs 2014-2017) to determine which comorbidities to include in the NTA component and the number of points to assign to each condition/service for purposes of resident classification and payment. Third, as discussed in section 1.3 of the SNF PDPM technical report, we conducted a series of investigations to test the robustness of our results across multiple years. We found that: The distribution of stays and resource utilization by each classifier used in the payment model (for example, clinical category, cognitive status, etc.), the case-mix groups generated by the CART algorithm, the costliest NTA comorbidities, and the distribution of stays across nursing RUGs was very similar across multiple years. Fourth, we examined changes in SNF resident characteristics over time in response to concerns raised by participants in technical expert panels, focusing on specific resident characteristics that TEP panelists identified as indicators of increasing acuity. These investigations generally found that resident characteristics changed little over time.

Finally, while we did not analyze the impact of PDPM on individual providers and beneficiaries across multiple years, we note that we also examined the impact of the RCS-I payment model, which has substantially similar classification criteria as PDPM, on various resident and provider subpopulations using FY 2014 data. The results of this analysis, shown in section 3.13 of the SNF PMR technical report and the 2017 ANPRM (82 FR 21008 through 21012), were consistent with the resident and provider subpopulation impact analysis conducted for PDPM (section 3.12 of the SNF PDPM technical report) in showing that a payment model based on the set of resident characteristics used to construct PDPM would be expected to increase payment associated with resident subpopulations with complex clinical needs, such as extensive services, high NTA utilization, IV medications, ESRD, diabetes, wound infections, amputation/prosthesis care, and longer inpatient stays. For all of the foregoing reasons, we expect PDPM to be robust and to have similar impacts on residents and providers across multiple years.

Comment: One commenter stated that there are several methodological issues that may affect the accuracy of PDPM impact calculations under budget neutrality. This includes:

(1) The use of hospital MS-DRGs in developing clinical categories will likely result in an inaccurate estimation of payment. Payment rates were set and impacts predicted based on using the MS-DRG assignment of the patient, whereas PDPM when implemented will rely on MDS responses. If SNFs report patients at a net lower acuity level in MDS data than the predicted clinical categorization, then the budget neutrality assumptions made by PDPM will be invalid.

(2) The conversion of charges to costs will likely result in an underestimation of payment. Because SNF charges have not driven payments under the SNF PPS before, it is possible SNFs will systematically re-evaluate their charges practices to bring them more in line with the cost information within their accounting systems. As a result, the use of SNF charges may need to be rapidly reevaluated once PDPM is implemented.

(3) The quality of FY 2017 section GG data is questionable due to the likely inaccuracies in newly implemented items. Thus, PDPM impacts may need to be re-run once more stable section GG data are available to ensure PDPM accurately accounts for patient functional characteristics.

Response: As stated in the proposed rule (83 FR 21074) and section 3.11.2 of the SNF PDPM technical report, the budget neutrality assumption refers to having total payments if PDPM had been in place be equal to total actual RUG-IV payments in FY 2017. It does not account for provider behavior change after the implementation of PDPM. We appreciate the concerns raised, and we will monitor the reporting of MDS clinical categories, charges, and section GG items under PDPM.

Comment: One commenter stated that PDPM does not adequately account for residents with behavioral health issues. The commenter stated that SNFs are treating younger patients with longer stays and complex behavioral needs. Further, the commenter said representatives of geriatric behavioral health services were not included in the TEPs that were convened during PDPM development. A few commenters requested that CMS study the impact of PDPM on beneficiaries with long stays, such as those exceeding 84 days in length to determine whether the payment model creates potential access issues for such beneficiaries.

Response: While our TEPs did not include a specific representative of geriatric behavioral health services, in response to the feedback received from TEP panelists, we investigated the impact of PDPM on residents with behavioral health issues. As discussed in section 3.12 of the SNF PDPM technical report, we found that PDPM is predicted to slightly increase payment associated with residents who have behavioral issues. Therefore, we believe the proposed payment model appropriately accounts for the resource needs of this subpopulation. Additionally, we found that PDPM is expected to notably increase payment associated with younger residents (below 65 and 65-74 years of age). However, we also estimated that payment associated with very long stays (utilization = 100 days) would decline by 1.9 percent under PDPM. We do plan to monitor the impact of PDPM on many different subpopulations, including those with long SNF stays.

Comment: Another commenter raised concerns about the provider-specific impact analysis included in the supplementary materials that were designed to aid stakeholders in reviewing and commenting on the proposed rule. The commenter stated that there were large differences in the estimated payment impact on individual providers between the provider-level impact file that accompanied the 2017 ANPRM and the provider-level impact file that accompanied the FY 2019 proposed rule. Additionally, the commenter stated that some providers have impact estimates in the RCS-I provider-level impact file (which accompanied the 2017 ANPRM) but are missing estimates in the PDPM provider-level impact file (which accompanied the FY 2019 SNF PPS proposed rule). According to the commenter, these discrepancies raise concerns about the reliability, accuracy, and completeness of the data used to develop PDPM.

Response: The commenter that raised concerns about changes in the provider-level impacts between the RCS-I and the PDPM provider-level impact files correctly notes that the provider-level impacts changed across the two files. There are two main reasons for changes in provider-level impacts across these two files that do not raise concerns about the quality of the data used to conduct the provider-specific impact analysis or to develop PDPM. First, the year of analysis is different across the two files. The RCS-I analysis uses data from FY 2014, which was also the year of data used to develop RCS-I, while the PDPM analysis uses data from FY 2017. Changes in the resident population of specific providers could contribute to changes in the estimated provider-level impact of the payment models.

Second, the two provider-level files provide impacts for two different payment models: The first displays impacts for RCS-I, while the second displays impacts for PDPM. While the two payment models are similar, differences between the two models also contribute to changes in estimated provider-level impacts. For the foregoing reasons, we should not expect the estimated payment impact for each provider be the same across the two payment models and data years. We further note that at the population level, the estimated impact on specific types of providers and residents was similar under RCS-I and PDPM, reflecting the similarity of the payment models. Specifically, for both models we estimate that payment would shift from stays receiving high amounts of therapy and providers that provide high amounts of therapy to stays associated with medically complex beneficiaries and providers that serve these beneficiaries.

Regarding providers that were included in the RCS-I provider-specific file but not in the PDPM provider-specific file, this occurs for three reasons: (1) The provider had no stays in FY 2017, the year of analysis for the PDPM file, (2) after applying matching and validity restrictions, the provider had no stays remaining in the dataset, or had fewer than 11 stays (and therefore could not be included for confidentiality reasons), or (3) after excluding stays that did not have sufficient information to be classified into a case-mix group for each PDPM component, the provider had fewer than 11 stays. Of the roughly 1,100 providers that were included in the RCS-I file but not included in the PDPM file, about 60 percent were excluded for reason (3); of the remaining excluded providers, about half were excluded for reason (1) and half were excluded for reason (2). It should also be noted that in total, there are about 700 fewer providers in the PDPM file than there are in the RCS-I file. Because this number is less than the number of providers included in the RCS-I file but not included in the PDPM file, this indicates that there are also a number of providers that are included in the PDPM file but not in the RCS-I file. To confirm the representativeness of our PDPM study population, we compared resident characteristics for the study population and the Medicare Part A SNF population, as shown in section 3.1.5 of the SNF PDPM technical report. As noted in the technical report, the two populations are similar in most respects, although the study population contains a higher proportion of stays from for-profit and freestanding facilities and a lower proportion of stays from non-profit, government, hospital-based, and swing bed facilities. Given the similarity of the two populations, we do not believe our population restrictions compromised the representativeness of our study population or the reliability of our results.

Comment: One commenter stated that there are apparent errors in the PDPM provider-specific impact file. The commenter states that the total numbers of days and stays shown in the file do not match the sum of the values in the respective columns. Additionally, the commenter states that the percentages of stays shown in the case-mix group distribution does not sum to 500 percent (as they should because 100 percent of days are assigned to a case-mix group within each of the five components) for three specific facilities. The commenter notes that all other rows in this tab correctly sum to 500 percent. The commenter recommends CMS research these issues and publish a corrected file as necessary.

Response: The commenter that stated the total stays and days shown in provider-specific file do not match the sum of the values in the respective columns is correct. The reason for this apparent discrepancy is that, while the total stays and days shown in this file include providers with fewer than 11 stays, these providers are not shown separately in the file for confidentiality reasons. As a result, the displayed totals across all facilities do not match the totals calculated from summing across rows. Regarding the three instances the commenter cites in which the percentages for the case-mix group distribution do not sum to 500 percent, we were unable to replicate this issue. We verified that the case-mix group distribution shown in the provider-specific file for each of these three providers does in fact sum to 500 percent and further verified that the case-mix group distribution sums to 500 percent for all providers shown in the file. Therefore, we do not believe a correction is warranted.

Comment: Some commenters supported CMS' decision not to propose a blended rate transition between RUG-IV and PDPM, but rather to make a full transition from one system to the other. Some commenters expressed support for a transition, requesting that CMS conduct a feasibility study to examine the impact of PDPM, particularly the therapy components, on access to medically necessary therapy. One commenter requested that CMS phase-in any negative impacts on providers from implementing PDPM. One commenter stated that, given the similarities between the RCS-I model and the PDPM, CMS should move forward with implementing PDPM in FY 2019. One commenter requested clarification on how a patient's reimbursement would be affected if the stay began under RUG-IV and ended under PDPM.

Response: We appreciate the support for our decision not to implement a transition strategy such as a blended rate option. We do not believe that such a transition, or one that would phase in negative impacts, would be beneficial for SNFs or their patients given the complexity of operating two systems simultaneously. With regard to the suggestion that CMS conduct a feasibility study to examine the impact of PDPM, we believe that the monitoring program we plan to undertake with implementation of PDPM will provide all of the necessary information in an efficient and expeditious manner that would negate the reasons for conducting a feasibility study. Finally, with regard to the comment that CMS implement PDPM in FY 2019, despite the similarities between RCS-I and PDPM, the education and training efforts necessary to ensure successful implementation of PDPM will likely require more time than such an implementation date would permit.

With regard to the comment about a patient that begins a stay under RUG-IV but ends under PDPM, given that there will be no transition period between RUG-IV and PDPM, providers would bill under RUG-IV for all days up to and including September 30, 2019 and then bill under PDPM for all days beginning October 1, 2019. Further, RUG-IV assessment scheduling and other RUG-IV payment-related policies would be in effect until September 30, 2019. Beginning on October 1, 2019, all PDPM related assessment scheduling and other PDPM payment-related policies would take effect.

Comment: One commenter stated that PDPM would require a minimum of 12 months for programming, testing, validating and deploying of software updates and tools. This commenter requested that CMS allow for our systems to report to providers RUG-IV payment data, such as associated HIPPS codes, up to 60 days after implementation of PDPM.

Response: We agree with the commenter regarding the timeframe for software development, which is part of the reason we are implementing PDPM on October 1, 2019, rather than in 2018. With regard to the comment that we report RUG-IV payment data after implementation of PDPM, we will consider this suggestion as part of transition planning.

Comment: Many commenters stressed the importance of provider education and training to support successful implementation of the PDPM. These commenters suggested that extensive education and training of all involved parties will be needed because PDPM is such a significant change from the existing system. These commenters recommend that CMS immediately begin work with stakeholders to identify and to plan for meeting these needs and to provide the necessary tools to implement the new system smoothly. Further, commenters suggested that, in Fall 2018, CMS should convene a PDPM Implementation Technical Expert Panel (TEP) comprised of SNF PPS stakeholders, representatives from states, referral sources, and payer representatives, and that the TEP Report should be made public and serve as the basis for a PDPM Transition Plan. Finally, several commenters urged CMS to release any technical specifications and manual revisions as soon as possible, to give providers and vendors as much time as possible to adapt to any PDPM-related changes.

Response: We agree with the comments regarding the importance of provider education and training and will be providing extensive opportunities and resources to accomplish this task. With regard to the suggestion for a TEP related to PDPM implementation, we appreciate this suggestion and will consider several methods to engage the stakeholder community in preparing for PDPM implementation. With regard to the comments on the need for transition planning and for CMS' timely release of any technical specifications and manual revisions, we agree with commenters and intend to release technical specifications and manual revisions as soon as possible, which will include specific instructions on operationalizing the transition from RUG-IV to PDPM.

Comment: A few commenters requested that CMS establish a formal and transparent process and timeline for refining the PDPM therapy components after implementation of PDPM.

Response: While we agree with using a transparent process for refining PDPM, as was used during its development, we believe it is premature at this time to provide such a timeframe for revisions to the model, until we are able to observe the impact of implementing this model.

Comment: One commenter requested that CMS consider providing additional funding during initial implementation of PDPM, given that providers will be under financial pressures associated with training, software purchases, as well as changes associated with other CMS initiatives.

Response: We do not believe that additional funding would be warranted for the activities described by the commenter. Given that CMS provides free grouper software, as well as a myriad of training and education resources, we believe that additional costs, such as software purchases, are private business decisions that exist outside the scope of SNF payments.

Accordingly, after considering the comments received, for the reasons discussed throughout section V of the FY 2019 SNF PPS proposed rule and for the reasons presented in this final rule, we are finalizing our proposals to implement the PDPM, as well as the other PDPM related changes discussed in this final rule, with the modifications previously discussed in this final rule, effective beginning October 1, 2019. Specifically, in section V.B of this final rule, we finalized our proposal, without modification, for updating the federal base payment rates and for adjusting the per diem rates for geographic differences under the PDPM. In section V.C.3.b of this final rule, we finalized the proposed PT and OT components under the PDPM and our proposals relating to the methodology for classifying residents under the PT and OT components, effective October 1, 2019, with the modifications discussed in that section. More specifically, in response to comments, rather than requiring providers to record the type of inpatient surgical procedure performed during the prior inpatient hospital stay by coding an ICD-10-PCS code in the second line of item I8000 as we proposed, we will instead require providers to select, as necessary, a surgical procedure category in a sub-item within Item J2000 which would identify the relevant surgical procedure that occurred during the patient's preceding hospital stay and which would augment the patient's PDPM clinical category. For purposes of calculating the function score, all missing values for section GG assessment items will receive zero points. Similarly, the function score will incorporate a new response “10. Not attempted due to environmental limitations” and we will assign it a point value of zero. Furthermore, consistent with a commenter's suggestion, we will adopt MDS item GG0170I1 (Walk 10 feet) as a substitute for retired item GG0170H1 (Does the resident walk), and we will use responses 07: “resident refused,” 09: “not applicable,” 10: “not attempted due to environmental limitations,” or 88: “not attempted due to medical condition or safety concerns” from MDS item GG0170I1 to identify residents who cannot walk. In section V.C.3.b of this final rule, we finalized, without modification, the proposed SLP component of PDPM and our proposals relating to the classification of residents under the SLP component. In section V.C.3.d of this final rule, we finalized, without modification, our proposals relating to the methodology for classifying patients under the nursing component of PDPM. In section V.C.3.e of this final rule, we finalized, without modification, our proposed NTA component of the PDPM and the proposed classification methodology for the NTA component. In section V.C.4 of this final rule, we finalized, without modification, to apply a variable per diem adjustment as part of the PDPM, utilizing the adjustment factors and schedule for the PT and OT components found in Table 30 and the adjustment factors and schedule for the NTA component found in Table 31. In section V.D.1 of this final rule, we finalized our proposed changes to the MDS assessment schedule and related assessment policies as discussed in the proposed rule, with the following modifications. As discussed in that section, rather than making the IPA a required assessment as we proposed, this assessment will be optional, and providers may determine whether and when an IPA is completed. In addition, because the IPA is an optional assessment and providers can determine their own criteria for when an IPA is completed, we are revising the ARD criteria such that the ARD will be the date the facility chooses to complete the assessment relative to the triggering event that makes the facility complete the IPA. Payment based on the IPA would begin the same day as the ARD. In section V.D.2 of this final rule, we finalized, without modification, our proposed additions to the Swing Bed PPS Assessment found in Table 34 of this final rule. In section V.D.3 of this final rule, we finalized, without modification, the proposed additions to the PPS Discharge Assessment found in Table 35 of this final rule. In section V.E of this final rule, we finalized, without modification, our proposed application of a combined 25 percent limit on group and concurrent therapy, per therapy discipline, as well as our proposal to implement a non-fatal warning edit on a provider's validation report when the amount of group and concurrent therapy exceeds 25 percent within a given therapy discipline. In section V.F of this final rule, we finalized, without modification, our proposed interrupted stay policy. In section V.G of this final rule, we finalized our proposed classifiers for purposes of applying the administrative presumption, with the following modification. As discussed in that section, we added 6 PT and OT classifiers and 8 SLP classifiers. In section V.H of this final rule, we finalized our proposal to replace the existing MMA add-on for patients with AIDS with the PDPM permanent adjustment in the case-mix that appropriate accounts for the increased costs of patients with AIDS.

As we proposed and as discussed in section V.I of this final rule, we will implement the PDPM and the other PDPM-related changes finalized in this rule in a budget neutral manner.

VI. Other Issues

A. Other Revisions to the Regulation Text

Along with our revisions to the regulations as discussed elsewhere in this final rule, we also proposed (83 FR 21080) to make two other revisions in the regulation text. The first involves § 411.15(p)(3)(iv), which specifies that whenever a beneficiary is formally discharged (or otherwise departs) from the SNF, this event serves to end that beneficiary's status as a “resident” of the SNF for purposes of consolidated billing (the SNF “bundling” requirement), unless he or she is readmitted (or returns) to that or another SNF “by midnight of the day of departure.” In initially establishing this so-called “midnight rule,” the FY 2001 SNF PPS final rule (65 FR 46770, July 31, 2000) noted in this particular context that, as we explained in the proposed rule, a patient “day” begins at 12:01 a.m. and ends the following midnight, so that the phrase “midnight of the day of departure” refers to the midnight that immediately follows the actual moment of departure, rather than to the midnight that immediately precedes it (65 FR 46792).

However, the Medicare program's standard practice for counting inpatient days is actually one in which an inpatient day would begin at midnight (see, for example, § 20.1 in the Medicare Benefit Policy Manual, Chapter 3, which specifies that in counting inpatient days, “. . . a day begins at midnight and ends 24 hours later” (emphasis added)). Accordingly, in order to ensure consistency with that approach, we proposed to revise § 411.15(p)(3)(iv) to specify that for consolidated billing purposes, a beneficiary's “resident” status ends whenever he or she is formally discharged (or otherwise departs) from the SNF, unless he or she is readmitted (or returns) to that or another SNF “before the following midnight.” We further noted that this revision would not alter the underlying principle that a beneficiary's SNF “resident” status in this context ends upon departure from the SNF unless he or she returns to that or another SNF later on that same day; rather, it would simply serve to conform the actual wording of the applicable regulations text with the Medicare manual's standard definition of the starting point of a patient “day.”

We also proposed a technical correction to § 424.20(a)(1)(i) (which describes the required content of the SNF level of care certification) in order to conform it more closely to that of the corresponding statutory requirements at section 1814(a)(2)(B) of the Act. This statutory provision defines the SNF level of care in terms of skilled services furnished on a daily basis which, as a practical matter, can only be provided on an inpatient basis in a SNF. In addition, it provides that the SNF-level care must be for either:

• An ongoing condition that was one of the conditions that the beneficiary had during the qualifying hospital stay; or
• A new condition that arose while the beneficiary was in the SNF for treatment of that ongoing condition.

