04-24787. Medicare Program; Prospective Payment System for Inpatient Psychiatric Facilities

AGENCY:

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

ACTION:

Final rule.

SUMMARY:

This final rule establishes a prospective payment system for Medicare payment of inpatient hospital services furnished in psychiatric hospitals and psychiatric units of acute care hospitals and critical access hospitals. It implements section 124 of the Medicare, Medicaid, and SCHIP Balanced Budget Refinement Act of 1999 (BBRA). The prospective payment system described in this final rule will replace the reasonable cost-based payment system under which psychiatric hospitals and psychiatric units are paid under Medicare.

DATES:

This rule is effective for cost reporting periods beginning on or after January 1, 2005.

FOR FURTHER INFORMATION CONTACT:

Janet Samen, [phone] (General information.) Phillip Cotterill, [phone] and Fred Thomas [phone], (For information regarding the regression analysis).

SUPPLEMENTARY INFORMATION:

Copies: To order copies of the Federal Register containing this document, send your request to: New Orders, Superintendent of Documents, P.O. Box 371954, Pittsburgh, PA 15250-7954. Specify the date of the issue requested and enclose a check or money order payable to the Superintendent of Documents, or enclose your Visa or Master Card number and expiration date. Credit card orders can also be placed by calling the order desk at [phone] or by faxing to [phone]. The cost for each copy is $10. As an alternative, you can view and photocopy the Federal Register document at most libraries designated as Federal Depository Libraries and at many other public and academic libraries throughout the country that receive the Federal Register.

This Federal Register document is also available from the Federal Register online database through GPO access, a service of the U.S. Government Printing Office. The Website address is http://www.access.gpo.gov/​nara/​index.html.

To assist readers in referencing sections contained in this document, we are providing he following table of contents.

Table of Contents

I. Background

A. General and Legislative History

B. Overview of the Payment System for Inpatient Psychiatric Hospitals and Psychiatric Units Before the BBRA

1. Description of the TEFRA Payment Methodology

2. BBA Amendments to TEFRA

3. BBRA Amendments to TEFRA

4. BIPA Amendments to TEFRA

II. Provisions of the Proposed Regulations

III. Analysis of and Responses to Public Comments

IV. Overview of the IPF PPS Proposed Payment Methodology

V. Development of the Budget-Neutral Federal Per Diem Base Rate.

A. Calculation of the Federal Per Diem Base Rate

B. Determining the Update Factors for the Budget-Neutrality Calculation

1. The 1997-Based Excluded Hospital with Capital Market Basket

2. Calculating the Budget-Neutrality Adjustment Factor

a. Cost Report Data for January 1, 2005 through June 30, 2006

b. Estimate of Total Payments Under the TEFRA Payment System

c. Payments Under the IPF PPS without a Budget-Neutrality Adjustment

C. Standardization and Budget-Neutrality Adjustments

D. Calculation of the Budget Neutrality Adjustments

1. Outlier Adjustment

2. Stop Loss Adjustment

3. Behavioral Offset

VI. Cost Regression Used To Develop Payment Adjustment Factors

A. Final Regression Analysis

B. Patient-Level Adjustments

1. Adjustment for DRG Assignment

2. Comorbidities

3. Other Coding Issues

4. Patient Age

5. Variable Per Diem Adjustments

6. Other Patient-Level Adjustments

a. Gender

b. Patients admitted through the Hospital's Emergency Department (ED)

c. Patients who Receive Electroconvulsive Therapy (ECT)

d. Patient's Involuntarily Committed to the IPF

e. Administrative Necessary Days

C. Facility-Level Adjustments

1. Wage Index

2. Rural Location

3. Teaching Status Adjustments

4. Other Facility-Level Adjustments

a. Adjustment for Psychiatric Units

b. Cost of Living Adjustment

i. IPFs Located in Alaska and Hawaii

ii. IPFs Located in California

c. Disproportionate Share Intensity

d. IPFs with Full-Service Emergency Departments

D. Other Proposed Adjustments and Policy Changes

1. Outlier Policy

a. Statistical Accuracy of Cost-for-Change Ratios

b. Adjustments of IPF Outlier Payments

2. Interrupted Stays

3. Stop-Loss Provision

4. Physician Recertification Requirements

VII. Implementation of the IPF PPS

A. Transition Period

1. Existing Providers

2. New Providers

B. Claims Processing

C. Annual Update

VIII. Future Refinements

IX. Comments Beyond the Scope of the Final Rule

X. Provisions of the Final Rule

XI. Collection of Information Requirements

XII. Regulatory Impact Analysis

A. Overall Impact

B. Anticipated Effects

1. Budgetary Impacts

2. Impacts on Providers

3. Results

a. Facility Type

b. Location

c. Teaching Status Adjustment

d. Census Region

e. Size

4. Effect on the Medicare Program

5. Effect on Beneficiaries

6. Computer Hardware and Software

C. Alternatives Considered

Regulation Text

Addendum A: Proposed Inpatient PPS Adjustments

Addendum B: Wages

Addendum C: “Code First” List

Acronyms

Because of the many terms to which we refer by acronym in this final rule, we are listing the acronyms used and their corresponding terms in alphabetical order below:

BBA Balanced Budget Act of 1997 (Pub. L. 105-33)

BBRA Medicare, Medicaid and SCHIP [State Children's Health Insurance Program] Balanced Budget Refinement Act of 1999 (Pub. L. 106-113)

BIPA Medicare, Medicaid, and SCHIP [State Children's Health Insurance Program] Benefits Improvement and Protection Act of 2000 (Pub. L. 106-554)

CMS Centers for Medicare & Medicaid Services

DSM-IV-TR Diagnostic and Statistical Manual of Mental Disorders Fourth Edition—Text Revision

DRGs Diagnosis-related groups

FY Federal fiscal year

HCRIS Hospital Cost Report Information System

ICD-9-CM International Classification of Diseases, 9th Revision, Clinical Modification

IPFs Inpatient psychiatric facilities

IPPS Hospital Inpatient Prospective Payment System

IRFs Inpatient rehabilitation facilities

LTCHs Long-term care hospitals

MedPAR Medicare provider analysis and review file

MMA Medicare Prescription Drug, Improvement, and Modernization Act of 2003 (Pub. L. 108-173)

PIP Periodic interim payments

PPS Prospective Payment System

TEFRA Tax Equity and Fiscal Responsibility Act of 1982, (Pub. L. 97-248)

I. Background

A. General and Legislative History

When the Medicare statute was originally enacted in 1965, Medicare payment for inpatient hospital services was based on the reasonable costs incurred in furnishing services to Medicare beneficiaries. Section 223 of the Social Security Act Amendments of 1972 (Pub. L. 92-603) amended section 1861(v)(1) of the Social Security Act (the Act) to set forth limits on reasonable costs for inpatient hospital services. The statute was later amended by section 101(a) of the Tax Equity and Fiscal Responsibility Act of 1982 (TEFRA) (Pub. L. 97-248) to limit payment by placing a limit on allowable costs per discharge.

The Congress directed implementation of a prospective payment system (PPS) for acute care hospitals in 1983, with the enactment of Public Law 98-21. Section 601 of the Social Security Amendments of 1983 (Pub. L. 98-21) added a new section 1886(d) to the Act that replaced the reasonable cost-based payment system for most inpatient hospital services with a PPS.

Although most inpatient hospital services became subject to the PPS, certain specialty hospitals were excluded from the PPS and continued to be paid reasonable costs subject to limits imposed by TEFRA. These hospitals included psychiatric hospitals and psychiatric units in acute care hospitals, long-term care hospitals (LTCH), children's hospitals, and rehabilitation hospitals and rehabilitation units in acute care hospitals. Cancer hospitals were added to the list of excluded hospitals by section 6004(a) of the Omnibus Budget Reconciliation Act of 1989 (Pub. L. 101-239).

The Congress enacted various provisions in the Balanced Budget Act of 1997 (BBA) (Pub. L. 105-33), the Medicare, Medicaid, and SCHIP [State Children's Health Insurance Program] Balanced Budget Refinement Act (BBRA) (Pub. L. 106-113), and the Medicare, Medicaid, and SCHIP Benefits Improvement and Protection Act of 2000 (BIPA) (Pub. L. 106-554) to replace the cost-based methods of reimbursement with a PPS for the following excluded hospitals:

• Rehabilitation hospitals and rehabilitation units in acute care hospitals.
• Psychiatric hospitals and psychiatric units in acute care hospitals.
• Long term care hospitals.

The BBA also imposed national limits (or caps) on hospital-specific target amounts (that is, annual per discharge limits) for these hospitals until cost reporting periods beginning on or after October 1, 2002. A detailed description of the TEFRA payment methodology is provided in section B.1. of this final rule.

Section 124 of the BBRA mandated that the Secretary—(1) develop a per diem PPS for inpatient hospital services furnished in psychiatric hospitals and psychiatric units (hereinafter referred to as inpatient psychiatric facilities (IPFs)); (2) include in the PPS an adequate patient classification system that reflects the differences in patient resource use and costs among psychiatric hospitals and psychiatric units; (3) maintain budget neutrality; (4) permit the Secretary to require psychiatric hospitals and psychiatric units to submit information necessary for the development of the PPS; and (5) submit a report to the Congress describing the development of the PPS.