In setting forth the SNF level of care definition itself, the implementing regulations at § 409.31 reflect both of the above two criteria (at paragraphs (b)(2)(i) and (b)(2)(ii), respectively); however, as we stated in the proposed rule (83 FR 21080), the regulations describing the content of the initial level of care certification at § 424.20(a)(1)(i) have inadvertently omitted the second criterion. Further, while that criterion admittedly might not be relevant in those instances where the initial certification is obtained promptly “at the time of admission” in accordance with the regulations at 42 CFR 424.20(b)(1), that same provision alternatively allows this requirement to be met “as soon thereafter as is reasonable and practicable.” Accordingly, in order to rectify this omission, we proposed to revise § 424.20(a)(1)(i) so that it more accurately tracks the language in the corresponding statutory authority at section 1814(a)(2)(B) of the Act.

We invited comments on our proposed revisions to § 411.15(p)(3)(iv) and § 424.20(a)(1)(i), but received no comments on either revision. Accordingly, in this final rule, we are finalizing both revisions as proposed, without further modification.

B. Skilled Nursing Facility (SNF) Quality Reporting Program (QRP)

1. Background

The Skilled Nursing Facility Quality Reporting Program (SNF QRP) is authorized by section 1888(e)(6) of the Act and it applies to freestanding SNFs, SNFs affiliated with acute care facilities, and all non-CAH swing-bed rural hospitals. Under the SNF QRP, the Secretary reduces by 2 percentage points the annual market basket percentage update described in section 1888(e)(5)(B)(i) of the Act applicable to a SNF for a fiscal year, after application of section 1888(e)(5)(B)(ii) of the Act (the MFP adjustment) and section 1888(e)(5)(B)(iii) of the Act (the 1 percent market basket increase for FY 2018), in the case of a SNF that does not submit data in accordance with sections 1888(e)(6)(B)(i) of the Act for that fiscal year. For more information on the requirements we have adopted for the SNF QRP, we refer readers to the FY 2016 SNF PPS final rule (80 FR 46427 through 46429), FY 2017 SNF PPS final rule (81 FR 52009 through 52010) and FY 2018 SNF PPS final rule (82 FR 36566).

Although we have historically used the preamble to the SNF PPS proposed and final rules each year to remind stakeholders of all previously finalized program requirements, we have concluded that repeating the same discussion each year is not necessary for every requirement, especially if we have codified it in our regulations. Accordingly, the following discussion is limited as much as possible to a discussion of our proposals, responses to comments submitted on those proposals, and policies we are finalizing for future years of the SNF QRP after consideration of the comments, and it represents the approach we intend to use in our rulemakings for this program going forward.

2. General Considerations Used for the Selection of Measures for the SNF QRP

a. Background

For a detailed discussion of the considerations we historically used for the selection of SNF QRP quality, resource use, and other measures, we refer readers to the FY 2016 SNF PPS final rule (80 FR 46429 through 46431).

We received several comments generally related to the SNF QRP. The comments and our responses are discussed below.

Comment: Several commenters expressed general support for CMS's proposals related to the SNF QRP, acknowledging CMS's goal of improving the quality of health care for Medicare beneficiaries through improvements to patient assessments and quality reporting. One commenter highlighted the need for additional transparency from CMS through this ongoing process. Another commenter requested that CMS ensure that the SNF QRP efforts do not negatively impact specialty populations.

Response: We appreciate commenters' general support for the SNF QRP proposals. To foster transparency, we continue to seek stakeholder input and will take into consideration the impact of specialty populations in the ongoing measure development and maintenance efforts of the SNF QRP.

Comment: One commenter expressed support for the IMPACT Act's objectives. However, the commenter expressed concern over the rapid development and implementation of the standardized patient assessment data element (SPADE) work, suggesting that further evaluation is necessary.

Response: We understand the concerns raised by commenters pertaining to the development and implementation of the SPADEs. As discussed in the FY 2018 SNF PPS Final Rule, we agreed that further evaluation of the data elements was necessary. Specifically, we thought that more time was needed to develop, test, to think through the implementation, and to reflect on how to maximize the time SNFs have to prepare for the reporting of standardized resident assessment data in these categories. We have worked to be responsive to the concerns raised by stakeholders while meeting our obligation to require the reporting of standardized resident assessment data with respect to the categories described in section 1899B(b)(1)(B) of the Act. Therefore, as outlined in the FY 2018 SNF PPS final rule, we did not finalize the standardized assessment data elements we proposed for three of the five categories under section 1899B(b)(1)(B) of the Act: Cognitive Function and Mental Status; Special Services, Treatments, and Interventions; and Impairments in that we felt this work needed more time for development and evaluation. Since the time of this proposal work, we have worked closely with stakeholders, solicited comments, reconvened our TEP, and are currently re-testing the SPADEs in a national field test (also known as the Alpha test). For more information on our prior proposal addressed in the FY 2018 SNF PPS final rule (82 FR 36568 through 36570, 36597 through 36605), we refer the reader to that detailed discussion. For more information on our national field test and associated work for SPADEs, please see: https://www.cms.gov/​Medicare/​Quality-Initiatives-Patient-Assessment-Instruments/​Post-Acute-Care-Quality-Initiatives/​IMPACT-Act-of-2014/​-IMPACT-Act-Standardized-Assessment-National-Testing-.html.

b. Accounting for Social Risk Factors in the SNF QRP

In the FY 2018 SNF PPS final rule (82 FR 36567 through 36568), we discussed the importance of improving beneficiary outcomes including reducing health disparities. We also discussed our commitment to ensuring that medically complex residents, as well as those with social risk factors, receive excellent care. We discussed how studies show that social risk factors, such as being near or below the poverty level as determined by HHS, belonging to a racial or ethnic minority group, or living with a disability, can be associated with poor health outcomes and how some of this disparity is related to the quality of health care.^[3] Among our core objectives, we aim to improve health outcomes, attain health equity for all beneficiaries, and ensure that complex residents, as well as those with social risk factors receive excellent care. Within this context, reports by the Office of the Assistant Secretary for Planning and Evaluation (ASPE) and the National Academy of Medicine have examined the influence of social risk factors in our value-based purchasing programs.^[4] As we noted in the FY 2018 SNF PPS final rule (82 FR 36567 through 36568), ASPE's report to Congress, which was required by the IMPACT Act, found that, in the context of value-based purchasing programs, dual eligibility was the most powerful predictor of poor health care outcomes among those social risk factors that they examined and tested. ASPE is continuing to examine this issue in its second report required by the IMPACT Act, which is due to Congress in the fall of 2019. In addition, as we noted in the FY 2018 SNF PPS final rule (82 FR 36567), the National Quality Forum (NQF) undertook a 2-year trial period in which certain new measures and measures undergoing maintenance review have been assessed to determine if risk adjustment for social risk factors is appropriate for these measures.^[5] The trial period ended in April 2017 and a final report is available at http://www.qualityforum.org/​SES_​Trial_​Period.aspx. The trial concluded that “measures with a conceptual basis for adjustment generally did not demonstrate an empirical relationship” between social risk factors and the outcomes measured. This discrepancy may be explained in part by the methods used for adjustment and the limited availability of robust data on social risk factors. NQF has extended the socioeconomic status (SES) trial,^[6] allowing further examination of social risk factors in outcome measures.

In the FY 2018/CY 2018 proposed rules for our quality reporting and value-based purchasing programs, we solicited feedback on which social risk factors provide the most valuable information to stakeholders and the methodology for illuminating differences in outcomes rates among patient groups within a provider that would also allow for a comparison of those differences, or disparities, across providers. Feedback we received across our quality reporting programs included encouraging us to explore whether factors that could be used to stratify or risk adjust the measures (beyond dual eligibility); to consider the full range of differences in resident backgrounds that might affect outcomes; to explore risk adjustment approaches; and to offer careful consideration of what type of information display would be most useful to the public.

We also sought public comment on confidential reporting and future public reporting of some of our measures stratified by resident dual eligibility. In general, commenters noted that stratified measures could serve as tools for SNFs to identify gaps in outcomes for different groups of residents, improve the quality of health care for all residents, and empower consumers to make informed decisions about health care. Commenters encouraged us to stratify measures by other social risk factors such as age, income, and educational attainment. With regard to value-based purchasing programs, commenters also cautioned CMS to balance fair and equitable payment while avoiding payment penalties that mask health disparities or discouraging the provision of care to more medically complex patients. Commenters also noted that value-based payment program measure selection, domain weighting, performance scoring, and payment methodology must account for social risk.

As a next step, we are considering options to improve health disparities among patient-groups within and across hospitals by increasing the transparency of disparities as shown by quality measures. We also are considering how this work applies to other CMS quality programs in the future. We refer readers to the FY 2018 IPPS/LTCH PPS final rule (82 FR 38403 through 38409) for more details, where we discuss the potential stratification of certain Hospital Inpatient Quality Reporting (IQR) Program outcome measures. Furthermore, we continue to consider options to address equity and disparities in our value-based purchasing programs.

We plan to continue working with ASPE, the public, and other key stakeholders on this important issue to identify policy solutions that achieve the goals of attaining health equity for all beneficiaries and minimizing unintended consequences.

Comment: Several commenters supported CMS' continuing evaluation of how social risk factors could impact SNF QRP measure rates and encouraged CMS to consider strategies and solutions in this area. Specific comments noted that lack of adjustment for social risk factors may negatively impact facility measure rates, and CMS should incorporate risk adjustment for sociodemographic and socioeconomic status into appropriate SNF QRP measures. We also received comments about the public display of measure information related to social risk factors, suggesting stratified measures be used and expressing concerns that publicly reported outcome measures could be misleading to consumers.

Response: We thank commenters for their comments and will take these comments into account as we further consider how to appropriately account for social risk factors in the SNF QRP. We also refer the reader to the FY 2018 SNF PPS final rule (82 FR 36567 through 36568) where we discussed in depth many of the issues raised by these commenters.

3. New Measure Removal Factor for Previously Adopted SNF QRP Measures

As a part of our Meaningful Measures Initiative discussed in section I.D. of this final rule, we strive to put patients first, ensuring that they, along with their clinicians, are empowered to make decisions about their own healthcare using data-driven information that is increasingly aligned with a parsimonious set of meaningful quality measures. We began reviewing the SNF QRP's measures in accordance with the Meaningful Measures Initiative, and we are working to identify how to move the SNF QRP forward in the least burdensome manner possible while continuing to incentivize improvement in the quality of care provided to patients.

Specifically, we believe the goals of the SNF QRP and the measures used in the program cover most of the Meaningful Measures Initiative priorities, including making care safer, strengthening person and family engagement, promoting coordination of care, promoting effective prevention and treatment, and making care affordable.

We also evaluated the appropriateness and completeness of the SNF QRP's current measure removal factors. We have previously finalized that we would use notice and comment rulemaking to remove measures from the SNF QRP based on the following factors ^[7]

• Factor 1. Measure performance among SNFs is so high and unvarying that meaningful distinctions in improvements in performance can no longer be made.
• Factor 2. Performance or improvement on a measure does not result in better resident outcomes.
• Factor 3. A measure does not align with current clinical guidelines or practice.
• Factor 4. A more broadly applicable measure (across settings, populations, or conditions) for the particular topic is available.
• Factor 5. A measure that is more proximal in time to desired resident outcomes for the particular topic is available.
• Factor 6. A measure that is more strongly associated with desired resident outcomes for the particular topic is available.
• Factor 7. Collection or public reporting of a measure leads to negative unintended consequences other than resident harm.

We continue to believe that these measure removal factors are appropriate for use in the SNF QRP. However, even if one or more of the measure removal factors applies, we may nonetheless choose to retain the measure for certain specified reasons. Examples of such instances could include when a particular measure addresses a gap in quality that is so significant that removing the measure could in turn result in poor quality, or in the event that a given measure is statutorily required. We note further that, consistent with other quality reporting programs, we apply these factors on a case-by-case basis.

In the FY 2019 SNF PPS proposed rule (83 FR 21082), we proposed to adopt an additional factor to consider when evaluating potential measures for removal from the SNF QRP measure set:

• Factor 8. The costs associated with a measure outweigh the benefit of its continued use in the program.

As we discussed in section I.D. of this final rule, with respect to our new Meaningful Measures Initiative, we are engaging in efforts to ensure that the SNF QRP measure set continues to promote improved health outcomes for beneficiaries while minimizing the overall costs associated with the program. We believe these costs are multifaceted and include not only the burden associated with reporting, but also the costs associated with implementing and maintaining the program. We have identified several different types of costs, including, but not limited to: (1) The provider and clinician information collection burden and burden associated with the submission/reporting of quality measures to CMS; (2) the provider and clinician cost associated with complying with other programmatic requirements; (3) the provider and clinician cost associated with participating in multiple quality programs, and tracking multiple similar or duplicative measures within or across those programs; (4) the cost to CMS associated with the program oversight of the measure including measure maintenance and public display; and (5) the provider and clinician cost associated with compliance with other federal and/or state regulations (if applicable).

For example, it may be needlessly costly and/or of limited benefit to retain or maintain a measure which our analyses show no longer meaningfully supports program objectives (for example, informing beneficiary choice). It may also be costly for health care providers to track the confidential feedback, preview reports, and publicly reported information on a measure where we use the measure in more than one program. CMS may also have to expend unnecessary resources to maintain the specifications for the measure, as well as the tools we need to collect, validate, analyze, and publicly report the measure data. Furthermore, beneficiaries may find it confusing to see public reporting on the same measure in different programs.

When these costs outweigh the evidence supporting the continued use of a measure in the SNF QRP, we believe it may be appropriate to remove the measure from the program. Although we recognize that one of the main goals of the SNF QRP is to improve beneficiary outcomes by incentivizing health care providers to focus on specific care issues and making data public related to those issues, we also recognize that those goals can have limited utility where, for example, the publicly reported data is of limited use because it cannot be easily interpreted by beneficiaries and used to influence their choice of providers. In these cases, removing the measure from the SNF QRP may better accommodate the costs of program administration and compliance without sacrificing improved health outcomes and beneficiary choice.

We proposed that we would remove measures based on this factor on a case-by-case basis. We might, for example, decide to retain a measure that is burdensome for health care providers to report if we conclude that the benefit to beneficiaries justifies the reporting burden. Our goal is to move the program forward in the least burdensome manner possible, while maintaining a parsimonious set of meaningful quality measures and continuing to incentivize improvement in the quality of care provided to patients.

We invited public comment on our proposal to adopt an additional measure removal Factor 8. The costs associated with a measure outweigh the benefit of its continued use in the program.

We also proposed to add a new § 413.360(b)(3) that would codify the removal factors we have previously finalized for the SNF QRP, as well as the new measure removal factor that we proposed to adopt in the proposed rule.

We sought comments on these proposals. A discussion of these comments, along with our responses, appears below.

Comment: Several commenters expressed support for an additional factor to consider when evaluating potential measures for removal from the SNF QRP measure set: Factor 8. The costs associated with a measure outweigh the benefit of its continued use in the program. One commenter acknowledged that removal of a measure from the program may be appropriate when the costs outweigh the evidence supporting its continued use. Another commenter supported the addition of the new measure removal factor because it reduces unnecessary administrative burden.

Response: We appreciate the commenters' support.

Comment: One commenter supported CMS' proposal to codify the proposed measure removal factor in the regulatory text.

Response: We appreciate the commenter's support.

Comment: Some commenters expressed concerns related to the new measure removal factor. One commenter did not support the addition of the factor, suggesting that the costs and benefits considered under this factor are not equivalent, as costs are typically imposed on providers while benefits are rendered to beneficiaries. This commenter expressed the concern that providers may argue for removal of a measure that is costly to collect and report despite its benefits. Another commenter suggested that using administrative cost to CMS as a basis for removal may be problematic if clinicians or patients believe the measure is important. Another commenter added that the proposed measure removal factor is subjective and recommended clearer guidelines and criteria for determining the costs and benefits of a measure before it is removed.

Response: We agree that it is possible that providers may recommend removal of measures they do not support based on the argument that these measures are costly to report. However, input from providers is only one element of our case-by-case analysis of measures that we would take into account when weighing the costs associated with a measure against the benefit of retaining the measure in a program. We will weigh input we receive from all stakeholders with our own analysis of each measure to make our case-by-case determination of whether it would be appropriate to remove a measure based on its costs outweighing the benefit of its continued use in the program. We wish to clarify that it is not our intent to remove measures that continue to benefit residents or providers solely because these measures incur administrative costs to CMS; this is only one example of costs that would be weighed against the benefits when considering each measure on a case-by-case basis.

Regarding concern over the subjectivity of the new measure removal factor and the suggestion for clearer guidelines and criteria for determining the costs and benefits of a measure before it is removed, we intend to be transparent in our assessment of measures under this measure removal factor. As described above, there are various considerations of costs and benefits, direct and indirect, financial and otherwise, that we will evaluate in applying removal Factor 8, and we will take into consideration the perspectives of multiple stakeholders. However, because we intend to evaluate each measure on a case-by-case basis, and each measure has been adopted to fill different needs in the SNF QRP, we do not believe it would be meaningful to identify a specific set of assessment criteria to apply to all measures. We believe costs include costs to stakeholders such as patients, caregivers, providers, CMS, and other entities. In addition, we note that the benefits we will consider center around benefits to residents and caregivers as the primary beneficiaries of our quality reporting program. When we propose through rulemaking to remove a measure under this measure removal factor, we will provide information on the costs and benefits we considered in evaluating the measure.

Comment: One commenter noted that the existing seven removal factors are sufficient for appropriate measure evaluation.

Response: While we acknowledge that there are seven factors currently adopted that may be used for considering measure removal from the SNF QRP, we believe the proposed new measure removal factor adds a new criterion that is not captured in the other seven factors. The proposed new measure removal factor will help advance the goals of the Meaningful Measures Initiative, which aims to improve outcomes for patients, their families, and health care providers while reducing burden and costs for clinicians and providers. We are also making minor grammatical edits to the SNF QRP measure removal factor language to align with the language of other CMS quality programs.

After considering the comments, we are finalizing our proposal to add an additional measure removal factor: Factor 8. The costs associated with a measure outweigh the benefit of its continued use in the program. We are also finalizing our proposal to the updates to the regulatory text and to codify the seven removal factors we have previously finalized for the SNF QRP, as well as the new measure removal factor, Factor 8 at new § 413.360(b)(3). We are also making minor grammatical edits to the SNF QRP measure removal factor language to align with the language of other CMS quality programs.

4. Quality Measures Currently Adopted for the FY 2020 SNF QRP

The SNF QRP currently has 12 measures for the FY 2020 program year, which are outlined in Table 39.

Table 39—Quality Measures Currently Adopted for the FY 2020 SNF QRP

Short name	Measure name and data source
Resident Assessment Instrument Minimum Data Set
Pressure Ulcer	Percent of Residents or Patients With Pressure Ulcers That Are New or Worsened (Short Stay) (NQF #0678).*
Pressure Ulcer/Injury	Changes in Skin Integrity Post-Acute Care: Pressure Ulcer/Injury.
Application of Falls	Application of Percent of Residents Experiencing One or More Falls with Major Injury (Long Stay) (NQF #0674).
Application of Functional Assessment/Care Plan	Application of Percent of Long-Term Care Hospital (LTCH) Patients with an Admission and Discharge Functional Assessment and a Care Plan That Addresses Function (NQF #2631).
Change in Mobility Score	Application of IRF Functional Outcome Measure: Change in Mobility Score for Medical Rehabilitation Patients (NQF #2634).
Discharge Mobility Score	Application of IRF Functional Outcome Measure: Discharge Mobility Score for Medical Rehabilitation Patients (NQF #2636).
Change in Self-Care Score	Application of the IRF Functional Outcome Measure: Change in Self-Care Score for Medical Rehabilitation Patients (NQF #2633).
Discharge Self-Care Score	Application of IRF Functional Outcome Measure: Discharge Self-Care Score for Medical Rehabilitation Patients (NQF #2635).
DRR	Drug Regimen Review Conducted With Follow-Up for Identified Issues—Post Acute Care (PAC) Skilled Nursing Facility (SNF) Quality Reporting Program (QRP).
Claims-Based
MSPB SNF	Medicare Spending Per Beneficiary (MSPB)—Post Acute Care (PAC) Skilled Nursing Facility (SNF) Quality Reporting Program (QRP).
DTC	Discharge to Community—Post Acute Care (PAC) Skilled Nursing Facility (SNF) Quality Reporting Program (QRP).
PPR	Potentially Preventable 30-Day Post-Discharge Readmission Measure for Skilled Nursing Facility (SNF) Quality Reporting Program (QRP).
* The measure will be replaced with the Changes in Skin Integrity Post-Acute Care: Pressure Ulcer/Injury measure, effective October 1, 2018.

Comment: While we did not solicit comment on currently adopted or future measures for the SNF QRP, we received multiple comments suggesting the removal or modification of measures finalized in previous rules as well as recommendations for future measure development.

Response: We thank commenters for their comments. We did not propose any changes to our previously finalized measures or to adopt any new measures for the SNF QRP. We will take these comments into consideration as we engage in future measure development and selection activities for the SNF QRP. The SNF QRP measures described in Table 39 were adopted in the FY 2016 SNF PPS final rule (80 FR 46432 through 46453), FY 2017 SNF PPS final rule (81 FR 52012 through 52039), or FY 2018 SNF PPS final rule (82 FR 36570 through 36594), and we refer the reader to those detailed discussions.

5. IMPACT Act Implementation Update

In the FY 2018 SNF PPS final rule (82 FR 36596 through 36597), we stated that we intended to specify two measures that would satisfy the domain of accurately communicating the existence and provision of the transfer of health information and care preferences under section 1899B(c)(1)(E) of the Act no later than October 1, 2018, and intended to propose to adopt them for the FY 2021 SNF QRP, with data collection beginning on or about October 1, 2019.

As stated in the FY 2019 SNF PPS proposed rule (83 FR 21083), as a result of the input provided during a public comment period between November 10, 2016 and December 11, 2016, input provided by a technical expert panel (TEP), and pilot measure testing conducted in 2017, we are engaging in continued development work on these two measures, including supplementary measure testing and providing the public with an opportunity for comment in 2018. We stated that we would reconvene a TEP for these measures in mid-2018 which occurred in April 2018. We stated that we now intend to specify the measures under section 1899B(c)(1)(E) of the Act no later than October 1, 2019 and intend to propose to adopt the measures for the FY 2022 SNF QRP, with data collection beginning with residents admitted as well as discharged on or after October 1, 2020. For more information on the pilot testing, we refer readers to https://www.cms.gov/​Medicare/​Quality-Initiatives-Patient-Assessment-Instruments/​Post-Acute-Care-Quality-Initiatives/​IMPACT-Act-of-2014/​IMPACT-Act-Downloads-and-Videos.html.

Comment: A few commenters supported the delayed implementation of the measures. One commenter supported the continued evaluation and testing of the measures prior to adoption. The commenter believed that this delay is appropriate as it allows more time for thorough measure development, continued field testing of the measures, and public input on the draft measures. This commenter noted that continued development of the measures will help to ensure they are measuring the domain of interest and will have a meaningful impact on the quality of care.

Response: We appreciate the commenters support.

6. Form, Manner, and Timing of Data Submission Under the SNF QRP

Under our current policy, SNFs report data on SNF QRP assessment-based measures and standardized resident assessment data by reporting the designated data elements for each applicable resident on the Minimum Data Set (MDS) resident assessment instrument and then submitting completed instruments to CMS using the Quality Improvement Evaluation System Assessment Submission and Processing (QIES ASAP) system. We refer readers to the FY 2018 SNF PPS final rule (82 FR 36601 through 36603) for the data collection and submission time frames for assessment-based measures and standardized resident assessment data that we finalized for the SNF QRP.

7. Changes to the SNF QRP Reconsideration Requirements

Section 413.360(d)(1) of our regulations states, in part, that SNFs that do not meet the SNF QRP requirements for a program year will receive a letter of non-compliance through the QIES ASAP system, as well as through the United States Postal Service.

In the FY 2019 SNF PPS proposed rule (83 FR 21083), we proposed to revise § 413.360(d)(1) to expand the methods by which we would notify a SNF of non-compliance with the SNF QRP requirements for a program year. Revised § 413.360(d)(1) would state that we would notify SNFs of non-compliance with the SNF QRP requirements via a letter sent through at least one of the following notification methods: The QIES ASAP system; the United States Postal Service; or via an email from the Medicare Administrative Contractor (MAC). We believe that this change will address feedback from providers who requested additional methods for notification.

In addition, § 413.360(d)(4) currently states that we will make a decision on the request for reconsideration and provide notice of the decision to the SNF through the QIES ASAP system and via letter sent through the United States Postal Service.

We proposed to revise § 413.360(d)(4) to state that we will notify SNFs, in writing, of our final decision regarding any reconsideration request via a letter sent through at least one of the following notification methods: The QIES ASAP system, the United States Postal Service, or via an email from the Medicare Administrative Contractor (MAC).

We invited public comments on these proposals.

Comment: Several commenters expressed support for CMS' efforts to expand the methods for notifying providers of non-compliance and decisions on reconsideration requests. One commenter acknowledged that the addition of email notifications from the Medicare Administrative Contractor (MAC) as a third notification method may help reduce burden, adding that providers should be notified via at least two of the three methods and that letters should require return receipt to ensure notifications are not lost in the mail. Another commenter recommended that CMS either specify a notification method that will always be used, allow providers to select a preferred method, or consistently use all three methods to ensure that notifications are received by appropriate organization leaders. Several commenters suggested that CMS provide additional information regarding how to specify appropriate recipients of email notifications from the Medicare Administrative Contractor (MAC). Another commenter recommended selecting a consistent notification process, using the same methods for all SNFs, noting that consistent and predictable notification will reduce provider burden and lower the risk of missing a notification.

Response: We thank commenters for their support. We will use at least one method of notification, and providers will be notified regarding the specific method of communication that CMS will use via the SNF QRP Reconsideration and Exception and Extension website at https://www.cms.gov/​Medicare/​Quality-Initiatives-Patient-Assessment-Instruments/​NursingHomeQualityInits/​Skilled-Nursing-Facility-Quality-Reporting-Program/​SNF-QR-Reconsideration-and-Exception-and-Extension.html and announcements via the PAC listserv. The announcements will be posted annually following the May 15th data submission deadline—prior to the distribution of the initial notices of non-compliance determination in late spring/early summer. Messaging will include the method of communication for the notices, instructions for sending a reconsideration request, and the final deadline for submitting the request. This policy would be effective October 1, 2018.

With regard to the comment about specifying the recipient of notification for a facility, our notifications are sent to the point of contact on file in the QIES database. This information is populated via ASPEN. It is the responsibility of the facility to ensure that this information is up-to-date. For information regarding how to update provider information in QIES, we refer providers to contact their Medicare Administrative Contractor or CMS Regional Office at https://www.cms.gov/​About-CMS/​Agency-Information/​RegionalOffices/​index.html. Downloads of contact information for each Regional Office are available at the bottom of the web page.

We disagree with the recommendation that CMS notify all SNFs using the same method in order to account for circumstances that are beyond our control, such as technical issues that may impede the delivery of electronic notifications. As discussed, providers will be notified in advance of the specific method of communication that CMS will use.

We are finalizing our proposal to revise § 413.360(d)(1) to state that we will notify a SNF of non-compliance with the SNF QRP requirements for a program year via a letter sent through at least one of the following notification methods: The QIES ASAP system; the United States Postal Service; or via an email from the Medicare Administrative Contractor (MAC).

We are also finalizing our proposal, to revise § 413.360(d)(4) to state that we will notify SNFs, in writing, of our final decision regarding any reconsideration request via a letter sent through at least one of the following notification methods: The QIES ASAP system, the United States Postal Service, or via an email from the Medicare Administrative Contractor (MAC).

8. Policies Regarding Public Display for the SNF QRP

Section 1899B(g) of the Act requires the Secretary to establish procedures for the public reporting of SNFs' performance on measures under sections 1899B(c)(1) and 1899B(d)(1) of the Act. SNF QRP measure data will be displayed on the Nursing Home Compare website, an interactive web tool that assists individuals by providing information on SNF quality of care to those who need to select a SNF.

In the FY 2018 SNF PPS final rule (82 FR 36606 through 36607), we finalized that we would publicly display the Medicare Spending Per Beneficiary-PAC SNF QRP and Discharge to Community-PAC SNF QRP measures in calendar year 2018 based on discharges from October 1, 2016 through September 30, 2017. In the FY 2019 SNF PPS proposed rule (83 FR 21084), we proposed to increase the number of years of data used to calculate the Medicare Spending Per Beneficiary-PAC SNF QRP and Discharge to Community-PAC SNF QRP measures for purposes of display from 1 year to 2 years. Under this proposal, data on these measures would be publicly reported in CY 2019, or as soon thereafter as operationally feasible, based on discharges from October 1, 2016 through September 30, 2018.

Increasing the measure calculation and public display periods from 1 to 2 years of data increases the number of SNFs with enough data adequate for public reporting for the Medicare Spending Per Beneficiary-PAC SNF QRP measure from 86 percent (based on 2016 Medicare FFS claims data) to 95 percent (based on 2015 through 2016 Medicare FFS claims data), and for the Discharge to Community-PAC SNF QRP measure from 83 percent (based on 2016 Medicare FFS claims data) to 94 percent (based on 2015 through 2016 Medicare FFS claims data). Increasing measure public display periods to 2 years also aligns with the public display periods of these measures in the IRF QRP and LTCH QRP.

We also proposed to begin publicly displaying data in CY 2020, or as soon thereafter as is operationally feasible, on the following four assessment-based measures: (1) Application of IRF Functional Outcome Measure: Change in Self-Care Score for Medical Rehabilitation Patients (NQF #2633) (Change in Self-Care Score); (2) Application of IRF Functional Outcome Measure: Change in Mobility Score for Medical Rehabilitation Patients (NQF #2634) (Change in Mobility Score); (3) Application of IRF Functional Outcome Measure: Discharge Self-Care Score for Medical Rehabilitation Patients (NQF #2635) (Discharge Self-Care Score); and (4) Application of IRF Functional Outcome Measure: Discharge Mobility Score for Medical Rehabilitation Patients (NQF #2636) (Discharge Mobility Score). SNFs are required to submit data on these four assessment-based measures with respect to admissions as well as discharges occurring on or after October 1, 2018. We proposed to display data for these assessment-based measures based on 4 rolling quarters of data, initially using 4 quarters of discharges from January 1, 2019 through December 31, 2019. To ensure the statistical reliability of the measure rates for these four assessment-based measures, we also proposed that if a SNF has fewer than 20 eligible cases during any 4 consecutive rolling quarters of data that we are displaying for any of these measures, then we would note in our public display of that measure that with respect to that SNF, the number of cases/resident stays is too small to publicly report.

Comment: One commenter supported the proposal to publicly display the four SNF functional outcome measures on the SNF Compare website in CY 2020.

Response: We thank the commenter for the support.

Comment: Several commenters, including MedPAC, supported increasing the number of years of data used to calculate the Medicare Spending per Beneficiary-PAC SNF QRP and Discharge to Community-PAC SNF QRP measures from 1 year to 2 years to increase the number of providers that can be included in public reporting and also to align the measurement period with that used in other PAC settings. One commenter was concerned that increasing the measurement period to 2 years would penalize facilities that showed improvement in a one-year period, as the data would be aggregated across 2 years. Two commenters agreed with increasing the measurement period from 1 to 2 years but questioned the usefulness of a measure that they stated required a significant adjustment in collection methods to acquire data necessary to calculate a rate.

Response: We thank MedPAC and the other commenters for their support to increase the number of years of data used to calculate the Medicare Spending per Beneficiary-PAC SNF QRP measure and Discharge to Community-PAC SNF QRP measure from 1 to 2 years. We appreciate the commenter's concern about the impact of aggregating data across 2 years on the ability to demonstrate improvement in a 1-year period; however, we believe that the benefit of increasing the number of SNFs in public reporting outweighs the expressed concern associated with increasing the measurement period to 2 years because it would provide more information to consumers who may have a limited number of SNFs in their area. Further, improvements in 1-year period will be included in the 2-year data, so providers' efforts to improve can still be reflected in their measure scores. The proposed change will also align with the measurement period of the three claims-based measures (Medicare Spending per Beneficiary, Discharge to Community, and Potentially Preventable Readmissions) across the IRF, LTCH, and SNF QRPs.

Comment: MedPAC suggested that if CMS increases the measurement period for the Medicare Spending per Beneficiary PAC SNF QRP measure and Discharge to Community PAC SNF QRP measure to 2 years, CMS could consider giving more weight to the most recent performance year. MedPAC also suggested that CMS reconsider the approach to establishing minimum counts of episodes for public reporting of the Medicare Spending per Beneficiary-PAC SNF QRP measure to ensure accurate representation of a provider's performance.

Response: We thank MedPAC for its suggestion to consider greater weighting of the most recent year of data and to reconsider the approach to establishing minimum counts of episodes for public reporting. We will consider testing these suggestions in the future.

Comment: A commenter noted the importance of understanding the relationship between the Medicare Spending per Beneficiary-PAC SNF QRP measure, quality, and beneficiary out-of-pocket expenses. The commenter also noted the importance of educating consumers on this measure. The commenter suggested that CMS analyze these relationships further and define a strategy for interpreting the results before making the measure results public. Another commenter noted that facilities should not be penalized for decisions made by physicians that are beyond providers' control.

Response: We thank the commenters for the suggestions for additional analyses on the relationship between the Medicare Spending per Beneficiary-PAC SNF QRP measure, quality, and out-of-pocket spending. We will consider analyses on these topics in the future. Regarding beneficiary education for interpreting results, we will continue to work to develop language to support beneficiary understanding of the measures in public reporting. Regarding the comment on facility penalty for physician decision-making, the measure is intended to promote care coordination and improve efficiency by creating a continuum of accountability between Medicare providers.

Comment: Some commenters suggested that the public reporting of the SNF functional outcome measures: (1) Change in Self-Care Score; (2) Change in Mobility Score; (3) Discharge Self-Care Score; and (4) Discharge Mobility Score, on the SNF Compare website be delayed beyond CY 2020. One commenter suggested that the reporting be delayed until additional measures that address the maintenance of functional abilities are also developed and reported alongside the functional improvement measures and also encouraged the development of measures related to other nursing goals. Other commenters suggested that CMS reconsider publicly reporting the SNF functional quality measures in CY 2020 if these measures do not receive NQF endorsement prior to public display.

Response: We thank the commenters for their suggestions. We addressed the importance of measuring functional maintenance for SNF residents in the FY 2018 SNF PPS final rule (82 FR 36588). We interpret the commenter's recommendation of “at least one nursing goals measure” to refer to the development of new measures relating to functional status for SNF residents. We support future quality measurement work that will address the development of other measures that focus on maintaining function and the slowing of functional decline. We agree that the NQF endorsement process is an important part of measure development. The four functional outcome quality measures that we proposed to publicly report are NQF-endorsed for the IRF setting, and we plan to submit these four assessment-based measures to NQF for endorsement consideration in the SNF setting as soon as feasible.

After consideration of public comments we received, we are finalizing our proposal, to increase the number of years of data used to calculate the Medicare Spending per Beneficiary-PAC SNF QRP measure and Discharge to Community-PAC SNF QRP measure for purposes of public display from 1 to 2 years, starting in CY 2019 or as soon thereafter as operationally feasible. We are also finalizing our proposal to begin publicly displaying data in CY 2020, or as soon thereafter as is operationally feasible, on the following four assessment-based measures: (1) Application of IRF Functional Outcome Measure: Change in Self-Care Score for Medical Rehabilitation Patients (NQF #2633); (2) Application of IRF Functional Outcome Measure: Change in Mobility Score for Medical Rehabilitation Patients (NQF #2634); (3) Application of IRF Functional Outcome Measure: Discharge Self-Care Score for Medical Rehabilitation Patients (NQF #2635); and (4) Application of IRF Functional Outcome Measure: Discharge Mobility Score for Medical Rehabilitation Patients (NQF #2636).

C. Skilled Nursing Facility Value-Based Purchasing Program (SNF VBP)

1. Background

Section 215(b) of the Protecting Access to Medicare Act of 2014 (PAMA) (Pub. L. 113-93) authorized the SNF VBP Program (the “Program”) by adding section 1888(h) to the Act. As a prerequisite to implementing the SNF VBP Program, in the FY 2016 SNF PPS final rule (80 FR 46409 through 46426), we adopted an all-cause, all-condition hospital readmission measure, as required by section 1888(g)(1) of the Act. In the FY 2017 SNF PPS final rule (81 FR 51986 through 52009), we adopted an all-condition, risk-adjusted potentially preventable hospital readmission measure for SNFs, as required by section 1888(g)(2) of the Act. In the FY 2018 SNF PPS final rule (82 FR 36608 through 36623), we adopted additional policies for the Program, including an exchange function methodology for disbursing value-based incentive payments.

Section 1888(h)(1)(B) of the Act requires that the SNF VBP Program apply to payments for services furnished on or after October 1, 2018. The SNF VBP Program applies to freestanding SNFs, SNFs affiliated with acute care facilities, and all non-CAH swing-bed rural hospitals. We believe the implementation of the SNF VBP Program is an important step towards transforming how care is paid for, moving increasingly towards rewarding better value, outcomes, and innovations instead of merely rewarding volume.

For additional background information on the SNF VBP Program, including an overview of the SNF VBP Report to Congress and a summary of the Program's statutory requirements, we refer readers to the FY 2016 SNF PPS final rule (80 FR 46409 through 46410). We also refer readers to the FY 2017 SNF PPS final rule (81 FR 51986 through 52009) for discussion of the policies that we adopted related to the potentially preventable hospital readmission measure, scoring, and other topics. Finally, we refer readers to the FY 2018 SNF PPS final rule (82 FR 36608 through 36623) for discussions of the policies that we adopted related to value-based incentive payments, the exchange function, and other topics.