Section 124 of the BBRA also required that the PPS for IPFs be implemented for cost reporting periods beginning on or after October 1, 2002. In general, the creation of a prospective payment system requires an extraordinary amount of lead-time in order to conduct the research that is required to create a completely new payment system. For example, we must create data files, develop models to test individual variables and those variables' ability to explain costs, as well as perform extensive empirical analysis of the collected data.

With respect to the creation of the IPF PPS, more lead time than usual was necessary. This is because the research we had conducted before the passage of the BBRA dated back to the 1980s and was focused on developing a per discharge IPF PPS. The research efforts to develop a discharged-based IPF PPS, however, failed to adequately explain cost variation among psychiatric cases. Because diagnosis in psychiatry is complicated and the criteria for diagnosis and treatment are less well defined in psychiatry than in general medicine and surgery, developing an IPF PPS was more elusive. Moreover, there have been significant changes in mental health treatment, for example, new medications and outpatient treatment options. Thus, to develop an adequate patient classification system that reflects the differences in patient resource use and costs, we had to embark on numerous courses of research that could be used as a possible foundation for the proposed IPF PPS.

When we began the process of developing a proposed IPF PPS, we believed pursuing an assessment instrument, incorporating key indicators of functional status, was the most logical place to begin. This approach is consistent with the approach we followed in developing patient classification systems for other Medicare prospective payment systems (for example., home health agencies, skilled nursing facilities, and inpatient rehabilitation facilities). Our administrative data was inadequate to develop other patient classification systems because, although it provides useful information on diagnoses, services, and procedures, it does not include many patient and clinical characteristics and functional status indicators, which have been established as key components of a patient classification system. Therefore, to obtain the patient-level data we needed to develop an assessment-based patient classification system, we contracted with the University of Michigan's Public Health Institute in September 2002. We selected this contractor because it had developed a protocol assessment instrument, precursors of which had shown promise in explaining variation in resource utilization among psychiatric patients. Although there continues to be progress in completing the initial phase of this research, that is, adoption of an initial assessment instrument for pilot testing, we are unable to delay implementation of the IPF PPS until the draft assessment instrument is completed.

Also, in our effort to meet the requirements of section 124 of the BBRA, we also pursued a second research project with the Health, Economics, Research, Inc. (now known as RTI International^®). RTI International^® embarked on a research project to identify patient characteristics and modes of practice believed to account for variation in per diem cost. It became apparent that, despite everyone's best efforts, the ongoing research projects being conducted by the University of Michigan and RTI International^®, could not be completed in time for us to engage in notice and comment rulemaking and achieve implementation of the IPF PPS by October 1, 2002.

In addition, shortly before October 1, 2002, the American Psychiatric Association (APA) informed us that The Health Economics and Outcomes Research Institute (THEORI) of the Greater New York Hospital Association had developed a potential IPF PPS classification model that was based on our currently available administrative data. Based on the model presented to us by the APA, we immediately began our own vigorous review of the “APA” model. We note, however, that although the information shared with us by the APA was extremely valuable in our formulation of a proposed IPF PPS, it came too late for us to be able to do the following: (1) Perform the analysis required to ensure that a system based on our administrative data would fulfill the statutory mandate of section 124 of the BBRA; and (2) engage in notice-and-comment rulemaking and implement the IPF PPS by October 1, 2002. As soon as we completed an analysis of the information presented by the APA and of our administrative data, we published the proposed IPF PPS regulation.

Initially, the proposed rule provided for a 60-day comment period. However, due to the complexity and scope of the proposed rule and because the public requested additional time to examine the rule so that it could provide meaningful comments, we extended the public comment period. The intricacy and complexity of the issues presented in the public comments required us to perform further substantial analysis to adequately address the issues raised by commenters, as well as our duty to satisfy section 124 of the BBRA. We have made every effort to complete this final rule as quickly as possible.

(We note that, even though the IPF PPS described in this final rule is effective for cost reporting periods beginning on or after January 1, 2005 and compliance with the IPF PPS requirements is required for cost reporting periods beginning on or after January 1, 2005, we will not have computer system changes in place that are necessary to accommodate claims processing under the IPF PPS until April 4, 2005 (claims processing updates will occur on the first Monday following April 1, 2005). Therefore, claims submitted after January 1, 2005, but before April 4, 2005, will be paid as if the TEFRA rate was still in effect. Payments will be reconciled with the appropriate IPF PPS amount. We have instructed the fiscal intermediaries (FIs) to reconcile the payments that are made to IPFs for covered inpatient hospital services furnished to Medicare beneficiaries for cost reporting periods beginning on or after January 1, 2005, until the date of the systems implementation on April 4, 2005, with the amounts that are payable under the IPF PPS system by May 1, 2005.