We proposed additional requirements for the FY 2021 SNF VBP Program in the FY 2019 SNF PPS proposed rule (83 FR 21084 through 21089). We received several general comments on the SNF VBP Program.

Comment: One commenter supported our goal of reducing preventable hospital readmissions, noting that those readmissions increase costs for the Medicare program, significantly affect beneficiaries, and increase the likelihood of medical errors related to care coordination.

Response: We thank the commenter for this support.

Comment: A commenter suggested that we consider developing an integrated approach that provides incentives to SNFs to accept more medically complex patients and promotes readmission prevention. The commenter suggested that, while the PDPM and SNF VBP Programs are authorized separately, integrating them might be helpful to that end, and could include payments for telemedicine, post-discharge care coordination, and training on readmission prevention protocols and refinements to Interrupted Stay policies. The commenter stated that readmissions prevention strategies can be very effective at saving Medicare spending and improving the patient experience, but can also require initial investments in technology and staff training.

Response: We agree that readmission prevention strategies can be effective at saving Medicare spending and improving the patient experience. At this time, we do not believe it is possible to integrate the PDPM and SNF VBP Program given their separate authorities and purposes. However, we will continue to monitor the effects of the SNF VBP Program and the case-mix classification methodology in the SNF prospective payment system, including the PDPM.

Comment: One commenter encouraged us to make public as much SNF VBP data as possible on Nursing Home Compare and data.medicare.gov, including individual facilities' baseline and performance period readmissions rates, achievement and improvement points, performance scores, rankings, and value-based incentive payment percentages. The commenter noted that CMS has provided most of this type of information for other programs, and that the public should expect the same level of transparency from SNF VBP.

Response: We agree with the comment and intend to be as transparent as possible in order to inform consumer decision-making, quality improvement initiatives, and high quality patient care. As required by section 1888(h)(9) of the Act, we will publish facility performance information, including SNF performance scores and rankings, the range of SNF performance scores, the number of SNFs receiving value-based incentive payments, and the range and total amounts of those payments, on the Nursing Home Compare website.

Comment: One commenter requested that we ensure that specialty populations such as children, patients with HIV/AIDS, ventilator-dependent patients, and those with Huntington's disease or other neurodegenerative disorders, do not experience unintended negative results based on the SNF VBP Program's incentives.

Response: We monitor numerous aspects of the SNF VBP Program, including trends in measure rates, SNF performance scores, and starting with FY 2019, value-based incentive payment percentages and their effects on SNFs' care quality and on beneficiaries' access to care. We understand the commenter's concerns about specialty patient populations, and we will continue working to ensure that such populations do not experience unintended consequences because of the SNF VBP Program.

2. Measures

For background on the measures we have adopted for the SNF VBP Program, we refer readers to the FY 2016 SNF PPS final rule (80 FR 46419), where we finalized the Skilled Nursing Facility 30-Day All-Cause Readmission Measure (SNFRM) (NQF #2510) that we are currently using for the SNF VBP Program. We also refer readers to the FY 2017 SNF PPS final rule (81 FR 51987 through 51995), where we finalized the Skilled Nursing Facility 30-Day Potentially Preventable Readmission Measure (SNFPPR) that we will use for the SNF VBP Program instead of the SNFRM as soon as practicable, as required by statute.

We did not propose any changes to the Program's measures. However, we received several comments on the Program's measures.

Comment: One commenter requested that we announce when we will transition the SNF VBP Program to a measure of potentially preventable readmissions rather than the current all-cause readmissions measure. One commenter recommended that we not replace the SNFRM with the SNFPPR before FY 2021 to allow SNFs time to adjust to the SNFRM and other measures of readmissions. Another commenter encouraged us to transition the Program to the measure of potentially preventable readmissions, stating that the PPR will exclude planned readmissions that are not considered a negative outcome, and therefore, should not be counted against SNFs. Other commenters urged us to seek NQF endorsement and input from the Measure Applications Partnership as soon as possible on the SNFPPR, and requested that we provide a timeline for when we will replace the all-cause measure with the SNFPPR. Another commenter requested that we consider standardizing and consolidating various SNF hospitalization measures used in Medicare to focus SNFs' quality improvement efforts. The commenter noted that state initiatives may also have similar measures based on different data, and that the multitude of hospitalization measures may be confusing for consumers and may dilute provider improvement efforts.

Response: We sought input from the MAP on the SNFPPR prior to proposing it for adoption in the SNF VBP. The MAP published its views in a February 2016 report, as we described in the FY 2017 SNF PPS final rule (81 FR 51989 through 51990). In that report,^[8] MAP noted the statutory requirement that we specify a measure of potentially preventable readmissions for the SNF VBP Program, and explained support for the importance of the measure and its acknowledgement that “readmission for the SNF setting is not an occasional occurrence.” MAP's report also noted public commenters' input, including general support for the recommendation to “encourage continued development” of the SNFPPR and some concerns about the measure's specifications and MAP's making a recommendation on a measure that is not fully tested. Regarding submission of the SNFPPR for consensus endorsement, we currently plan to submit the measure for NQF endorsement in 2019 upon completion of additional testing. We plan to propose transitioning to this measure after the completion of the endorsement process.

We also acknowledge the commenter's concern about the number of hospitalization measures in Medicare and in other quality programs, including those used by the states. We will consider how we might further streamline our quality programs, particularly under the Meaningful Measures Initiative. However, we note that all rehospitalization measures share the same underlying care focus—that is, avoiding rehospitalizations—even if they vary somewhat in the specifics of which hospitalizations they measure. We continue to believe that SNFs working to improve care quality and minimize rehospitalizations for their patients will perform well on hospitalization measures.

We continue to determine when it is practicable to transition the Program to the measure of potentially preventable readmissions, and we will propose that transition in future rulemaking, which we believe will provide sufficient notice to SNFs about the quality measure that will form the basis for the SNF VBP Program. We intend to take all of the views expressed by public commenters into account when we make that decision, as well as the operational necessities of the Program (such as the time needed to calculate measure rates on the SNFPPR and how that time interacts with the Program's performance and baseline periods). However, we would like to clarify that the SNFRM currently excludes certain planned readmissions.

Comment: One commenter stated that the SNF VBP Program should consist of more than just one hospital readmissions measure, and encouraged us to work with Congress to include additional measures in the Program, potentially including those currently displayed on Nursing Home Compare, were part of the SNF VBP demonstration, or are part of the SNF QRP. The commenter also specifically suggested measures including turnover as a percentage of nursing staff, total CNA hours per patient day, and total licensed nursing hours per patient day, noting that higher staffing levels are correlated with higher quality of care outcomes.

Response: We thank the commenter for these suggestions. As the commenter noted, any changes to expand the SNF VBP Program's measure set would require Congressional action.

Comment: One commenter expressed concern about the data elements that SNFs must document to track their performance on the SNFRM, noting that they are different than those used for the CMS Star Ratings. The commenter also urged us to better align the measures between the SNF QRP and SNF VBP Programs, stating that SNFs want harmonization in what they are required to collect, document, and extract for performance tracking and improvement purposes.

Response: SNFs may choose to track readmissions to the hospitals as part of their quality improvement efforts, and we note that the measures that we have specified for the SNF VBP program impose no data collection requirements on SNFs. Additionally, while we understand the potential benefits of quality measure alignment between the SNF QRP and SNF VBP Programs, we do not believe that this type of alignment meets the SNF VBP Program's needs at this time. While we generally agree that aligning measures across programs is ideal, we hesitate to do so when it is inappropriate to the programs and does not align with statutory direction. In this case, aligning with the SNF QRP readmission measure would require the SNF VBP Program to ignore readmissions that occur during the SNF stay, and we believe this is inappropriate to a value-based purchasing program intended to reduce readmissions among SNF patients in accordance with the statute. Likewise, the SNF QRP readmission measure must follow a statutory requirement to align with readmission measures in other post-acute QRPs that are not compatible with the needs of the SNF VBP program.

We thank the commenters for their feedback on SNF VBP measures.

a. Accounting for Social Risk Factors in the SNF VBP Program

In the FY 2018 SNF PPS final rule (82 FR 36611 through 36613), we discussed the importance of improving beneficiary outcomes including reducing health disparities. We also discussed our commitment to ensuring that medically complex patients, as well as those with social risk factors, receive excellent care. We discussed how studies show that social risk factors, such as being near or below the poverty level as determined by HHS, belonging to a racial or ethnic minority group, or living with a disability, can be associated with poor health outcomes and how some of this disparity is related to the quality of health care.^[9] Among our core objectives, we aim to improve health outcomes, attain health equity for all beneficiaries, and ensure that complex patients, as well as those with social risk factors, receive excellent care. Within this context, reports by the Office of the Assistant Secretary for Planning and Evaluation (ASPE) and the National Academy of Medicine have examined the influence of social risk factors in our value-based purchasing programs.^[10] As we noted in the FY 2018 SNF PPS final rule (82 FR 36611), ASPE's report to Congress found that, in the context of value-based purchasing programs, dual eligibility for Medicare and Medicaid was the most powerful predictor of poor health care outcomes among those social risk factors that they examined and tested. In addition, as noted in the FY 2018 SNF PPS final rule, the National Quality Forum (NQF) undertook a 2-year trial period in which certain new measures and measures undergoing maintenance review have been assessed to determine if risk adjustment for social risk factors is appropriate for these measures.^[11] The trial period ended in April 2017 and a final report is available at http://www.qualityforum.org/​SES_​Trial_​Period.aspx. The trial concluded that “measures with a conceptual basis for adjustment generally did not demonstrate an empirical relationship” between social risk factors and the outcomes measured. This discrepancy may be explained in part by the methods used for adjustment and the limited availability of robust data on social risk factors. NQF has extended the socioeconomic status (SES) trial,^[12] allowing further examination of social risk factors in outcome measures.

In the FY 2018/CY 2018 proposed rules for our quality reporting and value-based purchasing programs, we solicited feedback on which social risk factors provide the most valuable information to stakeholders and the methodology for illuminating differences in outcomes rates among patient groups within a provider that would also allow for a comparison of those differences, or disparities, across providers. Feedback we received across our quality reporting programs included encouraging CMS to explore whether factors could be used to stratify or risk adjust the measures (beyond dual eligibility); to consider the full range of differences in patient backgrounds that might affect outcomes; to explore risk adjustment approaches; and to offer careful consideration of what type of information display would be most useful to the public.

We also sought public comment on confidential reporting and future public reporting of some of our measures stratified by patient dual eligibility. In general, commenters noted that stratified measures could serve as tools for hospitals to identify gaps in outcomes for different groups of patients, improve the quality of health care for all patients, and empower consumers to make informed decisions about health care. Commenters encouraged us to stratify measures by other social risk factors such as age, income, and educational attainment. With regard to value-based purchasing programs, commenters also cautioned CMS to balance fair and equitable payment while avoiding payment penalties that mask health disparities, or discouraging the provision of care to more medically complex patients. Commenters also noted that value-based payment program measure selection, domain weighting, performance scoring, and payment methodology must account for social risk.

We stated in the FY 2019 SNF VBP PPS proposed rule that as a next step, we are considering options to improve health disparities among patient groups within and across hospitals, SNFs, and other health care providers by increasing the transparency of disparities as shown by quality measures. We also stated that we are considering how this work applies to other CMS quality programs in the future. We refer readers to the FY 2018 IPPS/LTCH PPS final rule (82 FR 38403 through 38409) for more details, where we discuss the potential stratification of certain Hospital Inpatient Quality Reporting Program outcome measures. Furthermore, we stated that we continue to consider options to address equity and disparities in our value-based purchasing programs.

We plan to continue working with ASPE, the public, and other key stakeholders on this important issue to identify policy solutions that achieve the goals of attaining health equity for all beneficiaries and minimizing unintended consequences.

We received several comments on our discussion of social risk factors.

Comment: One commenter suggested that we consider adjusting for social risk factors through peer grouping to avoid masking disparities in clinical performance. The commenter also suggested that we target technical assistance resources to low-performing providers and support research to reduce measurement bias. Another commenter was concerned that we had not yet adjusted the SNF Readmission Measure based on socioeconomic factors. The commenter expressed concern that we would score SNFs unfairly due to more challenging case mixes, and stated that we must adjust readmission scores to avoid unfair payment penalties for those SNFs serving patient populations with lower socioeconomic status. One commenter acknowledged that we are required by statute to adopt a measure of all-cause readmissions, but expressed concerns about the SNFRM due to its lack of risk adjustment for socioeconomic status, its lacking focus on preventable readmissions, and some design elements. The commenter encouraged us to create a socioeconomic status risk adjustment for this measure, noting that SNFs in underserved areas predominantly caring for low-income, dual-eligible residents may be penalized by measures of all-cause readmissions. Another commenter urged us to include risk adjustment for socioeconomic status for any readmission measures adopted under the SNF VBP Program. The commenter concurred with the December 2016 Report to Congress on Social Risk Factors' conclusion that social risk factors are essential determinants of health and stated that the IMPACT Act provides CMS with a wealth of patient-specific data that it can use to develop additional risk adjustment policies. The commenter encouraged us to use those data to adjust SNF VBP measures and provide incentives to SNFs caring for patients with social risk factors.

Response: We thank the commenters for these suggestions and will take them into account as we develop additional policies on social risk factors in the future. However, in response to the commenter who expressed concern about the current SNF Readmission Measure, we note that the SNF Readmission Measure includes the following case-mix adjustments that we believe promote fairness in the application of financial penalties: Demographic characteristics (age and sex), principal diagnosis from the Medicare claim corresponding to the prior proximal hospitalization as categorized by AHRQ's Clinical Classification Software (CCS) groupings, length of stay during the patient's prior proximal hospitalization, length of stay in the intensive care unit (ICU), end-stage renal disease (ESRD) status, the patient's disability status, the number of prior hospitalizations in the previous 365 days, system-specific surgical indicators, individual comorbidities as grouped by Hierarchical Condition Categories (HCCs) or other comorbidity indices, and a variable counting the number of comorbidities if the patient had more than two HCCs. We refer readers to the FY 2016 SNF PPS final rule (80 FR 46411 through 46419) for additional technical details on the SNFRM.

Comment: One commenter encouraged us to continue using findings from the NQF Sociodemographic Status trial to inform our efforts to address equity and disparities in our VBP Programs, but recommended that we not add SES covariates to the SNFRM risk adjustment model as that action may create biases in reporting, undermine system-based approaches to providing high quality care, and create access to care problems.

Response: We remain concerned about the possibility that additional risk adjustment may mask important performance differences for providers and suppliers that treat patients with additional comorbidities or complications, and we will continue studying the issue. We intend to monitor NQF's ongoing work on this topic carefully.

Comment: One commenter agreed with recommendations to incorporate social risk factors in risk adjustment, but was not sure about which risk characteristics are available in the Medicare eligibility files and whether those characteristics have been evaluated independently. The commenter also suggested that we coordinate research efforts with states that may already be conducting work in this area. One commenter urged us to incorporate risk adjustment for sociodemographic and socioeconomic status into SNF VBP measures, but expressed support for the continued use of unadjusted data for measures related to items that are within the SNF's control. The commenter urged us to make unadjusted performance measure data available to the public to ensure that analysis of health care disparities can continue, and until risk-adjusted measures are available, to report stratified measure rates to the public. The commenter also expressed support for alternative payment mechanisms that account for the complexities of extremely disadvantaged patients, and called on us to monitor the effects of our quality improvement programs carefully. Another commenter supported our continued evaluation of social determinants of health, including providers' commitments to caring for the Medicaid population, and their impact on our payment systems. The commenter encouraged us to ensure that our payment methodologies are updated consistently to account for these factors and maintain equitable access to care.

Response: We intend to continue working with states and other stakeholders to the greatest extent possible to understand the challenges associated with additional risk adjustment for socioeconomic and sociodemographic status in quality measurement programs, including assessing the appropriateness of incorporating specific risk factors in the risk adjustment models. That work includes identifying appropriate data sources for social risk factors, and we will consider the commenter's point about the Medicare eligibility files as a potential data source.

We agree with the commenters that studying health care disparities is critically important for the health care system, and we will continue to do so. We will also take that point under consideration as we consider social risk factors adjustment policies for the SNF VBP Program in the future. We will continue monitoring the SNF VBP Program to ensure that Medicare beneficiaries maintain access to needed SNF care.

We thank the commenters for this feedback, and will take it account as we consider the appropriateness of accounting for social risk factors in the SNF VBP Program.

3. Performance Standards

a. FY 2021 Performance Standards

We refer readers to the FY 2017 SNF PPS final rule (81 FR 51995 through 51998) for a summary of the statutory provisions governing performance standards under the SNF VBP Program and our finalized performance standards policy, as well as the numerical values for the achievement threshold and benchmark for the FY 2019 program year. We also responded to public comments on these policies in that final rule.

We published the final numerical values for the FY 2020 performance standards in the FY 2018 SNF PPS final rule (82 FR 36613), and for reference, we are displaying those values again in Table 40.

Table 40—Final FY 2020 SNF VBP Program Performance Standards

Measure ID	Measure description	Achievement threshold	Benchmark
SNFRM	SNF 30-Day All-Cause Readmission Measure (NQF #2510)	0.80218	0.83721

We will continue to use the achievement threshold and benchmark as previously finalized in the FY 2018 SNF PPS final rule. However, due to timing constraints associated with the compilation of the FY 2017 MedPAR file to include 3 months of data following the last discharge date, we were unable to provide estimated numerical values for the FY 2021 Program year's performance standards in the proposed rule. As discussed further below, we proposed to adopt FY 2017 as the baseline period for the FY 2021 program year. While we did not expect either the achievement threshold or benchmark to change significantly from what was finalized for the FY 2020 Program year, we stated our intent to publish the final numerical values for the performance standards based on the FY 2017 baseline period in the FY 2019 SNF PPS final rule.

We welcomed public comment on this approach.

Comment: One commenter urged us to score SNFs on achievement only, stating that Medicare's quality programs should reward providers based on clear, absolute, and prospectively set performance targets.

Response: While we appreciate this suggestion, we note that we are required by section 1888(h)(3)(B) of the Act to establish performance standards that include levels of achievement and improvement, and to use to the higher of either improvement or achievement when calculating the SNF performance score.

Comment: One commenter stated its understanding of the timing constraints that we discussed with respect to the MedPAR file for performance standards calculations, and reiterated its prior support for the Program's switch to fiscal year instead of calendar year performance periods.

Response: We appreciate the continued support for the policy we finalized in the FY 2018 SNF PPS final rule (82 FR 36613 through 36614) to change the SNF VBP Program's performance and baseline periods from calendar years to fiscal years. Additionally, as we note further below, we are finalizing FY 2019 (October 1, 2018 through September 30, 2019) as the performance period for the FY 2021 SNF VBP Program year.

Comment: One commenter supported our efforts to measure improvement and encouraged us to reward providers that consistently achieve high performance under the SNF VBP Program.

Response: We believe that the performance standards that we are adopting, which include levels of achievement and improvement as required by the SNF VBP Program's statute, continue to offer opportunities for us to recognize both SNFs that achieve high performance and those SNFs that improve over time.

After consideration of the public comments, we are finalizing the numerical values for the FY 2021 SNF VBP Program based on the FY 2017 baseline period. Those values follow below in Table 41.

Table 41—Final FY 2021 SNF VBP Program Performance Standards

Measure ID	Measure description	Achievement threshold	Benchmark
SNFRM	SNF 30-Day All-Cause Readmission Measure (NQF #2510)	0.79476	0.83212

b. Correction of Performance Standard Numerical Values in Cases of Errors

As noted previously, section 1888(h)(3)(C) of the Act requires that we establish and announce the performance standards for a fiscal year not later than 60 days prior to the performance period for the fiscal year involved. However, we currently do not have a policy that would address the situation where, subsequent to publishing the numerical values for the finalized performance standards for a program year, we discover an error that affects those numerical values. Examples of the types of errors that we could subsequently discover are inaccurate variables on Medicare claims, programming errors, excluding data should have been included in the performance standards calculations, and other technical errors that resulted in inaccurate achievement threshold and benchmark calculations. While we do not have reason to believe that the SNF VBP Program has previously published inaccurate numerical values for performance standards, in the FY 2019 SNF PPS proposed rule (83 FR 21086), we stated our concern about the possibility that we would discover an error in the future and have no ability to correct the numerical values.

We are aware that SNFs rely on the performance standards that we publicly display in order to target quality improvement efforts, and we do not believe that it would be fair to SNFs to repeatedly update our finalized performance standards if we were to identify multiple errors. In order to balance the need of SNFs to know what performance standards they will be held accountable to for a SNF VBP program year with our obligation to provide SNFs with the most accurate performance standards that we can based on the data available at the time, we proposed that if we discover an error in the calculations subsequent to having published the numerical values for the performance standards for a program year, we would update the numerical values to correct the error. We also proposed that we would only update the numerical values one time, even if we subsequently identified a second error, because we believe that a one-time correction would allow us to incorporate new information into the calculations without subjecting SNFs to multiple updates. Any update we would make to the numerical values based on a calculation error would be announced via the CMS website, listservs, and other available channels to ensure that SNFs are made fully aware of the update.

We welcomed public comments on this proposal.

Comment: One commenter supported our proposal to adopt correction authority for performance standards and agreed that making multiple changes to the performance standards in a given program year would be difficult for SNFs' quality improvement efforts. The commenter also urged us to be transparent if we find additional technical errors.

Response: We thank the commenter for the support, and we intend to be as transparent as possible if we identify any errors in the calculation of the numerical values of the SNF VBP Program's performance standards.

After consideration of the public comments we have received, we are finalizing our policy to correct performance standard numerical values in cases of errors as proposed.

4. FY 2021 Performance Period and Baseline Period and for Subsequent Years

a. Background

We refer readers to the FY 2016 SNF PPS final rule (80 FR 46422) for a discussion of our considerations for determining performance periods under the SNF VBP Program. Based on those considerations, as well as public comment, we adopted CY 2017 as the performance period for the FY 2019 SNF VBP Program, with a corresponding baseline period of CY 2015.

Additionally, in the FY 2018 SNF PPS final rule (82 FR 36613 through 36614), we adopted FY 2018 as the performance period for the FY 2020 SNF VBP Program, with a corresponding baseline period of FY 2016. We refer readers to that rule for a discussion of the need to shift the Program's measurement periods from the calendar year to the fiscal year.

b. FY 2021 Performance and Baseline Periods

As we discussed with respect to the FY 2019 and FY 2020 SNF VBP Program years, we continue to believe that a 12-month duration for the performance and baseline period is the most appropriate for the SNF VBP Program. Therefore, we proposed to adopt FY 2019 (October 1, 2018 through September 30, 2019) as the performance period for the FY 2021 SNF VBP Program year. We also proposed to adopt FY 2017 (October 1, 2016 through September 30, 2017) hospital discharges as the baseline period for the FY 2021 SNF VBP Program year.

We welcomed public comment on these proposals.

Comment: One commenter expressed concern about our proposal to use FY 2019 as the performance period for the FY 2021 SNF VBP Program year, stating that SNFs need more time to improve their data collection, reporting, and evaluation efforts. The commenter requested that we align our measures with the SNF QRP and other quality programs, which will allow SNFs additional time for performance tracking and improvement activities. The commenter also requested that we provide SNFs with more timely performance feedback to help them identify areas for improvement efforts. One commenter expressed concern about the proposed performance period, stating that SNFs do not believe they are ready for FY 2019 to be used as the performance period and indicated that the collection and reporting of quality measures is a significant administrative burden. The commenter urged us to move to an automated system to reduce the reporting burden on SNFs and requested that we provide SNFs with timely performance feedback that they can use to identify areas where they need to focus their improvement efforts.

Response: We would like to clarify for the commenter that the SNF VBP Program's measure is calculated based on hospital claims, and therefore, does not require data collection or impose any reporting burden on SNFs, though SNFs may choose to track readmissions to the hospital for their patients as part of their care coordination and quality improvement efforts. We do not believe that SNFs need additional time to track readmissions to the hospital for their patients or to undertake quality improvement efforts to minimize those readmissions because SNFs have had ample notice about the SNF VBP Program's operations and its focus on measures of hospital readmissions. We will, however, strive to provide as much timely information to SNFs as possible on their measured performance, but we note that the measure that we have specified for the Program includes significant calculations, including detailed risk adjustment, that complicates our intention to provide feedback more promptly than on a quarterly basis to SNFs.

Comment: One commenter supported our performance and baseline period proposals and agreed that 12-month periods are appropriate for both the SNFRM and the SNFPPR.

Response: We thank the commenter for the support.

After consideration of the public comments that we received, we are finalizing the performance period and baseline period for FY 2021 as proposed.

c. Performance Periods and Baseline Periods for Subsequent Program Years

As we have described in previous rules (see, for example, the FY 2016 SNF PPS final rule, 80 FR 46422), we strive to link performance furnished by SNFs as closely as possible to the program year to ensure clear connections between quality measurement and value-based payment. We also strive to measure performance using a sufficiently reliable population of patients that broadly represent the total care provided by SNFs.

Therefore, we proposed that beginning with the FY 2022 program year and for subsequent program years, we would adopt for each program year, a performance period that is the 1-year period following the performance period for the previous program year. We also proposed that beginning with the FY 2022 program year and for subsequent program years, we would adopt for each program year a baseline period that is the 1-year period following the baseline period for the previous year. Under this policy, the performance period for the FY 2022 program year would be FY 2020 (the 1-year period following the proposed FY 2021 performance period of FY 2019), and the baseline period for the FY 2022 program year would be FY 2018 (the 1-year period following the proposed FY 2021 baseline period of FY 2017). We believe adopting this policy will provide SNFs with certainty about the performance and baseline periods during which their performance will be assessed for future program years.

We welcomed public comments on this proposal.

Comment: One commenter supported our proposal to adopt performance and baseline periods automatically for subsequent program years.

Response: We thank the commenter for the support.

After consideration of the public comments that we have received, we are finalizing our policy to adopt performance periods and baseline periods for subsequent program years as proposed.

5. SNF VBP Performance Scoring

a. Background

We refer readers to the FY 2017 SNF PPS final rule (81 FR 52000 through 52005) for a detailed discussion of the scoring methodology that we have finalized for the Program, along with responses to public comments on our policies and examples of scoring calculations. We also refer readers to the FY 2018 SNF PPS final rule (82 FR 36614 through 36616) for discussion of the rounding policy we adopted, our request for comments on SNFs with zero readmissions, and our request for comments on a potential extraordinary circumstances exception policy.

b. Scoring Policy for SNFs Without Sufficient Baseline Period Data

In some cases, a SNF will not have sufficient baseline period data available for scoring for a Program year, whether due to the SNF not being open during the baseline period, only being open for a small portion of the baseline period, or other reasons (such as receiving an extraordinary circumstance exception, which we finalize below). The availability of baseline data for each SNF is an integral component of our scoring methodology, and we are concerned that the absence of sufficient baseline data for a SNF will preclude us from being able to score that SNF on improvement for a program year. As discussed further below, with respect to the proposed scoring adjustment for a SNF without sufficient data in the performance period to create a reliable SNF performance score, we are concerned that measuring SNFs with fewer than 25 eligible stays (or index SNF stays that would be included in the calculation of the SNF readmission measure) during the baseline period may result in unreliable improvement scores, and as a result, unreliable SNF performance scores. We considered policy options to address this issue.

We continue to believe it is important to compare SNF performance during the same periods to control for factors that may not be attributable to the SNF, such as increased patient case-mix acuity during colder weather periods when influenza, pneumonia, and other seasonal conditions and illnesses are historically more prevalent in the beneficiary population. Using a 12-month performance and baseline period for all SNFs ensures that, to the greatest extent possible, differences in performance can be attributed to the SNF's care quality rather than to exogenous factors.

Additionally, because we have proposed that for FY 2021 and future Program years, the start of the performance period for a Program year would begin exactly 12-months after the end of the baseline period for that Program year and there would not be sufficient time to compute risk-standardized readmission rates from another 12-month baseline period before the performance period if a SNF had insufficient data during the baseline period. For the FY 2021 Program, for example, the proposed baseline period would conclude at the end of FY 2017 (September 30, 2017) and the proposed performance period would begin on the first day of FY 2019 (October 1, 2018). We also do not believe it would be equitable to score SNFs without sufficient baseline period data using data from a different period. Doing so would, in our view, impede our ability to compare SNFs' performance on the Program's quality measure fairly, as additional factors that may affect SNFs' care could arise when comparing performance during different time periods. Therefore, we have concluded that it is not operationally feasible or equitable to use different baseline periods for purposes of awarding improvement scores to SNFs for a Program year.

We believe that SNFs without sufficient data from a single baseline period, which we would define for this purpose as SNFs with fewer than 25 eligible stays during the baseline period for a fiscal year based on an analysis of Pearson correlation coefficients at various denominator counts, should not be measured on improvement for that Program year. Accordingly, we are proposing to score these SNFs based only on their achievement during the performance period for any Program year for which they do not have sufficient baseline period data. The analysis of Pearson correlation coefficients at various denominator counts used in developing this proposal is available on our website at https://www.cms.gov/​Medicare/​Quality-Initiatives-Patient-Assessment-Instruments/​Value-Based-Programs/​Other-VBPs/​SNFRM-Reliability-Testing-Memo.pdf.

We proposed to codify this proposal by adding § 413.338(d)(1)(iv). We welcomed public comment on this proposal.

Comment: One commenter agreed with our proposal to not score SNFs on improvement when they do not have sufficient data during the baseline period for appropriate year-over-year comparisons. However, the commenter expressed concern that this approach is different for this group of low-volume SNFs compared to SNFs that are consistently low-volume. The commenter expressed continued concerns with the readmission rates awarded to SNFs when they have low case volume.

Response: We note that the policies that we have proposed for SNFs without sufficient baseline period data and for low-volume adjustment are intended to address separate permutations of the SNFRM reliability issue. In the first case, our intent is to ensure that we compare sufficiently-reliable SNFRM rates when assessing SNFs' improvement over time. That assessment relies on SNFRM rates being sufficiently reliable in both the baseline period and performance period to make the comparison that we use to award improvement points. In contrast, the low-volume scoring adjustment proposal focuses on the SNFRM's reliability during the performance period, which is necessary for both achievement and improvement scoring. We believe that these proposals ensure that SNFRM rates are sufficiently reliable for purposes of SNF VBP scoring, and as the commenter requested, ensure that SNFs are not scored on the SNFRM when the measure's case count is too low to produce sufficiently reliable scores.

Comment: One commenter supported our proposal to score SNFs without sufficient baseline period data on achievement only, agreeing with our view that measure results in those cases are susceptible to random variation and may not reliably represent quality in that facility.

Response: We thank the commenter for the support.

After consideration of the public comments that we received, we are finalizing our scoring policy for SNFs without sufficient baseline period data as proposed. We are also finalizing our regulation text on this policy as proposed.

c. SNF VBP Scoring Adjustment for Low-Volume SNFs

In previous rules, we have discussed and sought comment on policies related to SNFs with zero readmissions during the performance period. For example, in the FY 2018 SNF PPS rule (82 FR 36615 through 36616), we sought comment on policies we should consider for SNFs with zero readmissions during the performance period because under the risk adjustment and the statistical approach used to calculate the SNFRM, outlier values are shifted towards the mean, especially for smaller SNFs. As a result, SNFs with observed readmission rates of zero may receive risk-standardized readmission rates that are greater than zero. We continue to be concerned about the effects of the SNFRM's risk adjustment and statistical approach on the scores that we award to SNFs under the Program. We are specifically concerned that as a result of this approach, the SNFRM is not sufficiently reliable to generate accurate performance scores for SNFs with a low number of eligible stays during the performance period. We would like to ensure that the Program's scoring methodology results in fair and reliable SNF performance scores because those scores are linked to a SNF's ranking and payment.

Therefore, we considered whether we should make changes to our methodology for assessing the total performance of SNFs for a Program year that better accounts for SNFs with zero or low numbers of eligible stays during the performance period. Because the number of eligible SNF stays makes up the denominator of the SNFRM, we have concluded that the reliability of a SNF's measure rate and resulting performance score is adversely impacted if the SNF has less than 25 eligible stays during the performance period, as the Pearson correlation coefficient is lower at denominator counts of 5, 10, 15, and 20 eligible stays in comparison to 25 eligible stays. The analysis of Pearson correlation coefficients at various denominator counts used in developing this proposal is available on our website at https://www.cms.gov/​Medicare/​Quality-Initiatives-Patient-Assessment-Instruments/​Value-Based-Programs/​Other-VBPs/​SNFRM-Reliability-Testing-Memo.pdf.

We believe that the most appropriate way to ensure that low-volume SNFs (which we define for purposes of the SNF VBP Program as SNFs with fewer than 25 eligible stays during the performance period) receive sufficiently reliable SNF performance scores is to adopt an adjustment to the scoring methodology we use for the SNF VBP Program. We proposed that if a SNF has less than 25 eligible stays during a performance period for a Program year, we would assign a performance score to the SNF for that Program year. That assigned performance score would, when used to calculate the value-based incentive payment amount for the SNF, result in a value-based incentive payment amount that is equal to the adjusted Federal per diem rate that the SNF would have received for the fiscal year in the absence of the Program. The actual performance score that we would assign to an individual low-volume SNF for a Program year would be identified based on the distribution of all SNFs' performance scores for that Program year after calculating the exchange function. We would then assign that score to an individual low-volume SNF, and we would notify the low-volume SNF that it would be receiving an assigned performance score for the Program year in the SNF Performance Score Report that we provide not later than 60 days prior to the fiscal year involved.

We believe this scoring adjustment policy would appropriately ensure that our SNF performance score methodology is fair and reliable for SNFs with fewer than 25 eligible stays during the performance period for a Program year.

In section X.A.6. of the proposed rule, we estimated that $527.4 million would be withheld from SNFs' payments for the FY 2019 Program year based on the most recently available data. Additionally, the 60 percent payback percentage would result in an estimated $316.4 million being paid to SNFs in the form of value-based incentive payments with respect to FY 2019 services. Of the $316.4 amount, we estimated that $8.6 million will be paid to low-volume SNFs. However, if our proposal to adopt a scoring adjustment for low-volume SNFs were finalized, we estimated that we would redistribute an additional $6.7 million in value-based incentive payments to low-volume SNFs with respect to FY 2019 services, for a total of $15.3 million of the estimated $527.4 million available for value-based incentive payments for that Program year. The additional $6.7 million in value-based incentive payments that would result from finalizing this proposal would increase the 60 percent payback percentage for FY 2019 by approximately 1.28 percent, which would result in a payback percentage 61.28 percent of withheld funds. The payback percentage would similarly increase for all other Program years, however the actual amount of the increase for a particular Program year would vary based on the number of low-volume SNFs that we identify for that Program year and the distribution of all SNFs' performance scores for that Program year.

As an alternative, we considered assigning a performance score to SNFs with fewer than 25 eligible stays during the performance period that would result in a value-based incentive payment percentage of 1.2 percent, or 60 percent of the 2 percent withhold. This amount would match low-volume SNFs' incentive payment percentages with the finalized SNF VBP Program payback percentage of 60 percent, and would represent a smaller adjustment to low-volume SNFs' incentive payment percentages than the proposed policy described above. We estimated that this alternative would redistribute an additional $1 million with respect to FY 2019 services to low-volume SNFs. We also estimated that this alternative would increase the 60 percent payback percentage for FY 2019 by approximately 0.18 percent of the approximately $527.4 million of the total withheld from SNFs' payments, which would result in a payback percentage of 60.18 percent of the estimated $527.4 million in withheld funds for that Program year. However, as with the proposal above, we stated that the specific amount by which the payback percentage would increase for each Program year would vary based on the number of low-volume SNFs that we identify for each Program year and the distribution of all SNFs' performance scores for that Program year.

We welcomed public comments on this proposal and on the alternative that we considered. We also proposed to codify the definition of low-volume SNF at § 413.338(a)(16), and the definition of eligible stay at § 413.338(a)(17). We proposed to codify the low-volume scoring adjustment proposal at § 413.338(d)(3). We also proposed a conforming edit to the payback percentage policy at § 413.338(c)(2)(i).

Comment: One commenter expressed support for our proposed low-volume adjustment that would provide SNFs with a neutral value-based incentive payment percentage.

Response: We thank the commenter and appreciate the support.

Comment: One commenter requested clarification on how the SNF VBP will affect newly certified facilities that have no data from either the performance period or baseline period for the FY 2019 SNF VBP Program year.

Response: SNFs with zero eligible stays during both the baseline and performance periods are not covered by the low-volume adjustment policy. For the purposes of the SNF readmission measure, an eligible stay is an index SNF admission that would be included in the denominator of the measure. We will notify all SNFs of their incentive multipliers for the Program year, including SNFs with zero eligible stays during the baseline and performance periods. These SNFs will receive an incentive multiplier that results in the adjusted Federal per diem rate under the Medicare SNF PPS that they would otherwise have received absent the Program.

Comment: Commenter suggested as an alternative to our low-volume adjustment proposal that we consider adopting a 2-year performance period for low-volume SNFs only and weight the most recent year more highly.

Response: We thank the commenter for this feedback and will consider this in future rulemaking.

Comment: One commenter suggested that we consider assigning a 2 percent payment penalty to low-volume SNFs instead of adopting the low-volume scoring adjustment as proposed. The commenter suggested that this policy would encourage low-volume SNFs to increase their Medicare cases sizes, which would enable Medicare to adequately measure their care quality and hold all SNFs accountable for their care.

Response: We do not believe the intent of the SNF VBP was to incentivize SNFs to increase their Medicare case volume. We wish to avoid increasing possible healthcare disparities for smaller facilities when payment differences are driven solely by smaller measure denominators, and not quality of care as reflected in measure performance. Finally, we are concerned about the possibility of gaming this kind of policy, as SNFs might seek out Medicare cases to avoid the 2 percent penalty the commenter suggests.

Comment: One commenter expressed support for our proposed low-volume adjustment and opposition to the alternative that we presented, stating that performance scores under the Program can be skewed by a single readmission and that the alternative would reduce Medicare rates for low-volume SNFs regardless of their performance and with no opportunity to earn additional incentive payments. Another commenter supported our proposal to adopt a low-volume scoring adjustment, noting that the evidence shows that the SNFRM is not a reliable quality indicator when facilities have fewer than 25 qualifying admissions. The commenter also agreed with our proposal to adjust the Program's payback percentage to account for this policy.

Response: We thank the commenters for the support.

Comment: One commenter stated that low-volume SNFs should be excluded from the SNF VBP Program since they have no realistic opportunity to earn value-based incentive payments.

Response: We believe that the low-volume scoring adjustment policy ensures that these SNFs are adequately protected from being scored on insufficiently-reliable SNFRM rates.

Comment: One commenter appreciated our efforts to address low-volume SNFs and SNFs without baseline period data. However, the commenter was concerned that CMS had not provided enough information on these topics and requested additional clarity.

Response: We believe we have provided as much clarity as possible on the effects of the low-volume scoring adjustment policy in both the preamble of the proposed rule and the Regulatory Impact Analysis that was included in the proposed rule. We have also provided additional clarity in this final rule and in the Regulatory Impact Analysis that is included in this final rule. We will also ensure that affected SNFs are made fully aware when their SNF performance scores were assigned as a result of the policy and notify them of their value-based incentive payment percentage for the fiscal year, as required by section 1888(h)(7) of the Act. We believe that notification will ensure that SNFs are aware of the effects that this policy has on their SNF performance scores and incentive payments.

After consideration of the public comments that we have received, we are finalizing our scoring adjustment for low-volume SNFs as proposed. We are also finalizing our regulation text on this policy as proposed.

d. Extraordinary Circumstances Exception Policy for the SNF VBP Program

In the FY 2018 SNF PPS final rule (82 FR 36616), we summarized public comments that we received on the topic of a possible extraordinary circumstances exception policy for the SNF VBP Program. As we stated in that rule, in other value-based purchasing and quality reporting programs, we have adopted Extraordinary Circumstances Exceptions (ECE) policies intended to allow facilities to receive relief from program requirements due to natural disasters or other circumstances beyond the facility's control that may affect the facility's ability to provide high-quality health care.

In other programs, we have defined a “disaster” as any natural or man-made catastrophe which causes damages of sufficient severity and magnitude to partially or completely destroy or delay access to medical records and associated documentation or otherwise affect the facility's ability to continue normal operations. Natural disasters could include events such as hurricanes, tornadoes, earthquakes, volcanic eruptions, fires, mudslides, snowstorms, and tsunamis. Man-made disasters could include such events as terrorist attacks, bombings, flood caused by man-made actions, civil disorders, and explosions. A disaster may be widespread and impact multiple structures or be isolated and affect a single site only. As a result of either a natural or man-made disaster, we are concerned that SNFs' care quality and subsequent impact on measure performance in the SNF VBP Program may suffer, and as a result, SNFs might be penalized under the Program's quality measurement and scoring methodology. However, we do not wish to penalize SNFs in these circumstances. For example, we recognize that SNFs might receive patients involuntarily discharged from hospitals facing mandatory evacuation due to probable flooding, and these patients might be readmitted to inpatient acute care hospitals and result in poorer readmission measure performance in the SNF VBP Program. We therefore proposed to adopt an ECE policy for the SNF VBP Program to provide relief to SNFs affected by natural disasters or other circumstances beyond the facility's control that affect the care provided to the facility's patients. We proposed that if a SNF can demonstrate that an extraordinary circumstance affected the care that it provided to its patients and subsequent measure performance, we would exclude from the calculation of the measure rate for the applicable baseline and performance periods the calendar months during which the SNF was affected by the extraordinary circumstance. Under this proposal, a SNF requesting an ECE would indicate the dates and duration of the extraordinary circumstance in its request, along with any available evidence of the extraordinary circumstance, and if approved, we would exclude the corresponding calendar months from that SNF's measure rate for the applicable measurement period and by extension, its SNF performance score.

We further proposed that SNFs must submit this ECE request to CMS by filling out the ECE request form that we will place on the QualityNet website to the [e-mail] mailbox within 90 days following the extraordinary circumstance.

To accompany an ECE request, SNFs must provide any available evidence showing the effects of the extraordinary circumstance on the care they provided to their patients, including, but not limited to, photographs, newspaper and other media articles, and any other materials that would aid CMS in making its decision. We stated that we will review exception requests, and at our discretion based on our evaluation of the impact of the extraordinary circumstances on the SNF's care, provide a response to the SNF as quickly as feasible.

We stated our intent for this policy to offer relief to SNFs whose care provided to patients suffered as a result of the disaster or other extraordinary circumstance, and we believe that excluding calendar months affected by extraordinary circumstances from SNFs' measure performance under the Program appropriately ensures that such circumstances do not unduly affect SNFs' performance rates or performance scores. We developed this process to align with the ECE process adopted by the SNF Quality Reporting Program to the greatest extent possible and to minimize burden on SNFs. This policy is not intended to preclude us from granting exceptions to SNFs that have not requested them when we determine that an extraordinary circumstance, such as an act of nature, affects an entire region or locale. If we made the determination to grant an exception to all SNFs in a region or locale, we proposed to communicate this decision through routine communication channels to SNFs and vendors, including but not limited to, issuing memos, emails, and notices on our SNF VBP website at https://www.cms.gov/​Medicare/​Quality-Initiatives-Patient-Assessment-Instruments/​Value-Based-Programs/​Other-VBPs/​SNF-VBP.html.

We noted that if we finalize this policy, we would score any SNFs receiving ECEs on achievement and improvement for any remaining months during the performance period, provided the SNF had at least 25 eligible stays during both of those periods. If a SNF should receive an approved ECE for 6 months of the performance period, for example, we would score the SNF on its achievement during the remaining 6 months on the Program's measure as long as the SNF met the proposed 25 eligible stay threshold during the performance period. We would also score the SNF on improvement as long as it met the proposed 25 eligible stay threshold during the applicable baseline period.

We welcomed public comments on this proposal. We also proposed to codify this proposal at § 413.338(d)(4).

Comment: Two commenters expressed appreciation and support for our proposal to adopt an ECE policy for the SNF VBP Program. The commenters acknowledged that these exceptions are provided in other programs and agreed that we should align our ECE policy with the Hospital VBP Program as much as possible. A third commenter reiterated its previous support for an ECE policy in the SNF VBP Program.

Response: We thank the commenters for the support.

After consideration of the public comments that we received, we are finalizing our Extraordinary Circumstances Exception policy as proposed. We are also finalizing our regulation text on this policy as proposed.

6. SNF Value-Based Incentive Payments

We refer readers to the FY 2018 SNF PPS final rule (82 FR 36616 through 36621) for discussion of the exchange function methodology that we have adopted for the Program, as well as the specific form of the exchange function (logistic, or S-shaped curve) that we finalized, and the payback percentage of 60 percent. We adopted these policies for FY 2019 and subsequent fiscal years.

As required by section 1888(h)(7) of the Act, we will inform each SNF of the adjustments to its Medicare payments as a result of the SNF VBP Program that we will make not later than 60 days prior to the fiscal year involved. We will fulfill that requirement via SNF Performance Score Reports that we will circulate to SNFs using the QIES-CASPER system, which is also how we distribute the quarterly confidential feedback reports that we are required to provide to SNFs under section 1888(g)(5) of the Act. The SNF Performance Score Reports will contain the SNF's performance score, ranking, and value-based incentive payment adjustment factor that will be applied to claims submitted for the applicable fiscal year. Additionally, as we finalized in the FY 2018 SNF PPS final rule (82 FR 36622 through 36623), the provision of the SNF Performance Score Report will trigger the Phase Two Review and Corrections Process, and SNFs will have 30 days from the date we post the report on the QIES-CASPER system to submit corrections to their SNF performance score and ranking to the [e-mail] mailbox.

Finally, as we discussed in the FY 2018 SNF PPS final rule (82 FR 36618), beginning with FY 2019 (October 1, 2018) payments, we intend to make the 2 percent reduction and the SNF-specific value-based incentive payment adjustment to SNF claims simultaneously. Beginning with FY 2019, we will identify the adjusted federal per diem rate for each SNF for claims under the SNF PPS. We will then reduce that amount by 2 percent by multiplying the per diem amount by 0.98, in accordance with the requirements in section 1888(h)(6) of the Act. We will then multiply the result of that calculation by each SNF's specific value-based incentive payment adjustment factor, which will be based on each SNF's performance score for the program year and will be calculated by the exchange function, to generate the value-based incentive payment amount that applies to the SNF for the fiscal year. Finally, we will add the value-based incentive payment amount to the reduced rate, resulting in a new adjusted federal per diem rate that applies to the SNF for the fiscal year.

At the time of the publication of the proposed rule, we had not completed SNF performance score calculations for the FY 2019 program year. However, we stated our intent to provide the range of value-based incentive payment adjustment factors applicable to the FY 2019 program year in this final rule. For the FY 2019 SNF VBP Program Year, and incorporating the 2 percent reduction to SNFs' payments, we estimate the value-based incentive payment adjustment factors that we will award to SNFs range from 0.9802915381 to 1.02326809. That is, we estimate that SNFs may receive incentive payment percentages ranging from approximately −1.97 percent to approximately +2.33 percent, on a net basis.

We proposed to codify the SNF VBP Program's payment adjustments at § 413.337(f).

Comment: Two commenters urged us to revisit the payback percentage policy that we adopted in the FY 2018 SNF PPS final rule (82 FR 36619 through 36621), stating that we should distribute 70 percent of the funds withheld from SNFs' Medicare payments through the SNF VBP Program, the maximum amount allowable under the statute. One commenter requested that we return the remaining 30 percent of funds for SNF quality improvement initiatives, including programs to improve SNFs' performance when they have high readmission rates, while the other commenter stated that we should remit 100 percent of the Program's funds as is done in the Hospital Value-Based Purchasing Program.

Response: As we discussed in the FY 2018 SNF PPS final rule (82 FR 36621), we are not authorized to distribute the 30 percent of SNFs' Medicare payments that would remain after the payment withhold is determined for any purposes. Those funds are retained in the Medicare Trust Fund and used for other Medicare Program purposes authorized by statute. We are not allowed under current law to distribute 100 percent of the withheld funds for SNF VBP purposes.

Further, we do not believe it is appropriate to revisit the payback percentage policy at this time, with the exception of the low-volume policy, which we view as a narrow exception to the 60 percent payback percentage that would have no effect on the majority of facilities. At the time of the publication of this final rule, the SNF VBP Program will not yet have delivered its first incentive payments based on measured performance, and we do not believe we should consider whether to change the payback percentage further until we are able to more fully assess the effects that it has on the quality of care provided in SNFs. We refer readers to the FY 2018 SNF PPS final rule (82 FR 36619 through 36621) for our full discussion of the payback percentage policy that we have adopted for the SNF VBP Program.

We thank the commenters for their feedback. As noted in section III.B.5. of this final rule, we are finalizing the codification of the SNF VBP program payment adjustment as proposed.

D. Request for Information on Promoting Interoperability and Electronic Healthcare Information Exchange Through Possible Revisions to the CMS Patient Health and Safety Requirements for Skilled Nursing Facility Providers and Suppliers

In the FY 2019 SNF PPS proposed rule, we included a Request for Information (RFI) related to promoting interoperability and electronic healthcare information exchange (83 FR 21089). We received 22 comments on this RFI, and appreciate the input provided by commenters.

VII. Collection of Information Requirements

Under the Paperwork Reduction Act of 1995 (PRA) (44 U.S.C. 3501 et seq.), we are required to publish a 60-day notice in the Federal Register and solicit public comment before a collection of information requirement is submitted to the Office of Management and Budget (OMB) for review and approval.

To fairly evaluate whether an information collection should be approved by OMB, PRA section 3506(c)(2)(A) requires that we solicit comment on the following issues:

• The need for the information collection and its usefulness in carrying out the proper functions of our agency.
• The accuracy of our burden estimates.
• The quality, utility, and clarity of the information to be collected.
• Our effort to minimize the information collection burden on the affected public, including the use of automated collection techniques.

In our May 8, 2018 proposed rule (83 FR 21018), we solicited public comment on each of the section 3506(c)(2)(A)-required issues for the following information collection requirements (ICRs).We did not receive any comments on the ICR section of the proposed rule.

A. Wages

To derive average costs, we used data from the U.S. Bureau of Labor Statistics' May 2017 National Occupational Employment and Wage Estimates for all salary estimates (as compared to the FY 2019 SNF PPS proposed rule when we used May 2016 estimates) (http://www.bls.gov/​oes/​current/​oes_​nat.htm). In this regard, Table 42 presents the mean hourly wage, the cost of fringe benefits and overhead (calculated at 100 percent of salary), and the adjusted hourly wage. We are using the adjusted wages to derive our cost estimates.

Table 42—National Occupational Employment and Wage Estimates

Occupation title	Occupation code	Mean hourly wage ($/hr)	Fringe benefits and overhead ($/hr)	Adjusted hourly wage ($/hr)
Health Information Technician	29-2071	20.59	20.59	41.18
Registered Nurse	29-1141	35.36	35.36	70.72

As indicated, we are adjusting our employee hourly wage estimates by a factor of 100 percent. This is necessarily a rough adjustment, both because fringe benefits and overhead costs vary significantly from employer to employer, and because methods of estimating these costs vary widely from study to study. We believe that doubling the hourly wage to estimate total cost is a reasonably accurate estimation method.

B. Information Collection Requirements (ICRs)

1. ICRs Regarding the SNF PPS Assessment Schedule Under the PDPM

The following sets out the requirements and burden associated with the MDS assessment schedule that will be effective October 1, 2019 under the SNF PPS in conjunction with implementation of the PDPM. The requirements and burden will be submitted to OMB for approval under control number 0938-1140 (CMS-10387).

Section V.D. of this final rule finalizes revisions to the current SNF PPS assessment schedule to require only two scheduled assessments (as opposed to the current requirement for five scheduled assessments) for each SNF stay: A 5-day scheduled PPS assessment and a discharge assessment.

The current 5-day scheduled PPS assessment will be used as the admission assessment under this rule's finalized PDPM and set the resident's case-mix classification for the resident's SNF stay. The PPS discharge assessment (which is already required for all SNF Part A residents) will serve as the discharge assessment and be used for monitoring purposes. In section V.D. of this final rule, we discuss that while we proposed to require SNFs to reclassify residents under the PDPM using the Interim Payment Assessment (IPA) if certain criteria are met, we have decided in this final rule to make this assessment optional, thereby leaving completion of this assessment at the discretion of the individual provider. Thus, the 5-day SNF PPS scheduled assessment will be the only PPS assessment required to classify a resident under the PDPM for payment purposes, while the IPA may also be completed, as discussed in section V.D. of this final rule. This eliminates the requirement for the following assessments under the SNF PPS: 14-day scheduled PPS assessment, 30-day scheduled PPS assessment, 60-day scheduled PPS assessment, 90-day scheduled PPS assessment, Start of Therapy Other Medicare Required Assessment (OMRA), End of Therapy OMRA, and Change of Therapy OMRA.

In estimating the amount of time to complete a PPS assessment, we utilize the OMRA assessment, or the NO/SO item set (this is consistent with the current information collection request as approved by OMB on July 28, 2017; see https://www.reginfo.gov/​public/​do/​PRAViewDocument?​ref_​nbr=​201703-0938-018) as a proxy for all assessments. In section V.D. of this final rule, we finalized the addition of 18 items to the PPS discharge assessment in order to calculate and monitor the total amount of therapy provided during a SNF stay. These items are listed in Table 35 under section V.D. of this final rule. Given that the PPS OMRA assessment has 272 items (as compared to 125 items currently on the PPS discharge assessment) we believe the items that we are adding to the PPS discharge assessment, while increasing burden for each of the respective assessments, is accounted for by using the longer PPS OMRA assessment as a proxy for the time required to complete all assessments.

When calculating the burden for each assessment, we estimated that it will take 40 minutes (0.6667 hours) at $70.72/hr for an RN to collect the information necessary for preparing the assessment, 10 minutes (0.1667 hours) at $55.95/hr (the average hourly wage for RN ($70.72/hr) and health information technician ($41.18/hr)) for staff to code the responses, and 1 minute (0.0167 hours) at $41.18/hr for a health information technician to transmit the results. In total, we estimate that it would take 51 minutes (0.85 hours) to complete a single PPS assessment. Based on the adjusted hourly wages for the noted staff, we estimate that it would cost $57.17 [($70.72/hr × 0.6667 hr) + ($55.95/hr × 0.1667 hr) + ($41.18/hr × 0.0167 hr)] to prepare, code, and transmit each PPS assessment.

The ongoing burden associated with the revisions to the SNF PPS assessment schedule is the time and effort it would take each Medicare Part A SNF to complete the 5-day PPS and discharge assessments. Based on our most current data, there are 15,471 Medicare Part A SNFs (as opposed to the 15,455 discussed in the proposed rule). Based on FY 2017 data, we estimate that 2,406,401 5-day PPS assessments will be completed and submitted by Part A SNFs each year under the PDPM. We used the same number of assessments (2,406,401) as a proxy for the number of PPS discharge assessments that would be completed and submitted each year, since all residents who require a 5-day PPS assessment will also require a discharge assessment under the SNF PDPM.

As compared to the FY 2019 SNF PPS proposed rule, in which we used the Significant Change in Status Assessment (SCSA) as a proxy to estimate the number of IPAs (83 FR 21093), we have eliminated this portion of our burden estimate as this assessment would not be required, per the discussion in section V.D. of this final rule. Therefore, we estimate that the total number of 5-day scheduled PPS assessments, and PPS discharge assessments that would be completed across all facilities is 4,812,802 assessments (2,406,401 + 2,406,401, respectively) instead of 4,905,042 assessments (2,406,401 + 92,240 + 2,406,401) that was set out in the proposed rule. For all assessments under the PDPM, we estimated a burden of 4,090,882 hours (4,812,802 assessments × 0.85 hr/assessment) at a cost of $275,147,890 (4,812,802 assessments × $57.17/assessment).

Based on the same FY 2017 data, there were 5,833,476 non-discharge related assessments (scheduled and unscheduled PPS assessments) completed under the RUG- IV payment system. To this number we add the same proxy as above for the number of discharge assessments (2,406,401), since every resident under RUG-IV who required a 5-day scheduled PPS assessment would also require a discharge assessment. This brings the total number of estimated assessments under RUG-IV to 8,239,877. Using the same wage and time figures (per assessment), we estimated a burden of 7,003,895 hours (8,239,877 assessments × 0.85 hr/assessment) at a cost of $471,073,768 (8,239,877 assessments × $57.17/assessment).

When comparing the currently approved RUG-IV burden with the PDPM burden, we estimate a savings of 2,913,013 administrative hours (7,003,895 RUG-IV hours − 4,090,882 PDPM hours) or approximately 188 hours per provider per year (2,913,013 hours/15,471 providers). As depicted in Table 43, we also estimate a cost savings of $195,925,878 ($471,073,768 RUG-IV costs − $275,147,890 PDPM costs) or $12,664 per provider per year ($195,925,878/15,471 providers). This represents a significant decrease in administrative burden to providers under PDPM.

Table 43—PDPM Savings

Burden reconciliation	Respondents *	Responses (assessments)	Burden per response (hours)	Total annual burden (hours)	Cost ($)
RUG-IV	15,455	8,239,877	0.85	7,003,895	471,073,768
PDPM	15,471	4,812,802	0.85	4,090,882	275,147,890
SAVINGS	(16)	(3,427,075)	No change	(2,913,013)	(195,925,878)
* The RUG-IV number of respondents is based on the last approved PRA package in 2017. Numbers of respondents changes from year to year.

Finally, in section V.D. of this final rule, we finalized the addition of 3 items, as listed in Table 34 of this final rule), to the MDS 3.0 for Nursing Homes and Swing Bed Providers. Based on the small number of items being added and the small percentage of assessments that Swing Bed providers make up, we do not believe this action will cause any measurable adjustments to our currently approved burden estimates. Consequently, we are not revising any of those estimates.

2. ICRs Regarding the SNF VBP Program

In section VI.C.5.d. of this final rule, we are adopting an Extraordinary Circumstances Exception (ECE) process for the SNF VBP. Because the same CMS Extraordinary Circumstances Exceptions (ECE) Request Form would be used across ten quality programs: Hospital IQR Program, Hospital Outpatient Reporting Program, Inpatient Psychiatric Facility Quality Reporting Program, PPS-Exempt Cancer Hospital Quality Reporting Program, Ambulatory Surgical Center Quality Reporting Program, Hospital VBP Program, Hospital-Acquired Condition Reduction Program, Hospital Readmissions Reduction Program, End Stage Renal Disease Quality Incentive Program, and Skilled Nursing Facility Value-Based Purchasing Program—the form and its associated requirements/burden will be submitted to OMB for approval under one information collection request (CMS-10210, OMB control number: 0938-1022) and in association with our IPPS final rule (CMS-1694-F; RIN 0938-AT27). To avoid double counting we are not setting out the form's SNF-related burden in this final rule. Separately, we are not removing or adding any new or revised SNF VBP measure-related requirements or burden in this rule. Consequently, this final rule does not set out any new VBP-related collections of information that would be subject to OMB approval under the authority of the PRA.

3. ICRs for the SNF Quality Reporting Program (QRP)

We are not removing or adding any new or revised SNF QRP measure-related requirements or burden in this rule. Consequently, this final rule does not set out any new QRP-related collections of information that would be subject to OMB approval under the authority of the PRA.

C. Summary of Requirements and Annual Burden Estimates

Table 44—Information Collection Requirements and Burden Estimates

Requirement	OMB control No.	Respondents	Responses (per respondent)	Total responses	Time per response (hr)	Total time (hr)	Labor cost per hour ($/hr)	Annualized cost ($)
SNF PPS Assessment Schedule	0938-1140	15,471	(311)	(4,812,802)	0.85	(4,090,882)	Varies	(275,147,890)

VIII. Economic Analyses

A. Regulatory Impact Analysis

1. Statement of Need

This final rule will update the FY 2018 SNF prospective payment rates as required under section 1888(e)(4)(E) of the Act. It also responds to section 1888(e)(4)(H) of the Act, which requires the Secretary to provide for publication in the Federal Register before the August 1 that precedes the start of each FY, the unadjusted federal per diem rates, the case-mix classification system, and the factors to be applied in making the area wage adjustment. As these statutory provisions prescribe a detailed methodology for calculating and disseminating payment rates under the SNF PPS, we do not have the discretion to adopt an alternative approach on these issues. We did not include the impacts of the proposed PDPM and related policies in the sections that follow, as we have included this discussion in section V.I. of this final rule.

2. Introduction

We have examined the impacts of this final rule as required by Executive Order 12866 on Regulatory Planning and Review (September 30, 1993), Executive Order 13563 on Improving Regulation and Regulatory Review (January 18, 2011), the Regulatory Flexibility Act (RFA, September 19, 1980, Pub. L. 96-354), section 1102(b) of the Act, section 202 of the Unfunded Mandates Reform Act of 1995 (UMRA, March 22, 1995; Pub. L. 104-4), Executive Order 13132 on Federalism (August 4, 1999), the Congressional Review Act (5 U.S.C. 804(2)), and Executive Order 13771 on Reducing Regulation and Controlling Regulatory Costs (January 30, 2017).

Executive Orders 12866 and 13563 direct agencies to assess all costs and benefits of available regulatory alternatives and, if regulation is necessary, to select regulatory approaches that maximize net benefits (including potential economic, environmental, public health and safety effects, distributive impacts, and equity). Executive Order 13563 emphasizes the importance of quantifying both costs and benefits, of reducing costs, of harmonizing rules, and of promoting flexibility. This rule has been designated an economically significant rule, under section 3(f)(1) of Executive Order 12866. Accordingly, we have prepared a regulatory impact analysis (RIA) as further discussed below. Also, the rule has been reviewed by OMB.

Executive Order 13771, titled Reducing Regulation and Controlling Regulatory Costs, was issued on January 30, 2017. OMB's implementation guidance, issued on April 5, 2017, explains that “Federal spending regulatory actions that cause only income transfers between taxpayers and program beneficiaries (for example, regulations associated with . . . Medicare spending) are considered `transfer rules' and are not covered by E.O. 13771. . . . However . . . such regulatory actions may impose requirements apart from transfers . . . In those cases, the actions would need to be offset to the extent they impose more than de minimis costs. Examples of ancillary requirements that may require offsets include new reporting or recordkeeping requirements.” As discussed in section VII. of this final rule, we estimate that this final rule will lead to paperwork cost savings of approximately $196 million per year, or $171 million per year on an ongoing basis discounted at 7 percent relative to year 2016, over a perpetual time horizon. This final rule is considered an E.O. 13771 deregulatory action.

3. Overall Impacts

This final rule sets forth updates of the SNF PPS rates contained in the SNF PPS final rule for FY 2018 (82 FR 36530). We estimate that the aggregate impact will be an increase of approximately $820 million in payments to SNFs in FY 2019, resulting from the SNF market basket update to the payment rates, as required by section 53111 of the BBA 2018. Absent the application of section 53111 of the BBA 2018, as discussed in section III.A.2. of this final rule, the aggregate impact from the 2.0 percentage point market basket increase factor would have been approximately $680 million. We note that these impact numbers do not incorporate the SNF VBP reductions that we estimate will total $527.4 million for FY 2019.

We would note that events may occur to limit the scope or accuracy of our impact analysis, as this analysis is future-oriented, and thus, very susceptible to forecasting errors due to events that may occur within the assessed impact time period. In accordance with sections 1888(e)(4)(E) and 1888(e)(5) of the Act, we update the FY 2018 payment rates by a factor equal to the market basket index percentage change adjusted by the MFP adjustment to determine the payment rates for FY 2019. As discussed previously, section 53111 of the BBA 2018 stipulates a market basket increase factor of 2.4 percent. The impact to Medicare is included in the total column of Table 45. In updating the SNF PPS rates for FY 2019, we made a number of standard annual revisions and clarifications mentioned elsewhere in this final rule (for example, the update to the wage and market basket indexes used for adjusting the federal rates).

The annual update set forth in this final rule applies to SNF PPS payments in FY 2019. Accordingly, the analysis of the impact of the annual update that follows only describes the impact of this single year. Furthermore, in accordance with the requirements of the Act, we will publish a rule or notice for each subsequent FY that will provide for an update to the payment rates and include an associated impact analysis.

4. Detailed Economic Analysis

The FY 2019 SNF PPS payment impacts appear in Table 45. Using the most recently available data, in this case FY 2017, we apply the current FY 2018 wage index and labor-related share value to the number of payment days to simulate FY 2018 payments. Then, using the same FY 2017 data, we apply the proposed FY 2019 wage index and labor-related share value to simulate FY 2019 payments. We tabulate the resulting payments according to the classifications in Table 45 (for example, facility type, geographic region, facility ownership), and compare the simulated FY 2018 payments to the simulated FY 2019 payments to determine the overall impact. The breakdown of the various categories of data Table 45 follows:

• The first column shows the breakdown of all SNFs by urban or rural status, hospital-based or freestanding status, census region, and ownership.
• The first row of figures describes the estimated effects of the various changes on all facilities. The next six rows show the effects on facilities split by hospital-based, freestanding, urban, and rural categories. The next nineteen rows show the effects on facilities by urban versus rural status by census region. The last three rows show the effects on facilities by ownership (that is, government, profit, and non-profit status).
• The second column shows the number of facilities in the impact database.
• The third column shows the effect of the annual update to the wage index. This represents the effect of using the most recent wage data available. The total impact of this change is 0 percent; however, there are distributional effects of the change.
• The fourth column shows the effect of all of the changes on the FY 2019 payments. The update of 2.4 percent is constant for all providers and, though not shown individually, is included in the total column. It is projected that aggregate payments will increase by 2.4 percent, assuming facilities do not change their care delivery and billing practices in response.

As illustrated in Table 45, the combined effects of all of the changes vary by specific types of providers and by location. For example, due to changes in this rule, providers in the urban Pacific region will experience a 3.4 percent increase in FY 2019 total payments.

Table 45—Impact to the SNF PPS for FY 2019

	Number of facilities FY 2019 (%)	Update wage data (%)	Total change (%)
Group:
Total	15,471	0.0	2.4
Urban	11,042	0.0	2.4
Rural	4,429	0.1	2.5
Hospital-based urban	498	0.0	2.4
Freestanding urban	10,544	0.0	2.4
Hospital-based rural	555	0.0	2.4
Freestanding rural	3,874	0.2	2.6
Urban by region:
New England	790	−0.7	1.7
Middle Atlantic	1,481	0.0	2.4
South Atlantic	1,869	−0.1	2.3
East North Central	2,127	−0.4	2.0
East South Central	555	−0.2	2.2
West North Central	920	−0.4	2.0
West South Central	1,346	0.3	2.7
Mountain	527	−0.8	1.6
Pacific	1,421	1.0	3.4
Outlying	6	−0.5	1.9
Rural by region:
New England	134	−0.7	1.6
Middle Atlantic	215	0.1	2.5
South Atlantic	494	0.1	2.5
East North Central	931	0.1	2.5
East South Central	523	−0.3	2.1
West North Central	1,074	0.3	2.7
West South Central	734	1.0	3.5
Mountain	229	0.2	2.6
Pacific	95	−0.5	1.9
Ownership:
Government	10,887	0.0	2.4
Profit	3,570	−0.1	2.3
Non-Profit	1,014	0.0	2.4
Note: The Total column includes the 2.4 percent market basket increase required by section 53111 of the BBA 2018. Additionally, we found no SNFs in rural outlying areas.

5. Impacts for the SNF QRP

We did not propose to add, remove, or revise any measures in the SNF QRP. Consequently, this final rule does not set out any new QRP-related impacts associated with the SNF QRP.

6. Impacts for the SNF VBP Program

In Table 44 of the FY 2019 SNF PPS proposed rule (83 FR 21096 through 20197), we estimated the impacts of the FY 2019 SNF VBP Program without taking into account our low-volume scoring adjustment proposal. We modeled SNFs' performance in the Program using SNFRM data from CY 2014 as the baseline period and FY 2016 as the performance period. Additionally, we modeled a logistic exchange function with a payback percentage of 60 percent, as we finalized in the FY 2018 SNF PPS final rule (82 FR 36619 through 36621), and based the following analyses on payments to SNFs in FY 2016. We estimated the total reductions to payments required by section 1888(h)(6) of the Act, to be $527.4 million for FY 2019. Based on the 60 percent payback percentage, we estimated that we would disburse approximately $316.4 million in value-based incentive payments to SNFs in FY 2019, which we estimated would result in approximately $211 million in savings to the Medicare program in FY 2019.

In Table 45 of the FY 2019 SNF PPS proposed rule (83 FR 21097), we also modeled the estimated impacts of the FY 2019 SNF VBP Program and included in that model the impacts of our proposed scoring adjustment for low-volume SNFs. We estimated that the scoring adjustment policy proposal would redistribute an additional $6.7 million to the group of low volume SNFs. As we discuss further in section II.E.3.e. of this final rule, we are finalizing our low-volume scoring adjustment policy, and our estimated FY 2019 SNF VBP impacts, which we described in Table 45 of the proposed rule, are reproduced as Table 46 below.

We continue to estimate that this policy will result in increasing low-volume SNFs' value-based incentive payment percentages by approximately 0.99 percent, on average, from the value-based incentive payment percentage that they would receive in the absence of the low-volume adjustment. An increase in value-based incentive payment percentages by 0.99 percent is needed to bring low-volume SNFs back to the 2.0 percent that was withheld from their payments. We also continue to estimate that we will pay an additional $6.7 million in incentive payments to low-volume SNFs, which would increase the 60 percent payback percentage for FY 2019 by approximately 1.28 percent, making the new payback percentage for FY 2019 equal to 61.28 percent of the estimated $527.4 million in withheld funds for that fiscal year.

Our detailed analysis of the impacts of the FY 2019 SNF VBP Program, including the finalized low-volume scoring adjustment policy, follows in Table 46.

Table 46—Estimated SNF VBP Program Impacts Including Effects of the Finalized Low-Volume Scoring Adjustment

Category	Criterion	Number of facilities	RSRR (mean)	Mean SNF performance score	Mean incentive multiplier (60% payback)	Percent of proposed payback
Group	Total	12,845	0.18912	41.371	1.192	99.9 *
	Urban	9,604	0.18957	40.956	1.177	84.4
	Rural	3,241	0.18779	41.011	1.181	15.4
Urban by Region	Total	9,604
	01=Boston	713	0.19089	37.26777	1.059	4.9
	02=New York	836	0.19029	40.90383	1.165	11.8
	03=Philadelphia	1,040	0.18601	45.31896	1.325	10.1
	04=Atlanta	1,767	0.19332	37.28735	1.052	13.3
	05=Chicago	1,961	0.18784	43.06368	1.246	16.0
	06=Dallas	1,134	0.19416	34.53275	0.949	6.1
	07=Kansas City	510	0.19057	39.26278	1.132	2.6
	08=Denver	241	0.17832	57.62596	1.790	2.9
	09=San Francisco	1,098	0.18908	40.80722	1.176	12.5
	10=Seattle	304	0.17808	56.67839	1.713	4.2
Rural by Region	Total	3,241
	01=Boston	115	0.18133	51.89294	1.568	0.9
	02=New York	77	0.18366	50.48193	1.569	0.5
	03=Philadelphia	240	0.18789	42.12621	1.218	1.3
	04=Atlanta	764	0.19283	36.51452	1.032	3.3
	05=Chicago	818	0.18397	47.85089	1.399	4.5
	06=Dallas	557	0.19355	34.00868	0.952	1.7
	07=Kansas City	421	0.18634	42.64769	1.236	1.2
	08=Denver	132	0.18000	52.38900	1.544	0.7
	09=San Francisco	48	0.17780	61.50419	1.931	0.6
	10=Seattle	69	0.17628	60.70084	1.836	0.7
Ownership Type	Total	12,847
	Government	688	0.18529	46.450	1.380	5.2
	Profit	9,250	0.19039	39.526	1.127	72.0
	Non-Profit	2,909	0.18597	46.038	1.353	22.9
Number of Beds	Total	12,847
	1st Quartile:	3,222	0.18760	42.466	1.226	24.6
	2nd Quartile:	3,221	0.18878	40.971	1.175	24.4
	3rd Quartile:	3,197	0.19048	40.242	1.153	23.3
	4th Quartile:	3,207	0.18963	41.800	1.212	27.7
* This category does not add to 100% because a small number of SNFs did not have urban/rural designations in our data.

7. Alternatives Considered

As described in this section, we estimated that the aggregate impact for FY 2019 under the SNF PPS will be an increase of approximately $820 million in payments to SNFs, resulting from the SNF market basket update to the payment rates, as required by section 53111 of the BBA 2018. Absent application of section 53111 of the BBA 2018, as discussed in section III.A.2. of this final rule, the market basket increase factor of 2.0 percent would have resulted in an aggregate increase in payments to SNFs of approximately $680 million.

Section 1888(e) of the Act establishes the SNF PPS for the payment of Medicare SNF services for cost reporting periods beginning on or after July 1, 1998. This section of the statute prescribes a detailed formula for calculating base payment rates under the SNF PPS, and does not provide for the use of any alternative methodology. It specifies that the base year cost data to be used for computing the SNF PPS payment rates must be from FY 1995 (October 1, 1994, through September 30, 1995). In accordance with the statute, we also incorporated a number of elements into the SNF PPS (for example, case-mix classification methodology, a market basket index, a wage index, and the urban and rural distinction used in the development or adjustment of the federal rates). Further, section 1888(e)(4)(H) of the Act specifically requires us to disseminate the payment rates for each new FY through the Federal Register, and to do so before the August 1 that precedes the start of the new FY; accordingly, we are not pursuing alternatives for this process.

As discussed in section VI.C. of this final rule, we also considered an alternative SNF VBP low-volume scoring policy. This alternative scoring assignment would result in a value-based incentive payment percentage of 1.2 percent, or 60 percent of the 2 percent withhold. This amount would match low-volume SNFs' incentive payment percentages with the finalized SNF VBP Program payback percentage of 60 percent, and would represent a smaller adjustment to low-volume SNFs' incentive payment percentages than the proposed policy described above. We estimated that this alternative would redistribute an additional $1 million with respect to FY 2019 services to low-volume SNFs. We also estimated that this alternative would increase the 60 percent payback percentage for FY 2019 by approximately 0.18 percent of the approximately $527.4 million of the total withheld from SNFs' payments, which would result in a payback percentage of 60.18 percent of the estimated $527.4 million in withheld funds for that Program year. We estimated that this alternative would pay back SNFs about $5.7 million less than the proposed low-volume scoring methodology adjustment in total estimated payments on an annual basis. However, under this alternative, like the policy we are finalizing, the specific amount by which the payback percentage would increase for each Program year would vary based on the number of low-volume SNFs that we identify for each Program year and the distribution of all SNFs' performance scores for that Program year.

We discussed the comments that we received on this alternative and our responses to those comments in section II.E.3.e. of this final rule in our discussion of the low-volume scoring adjustment policy.

8. Accounting Statement

As required by OMB Circular A-4 (available online at https://obamawhitehouse.archives.gov/​omb/​circulars_​a004_​a-4/​), in Tables 48 and 49, we have prepared an accounting statement showing the classification of the expenditures associated with the provisions of this final rule for FY 2019. Tables 45 and 48 provides our best estimate of the possible changes in Medicare payments under the SNF PPS as a result of the policies in this final rule, based on the data for 15,471 SNFs in our database. Tables 46 and 49 provide our best estimate of the possible changes in Medicare payments under the SNF VBP as a result of the policies in this final rule.

Table 47—Accounting Statement: Classification of Estimated Expenditures, From the 2018 SNF PPS Fiscal Year to the 2019 SNF PPS Fiscal Year

Category	Transfers
Annualized Monetized Transfers	$820 million.*
From Whom To Whom?	Federal Government to SNF Medicare Providers.
* The net increase of $820 million in transfer payments is a result of the market basket increase of $820 million.

Table 48—Accounting Statement: Classification of Estimated Expenditures for the FY 2019 SNF VBP Program

Category	Transfers
Annualized Monetized Transfers	$316.4 million.*
From Whom To Whom?	Federal Government to SNF Medicare Providers.
* This estimate does not include the two percent reduction to SNFs' Medicare payments (estimated to be $527.4 million) required by statute.

9. Conclusion

This final rule sets forth updates of the SNF PPS rates contained in the SNF PPS final rule for FY 2018 (82 FR 36530). Based on the above, we estimate the overall estimated payments for SNFs in FY 2019 are projected to increase by approximately $820 million, or 2.4 percent, compared with those in FY 2018. We estimate that in FY 2019 under RUG-IV, SNFs in urban and rural areas will experience, on average, a 2.4 percent increase and 2.5 percent increase, respectively, in estimated payments compared with FY 2018. Providers in the urban rural West South Central region will experience the largest estimated increase in payments of approximately 3.5 percent. Providers in the urban Mountain and rural New England regions will experience the smallest estimated increase in payments of 1.6 percent.

B. Regulatory Flexibility Act Analysis

The RFA requires agencies to analyze options for regulatory relief of small entities, if a rule has a significant impact on a substantial number of small entities. For purposes of the RFA, small entities include small businesses, non-profit organizations, and small governmental jurisdictions. Most SNFs and most other providers and suppliers are small entities, either by reason of their non-profit status or by having revenues of $27.5 million or less in any 1 year. We utilized the revenues of individual SNF providers (from recent Medicare Cost Reports) to classify a small business, and not the revenue of a larger firm with which they may be affiliated. As a result, for the purposes of the RFA, we estimate that almost all SNFs are small entities as that term is used in the RFA, according to the Small Business Administration's latest size standards (NAICS 623110), with total revenues of $27.5 million or less in any 1 year. (For details, see the Small Business Administration's website at http://www.sba.gov/​category/​navigation-structure/​contracting/​contracting-officials/​eligibility-size-standards). In addition, approximately 20 percent of SNFs classified as small entities are non-profit organizations. Finally, individuals and states are not included in the definition of a small entity.

This final rule sets forth updates of the SNF PPS rates contained in the SNF PPS final rule for FY 2018 (82 FR 36530). Based on the above, we estimate that the aggregate impact for FY 2019 will be an increase of $820 million in payments to SNFs, resulting from the SNF market basket update to the payment rates. While it is projected in Table 45 that providers will experience a net increase in payments, we note that some individual providers within the same region or group may experience different impacts on payments than others due to the distributional impact of the FY 2019 wage indexes and the degree of Medicare utilization.

Guidance issued by the Department of Health and Human Services on the proper assessment of the impact on small entities in rulemakings, utilizes a cost or revenue impact of 3 to 5 percent as a significance threshold under the RFA. In their March 2017 Report to Congress (available at http://medpac.gov/​docs/​default-source/​reports/​mar17_​medpac_​ch8.pdf), MedPAC states that Medicare covers approximately 11 percent of total patient days in freestanding facilities and 21 percent of facility revenue (March 2017 MedPAC Report to Congress, 202). As a result, for most facilities, when all payers are included in the revenue stream, the overall impact on total revenues should be substantially less than those impacts presented in Table 45. As indicated in Table 45, the effect on facilities is projected to be an aggregate positive impact of 2.4 percent for FY 2019. As the overall impact on the industry as a whole, and thus on small entities specifically, is less than the 3 to 5 percent threshold discussed previously, the Secretary has determined that this final rule will not have a significant impact on a substantial number of small entities for FY 2019.

In addition, section 1102(b) of the Act requires us to prepare a regulatory impact analysis if a rule may have a significant impact on the operations of a substantial number of small rural hospitals. This analysis must conform to the provisions of section 604 of the RFA. For purposes of section 1102(b) of the Act, we define a small rural hospital as a hospital that is located outside of an MSA and has fewer than 100 beds. This final rule will affect small rural hospitals that (1) furnish SNF services under a swing-bed agreement or (2) have a hospital-based SNF. We anticipate that the impact on small rural hospitals will be similar to the impact on SNF providers overall. Moreover, as noted in previous SNF PPS final rules (most recently, the one for FY 2018 (82 FR 36530)), the category of small rural hospitals is included within the analysis of the impact of this final rule on small entities in general. As indicated in Table 45, the effect on facilities for FY 2019 is projected to be an aggregate positive impact of 2.4 percent. As the overall impact on the industry as a whole is less than the 3 to 5 percent threshold discussed above, the Secretary has determined that this final rule will not have a significant impact on a substantial number of small rural hospitals for FY 2019.

C. Unfunded Mandates Reform Act Analysis

Section 202 of the Unfunded Mandates Reform Act of 1995 also requires that agencies assess anticipated costs and benefits before issuing any rule whose mandates require spending in any 1 year of $100 million in 1995 dollars, updated annually for inflation. In 2018, that threshold is approximately $150 million. This final rule will impose no mandates on state, local, or tribal governments or on the private sector.

D. Federalism Analysis

Executive Order 13132 establishes certain requirements that an agency must meet when it issues a final rule that imposes substantial direct requirement costs on state and local governments, preempts state law, or otherwise has federalism implications. This final rule will have no substantial direct effect on state and local governments, preempt state law, or otherwise have federalism implications.

E. Congressional Review Act

This final regulation is subject to the Congressional Review Act provisions of the Small Business Regulatory Enforcement Fairness Act of 1996 (5 U.S.C. 801 et seq.) and has been transmitted to the Congress and the Comptroller General for review.

F. Regulatory Review Costs

If regulations impose administrative costs on private entities, such as the time needed to read and interpret this final rule, we should estimate the cost associated with regulatory review. Due to the uncertainty involved with accurately quantifying the number of entities that will review the rule, we assume that the total number of unique commenters on last year's proposed rule will be the number of reviewers of this year's proposed rule. We acknowledge that this assumption may understate or overstate the costs of reviewing this rule. It is possible that not all commenters reviewed last year's rule in detail, and it is also possible that some reviewers chose not to comment on the proposed rule. For these reasons, we thought that the number of past commenters is a fair estimate of the number of reviewers of this rule. In the FY 2019 SNF PPS proposed rule (83 FR 21099), we welcomed any comments on the approach in estimating the number of entities which will review the proposed rule.

We also recognize that different types of entities are in many cases affected by mutually exclusive sections of this final rule, and therefore, for the purposes of our estimate we assume that each reviewer reads approximately 50 percent of the rule. We sought comments on this assumption in the FY 2019 SNF PPS proposed rule (83 FR 21099).

Using the wage information from the BLS for medical and health service managers (Code 11-9111), we estimate that the cost of reviewing this rule is $107.38 per hour, including overhead and fringe benefits https://www.bls.gov/​oes/​current/​oes_​nat.htm. Assuming an average reading speed, we estimate that it would take approximately 4 hours for the staff to review half of the proposed rule. For each SNF that reviews the rule, the estimated cost is $429.52 (4 hours × $107.38). Therefore, we estimate that the total cost of reviewing this regulation is $124,561 ($429.52 × 247 reviewers).

In accordance with the provisions of Executive Order 12866, this final rule was reviewed by the Office of Management and Budget.

List of Subjects

42 CFR Part 411

• Diseases
• Medicare
• Reporting and recordkeeping requirements

42 CFR Part 413

• Health facilities
• Diseases
• Medicare
• Reporting and recordkeeping requirements

42 CFR Part 424

• Emergency medical services
• Health facilities
• Health professions
• Medicare
• Reporting and recordkeeping requirements

For the reasons set forth in the preamble, the Centers for Medicare & Medicaid Services amends 42 CFR chapter IV as set forth below:

PART 411—EXCLUSIONS FROM MEDICARE AND LIMITATIONS ON MEDICARE PAYMENT

1. The authority citation for part 411 is revised to read as follows:

Authority: 42 U.S.C. 1302, 1395w-101 through 1395w-152, 1395hh, and 1395nn.

§ 411.15

[Amended]

2. Section 411.15 is amended in paragraph (p)(3)(iv) by removing the phrase “by midnight of the day of departure” and adding in its place the phrase “before the following midnight”.

PART 413—PRINCIPLES OF REASONABLE COST REIMBURSEMENT; PAYMENT FOR END-STAGE RENAL DISEASE SERVICES; PROSPECTIVELY DETERMINED PAYMENT RATES FOR SKILLED NURSING FACILITIES; PAYMENT FOR ACUTE KIDNEY INJURY DIALYSIS

3. The authority citation for part 413 is revised to read as follows:

Authority: 42 U.S.C. 1302, 1395d(d), 1395f(b), 1395g, 1395l(a), (i), and (n), 1395x(v), 1395hh, 1395rr, 1395tt, and 1395ww; and sec. 124 of Public Law 106-113, 113 Stat. 1501A-332; sec. 3201 of Public Law 112-96, 126 Stat. 156; sec. 632 of Public Law 112-240, 126 Stat. 2354; sec. 217 of Public Law 113-93, 129 Stat. 1040; and sec. 204 of Public Law 113-295, 128 Stat. 4010; and sec. 808 of Public Law 114-27, 129 Stat. 362.

4. Section 413.337 is amended by revising paragraph (d)(1)(v) and adding paragraphs (d)(1)(vi) and (vii) and (f) to read as follows:

§ 413.337

Methodology for calculating the prospective payment rates.

* * * * *

(d) * * *

(1) * * *

(v) For each subsequent fiscal year, the unadjusted Federal payment rate is equal to the rate computed for the previous fiscal year increased by a factor equal to the SNF market basket index percentage change for the fiscal year involved, except as provided in paragraphs (d)(1)(vi) and (vii) of this section.

(vi) For fiscal year 2018, the unadjusted Federal payment rate is equal to the rate computed for the previous fiscal year increased by a SNF market basket index percentage change of 1 percent (after application of paragraphs (d)(2) and (3) of this section).

(vii) For fiscal year 2019, the unadjusted Federal payment rate is equal to the rate computed for the previous fiscal year increased by a SNF market basket index percentage change of 2.4 percent (after application of paragraphs (d)(2) and (3) of this section).

* * * * *

(f) Adjustments to payment rates under the SNF Value-Based Purchasing Program. Beginning with payment for services furnished on October 1, 2018, the adjusted Federal per diem rate (as defined in § 413.338(a)(2)) otherwise applicable to a SNF for the fiscal year is reduced by the applicable percent (as defined in § 413.338(a)(3)). The resulting amount is then adjusted by the value-based incentive payment amount (as defined in § 413.338(a)(14)) based on the SNF's performance score for that fiscal year under the SNF Value-Based Purchasing Program, as calculated under § 413.338.

5. Section 413.338 is amended by—

a. Revising the section heading;

b. Adding paragraphs (a)(16) and (17);

c. Revising paragraph (c)(2)(i); and

d. Adding paragraphs (d)(1)(iv) and (d)(3) and (4).

The additions and revision read as follows:

§ 413.338

Skilled nursing facility value-based purchasing program.

(a) * * *

(16) Low-volume SNF means a SNF with fewer than 25 eligible stays included in the SNF readmission measure denominator during the performance period for a fiscal year.

(17) Eligible stay means, for purposes of the SNF readmission measure, an index SNF admission that would be included in the denominator of that measure.

* * * * *

(c) * * *

(2) * * *

(i) Total amount available for a fiscal year. The total amount available for value-based incentive payments for a fiscal year is at least 60 percent of the total amount of the reduction to the adjusted SNF PPS payments for that fiscal year, as estimated by CMS, and will be increased as appropriate for each fiscal year to account for the assignment of a performance score to low-volume SNFs under paragraph (d)(3) of this section.

(d) * * *

(1) * * *

(iv) CMS will not award points for improvement to a SNF that has fewer than 25 eligible stays during the baseline period.

* * * * *

(3) If CMS determines that a SNF is a low-volume SNF with respect to a fiscal year, CMS will assign a performance score to the SNF for the fiscal year that, when used to calculate the value-based incentive payment amount (as defined in paragraph (a)(14) of this section), results in a value-based incentive payment amount that is equal to the adjusted Federal per diem rate (as defined in paragraph (a)(2) of this section) that would apply to the SNF for the fiscal year without application of § 413.337(f).

(4)(i) A SNF may request and CMS may grant exceptions to the SNF Value-Based Purchasing Program's requirements under this section for one or more calendar months when there are certain extraordinary circumstances beyond the control of the SNF.

(ii) A SNF may request an exception within 90 days of the date that the extraordinary circumstances occurred by sending an email to [e-mail] that includes a completed Extraordinary Circumstances Request form (available on the SNF VBP section of QualityNet at https://www.qualitynet.org/​) and any available evidence of the impact of the extraordinary circumstances on the care that the SNF furnished to patients, including, but not limited to, photographs, newspaper, and other media articles.

(iii) Except as provided in paragraph (d)(4)(iv) of this section, CMS will not consider an exception request unless the SNF requesting such exception has complied fully with the requirements in this paragraph (d).

(iv) CMS may grant exceptions to SNFs without a request if it determines that an extraordinary circumstance affects an entire region or locale.

(v) CMS will calculate a SNF performance score for a fiscal year for a SNF for which it has granted an exception request that does not include its performance on the SNF readmission measure during the calendar months affected by the extraordinary circumstance.

* * * * *

6. Section 413.360 is amended by adding paragraph (b)(3) and revising paragraphs (d)(1) and (4) to read as follows:

§ 413.360

Requirements under the Skilled Nursing Facility (SNF) Quality Reporting Program (QRP).

* * * * *

(b) * * *

(3) CMS may remove a quality measure from the SNF QRP based on one or more of the following factors:

(i) Measure performance among SNFs is so high and unvarying that meaningful distinctions in improvements in performance can no longer be made.

(ii) Performance or improvement on a measure does not result in better resident outcomes.

(iii) A measure does not align with current clinical guidelines or practice.

(iv) The availability of a more broadly applicable (across settings, populations, or conditions) measure for the particular topic.

(v) The availability of a measure that is more proximal in time to desired resident outcomes for the particular topic.

(vi) The availability of a measure that is more strongly associated with desired resident outcomes for the particular topic.

(vii) Collection or public reporting of a measure leads to negative unintended consequences other than resident harm.

(viii) The costs associated with a measure outweigh the benefit of its continued use in the program.

* * * * *

(d) * * *

(1) SNFs that do not meet the requirements in paragraph (b) of this section for a program year will receive a written notification of non-compliance through at least one of the following methods: Quality Improvement Evaluation System (QIES) Assessment Submission and Processing (ASAP) system, the United States Postal Service, or via an email from the Medicare Administrative Contractor (MAC). A SNF may request reconsideration no later than 30 calendar days after the date identified on the letter of non-compliance.

* * * * *

(4) CMS will notify SNFs, in writing, of its final decision regarding any reconsideration request through at least one of the following notification methods: QIES ASAP system, the United States Postal Service, or via email from the Medicare Administrative Contractor (MAC).

* * * * *

PART 424—CONDITIONS FOR MEDICARE PAYMENT

7. The authority citation for part 424 is revised to read as follows:

Authority: 42 U.S.C. 1302 and 1395hh.

8. Section 424.20 is amended by revising paragraph (a)(1)(i) to read as follows:

§ 424.20

Requirements for posthospital SNF care.

* * * * *

(a) * * *

(1) * * *

(i) The individual needs or needed on a daily basis skilled nursing care (furnished directly by or requiring the supervision of skilled nursing personnel) or other skilled rehabilitation services that, as a practical matter, can only be provided in an SNF or a swing-bed hospital on an inpatient basis, and the SNF care is or was needed for a condition for which the individual received inpatient care in a participating hospital or a qualified hospital, as defined in § 409.3 of this chapter, or for a new condition that arose while the individual was receiving care in the SNF or swing-bed hospital for a condition for which he or she received inpatient care in a participating or qualified hospital; or.

* * * * *

Dated: July 26, 2018.

Seema Verma,

Administrator, Centers for Medicare & Medicaid Services.

Dated: July 26, 2018.

Alex M. Azar II,

Secretary, Department of Health and Human Services.

Footnotes

1.  Meaningful Measures web page: https:/www.cms.gov/​Medicare/​Quality-Initiatives-Patient-Assessment-Instruments/​QualityInitativesGenIndo/​MMF/​General-info-Sub-Page.html.

2.  See Remarks by Administrator Seema Verma at the Health Care Payment Learning and Action Network (LAN) Fall Summit, as prepared for delivery on October 30, 2017 https://www.cms.gov/​Newsroom/​MediaReleaseDatabase/​Fact-sheets/​2017-Fact-Sheet-items/​2017-10-30.html

3.  See, for example, United States Department of Health and Human Services. “Healthy People 2020: Disparities. 2014.” Available at http://www.healthypeople.gov/​2020/​about/​foundation-health-measures/​Disparities; or National Academies of Sciences, Engineering, and Medicine. Accounting for Social Risk Factors in Medicare Payment: Identifying Social Risk Factors. Washington, DC: National Academies of Sciences, Engineering, and Medicine 2016.

4.  Department of Health and Human Services Office of the Assistant Secretary for Planning and Evaluation (ASPE), “Report to Congress: Social Risk Factors and Performance Under Medicare's Value-Based Purchasing Programs.” December 2016. Available at https://aspe.hhs.gov/​pdf-report/​report-congress-social-risk-factors-and-performance-under-medicares-value-based-purchasing-programs.

5.  Available at http://www.qualityforum.org/​SES_​Trial_​Period.aspx.

6.  Available at http://www.qualityforum.org/​WorkArea/​linkit.aspx?​LinkIdentifier=​ id&ItemID=86357 .

7.  We refer readers to the FY 2016 SNF PPS final rule (80 FR 46431 through 46432) for more information on the factors we consider for removing measures.

8.  Available at http://www.qualityforum.org/​Publications/​2016/​02/​MAP_​2016_​Considerations_​for_​Implementing_​Measures_​in_​Federal_​Programs_​-_​PAC-LTC.aspx.

9.  See, for example United States Department of Health and Human Services. “Healthy People 2020: Disparities. 2014.” Available at http://www.healthypeople.gov/​2020/​about/​foundation-health-measures/​Disparities; or National Academies of Sciences, Engineering, and Medicine. Accounting for Social Risk Factors in Medicare Payment: Identifying Social Risk Factors. Washington, DC: National Academies of Sciences, Engineering, and Medicine 2016.

10.  Department of Health and Human Services Office of the Assistant Secretary for Planning and Evaluation (ASPE), “Report to Congress: Social Risk Factors and Performance Under Medicare's Value-Based Purchasing Programs.” December 2016. Available at https://aspe.hhs.gov/​pdf-report/​report-congress-social-risk-factors-and-performance-under-medicares-value-based-purchasing-programs.

11.  Available at http://www.qualityforum.org/​SES_​Trial_​Period.aspx.

12.  Available at http://www.qualityforum.org/​WorkArea/​linkit.aspx?​LinkIdentifier=​id&​ ItemID=86357.