Since IPFs will receive payment under the IPF PPS starting with their first cost reporting period beginning on or after January 1, 2005, only those IPFs with cost reporting periods beginning on or after January 1, 2005 but before April 1, 2005 will experience payment reconciliation.

Requirements for Issuance of Regulations

Section 902 of the Medicare Prescription Drug, Improvement, and Modernization Act of 2003 (MMA) amended section 1871(a) of the Act and requires the Secretary, in consultation with the Director of the Office of Management and Budget, to establish and publish timelines for the publication of Medicare final regulations based on the previous publication of a Medicare proposed or interim final regulation. Section 902 of the MMA also states that the timelines for these regulations may vary but will not exceed 3 years after publication of the preceding proposed or interim final regulation except under exceptional circumstances.

This rule finalizes provisions set forth in the November 28, 2003 proposed rule (68 FR 66920). In addition, this final rule has been published within the 3-year time limit imposed by section 902 of the MMA. Therefore, we believe that the final rule is in accordance with the Congress' intent to ensure timely publication of final regulations.

B. Overview of the Payment System for Inpatient Psychiatric Hospitals and Psychiatric Units Before the BBRA

1. Description of the TEFRA Payment Methodology

Hospitals and units that are excluded from the hospital inpatient prospective payment system (IPPS) under section 1886(d)(1)(B) of the Act are paid for their inpatient operating costs under the provisions of the TEFRA (Pub. L. 97-248).

The TEFRA provisions are found in section 1886(b) of the Act and implemented in regulations at 42 CFR 413. TEFRA established payments based on hospital-specific limits for inpatient operating costs. As specified in § 413.40, TEFRA established a ceiling on payments for hospitals excluded from the IPPS. The ceiling on payments is determined by calculating the product of a facility's base year costs (the year in which its target reimbursement limit is based) per discharge, updated to the current year by a rate-of-increase percentage, and multiplied by the number of total current year discharges. A detailed discussion of target amount payment limits under TEFRA can be found in the final rule concerning the IPPS published in the Federal Register on September 1, 1983 (48 FR 39746).

The base year for a facility varied, depending on when the facility was initially determined to be an IPPS excluded provider. The base year for facilities that were established before the implementation of the TEFRA provision was 1982. For facilities established after the implementation of the TEFRA provision, facilities were allowed to choose which of their first 3 cost reporting years would be used in the future to determine their target limit. In 1992, the “new provider” period was shortened to 2 full years of cost reporting periods (§ 413.40(f)(1)).

Excluded facilities whose costs were below their target amounts would receive bonus payments equal to the lesser of half of the difference between costs and the target amount, up to a maximum of 5 percent of the target amount, or the hospital's costs. For excluded hospitals whose costs exceeded their target amounts, Medicare provided relief payments equal to half of the amount by which the hospital's costs exceeded the target amount up to 10 percent of the target amount. Excluded facilities that experienced a more significant increase in patient acuity could also apply for an additional amount as specified in § 413.40(d) for Medicare exception payments.

2. BBA Amendments to TEFRA

The BBA amendments to section 1886 of the Act significantly altered the payment provisions for hospitals and units paid under the TEFRA provisions and added other qualifying criteria for certain hospitals excluded from the IPPS. A complete explanation of these amendments can be found in the final rule concerning the IPPS we published in the Federal Register on August 29, 1997 (62 FR 45966).

The BBA made the following changes to section 1886 of the Act for TEFRA hospitals:

• Section 4411 of the BBA amended section 1886(b)(3)(B) of the Act and restricted the rate-of-increase percentages that are applied to each provider's target amount so that excluded hospitals and units experiencing lower inpatient operating costs relative to their target amounts receive lower rates of increase.
• Section 4412 of the BBA amended section 1886(g) of the Act to establish a 15-percent reduction in capital payments for excluded psychiatric and rehabilitation hospitals and units and LTCHs, for portions of cost reporting periods occurring during the period of October 1, 1997, through September 30, 2002.
• Section 4414 of the BBA amended section 1886(b)(3) of the Act to establish caps on the target amounts for excluded hospitals and units at the 75th percentile of target amounts for similar facilities for cost reporting periods beginning on or after October 1, 1997, through September 30, 2002. The caps on these target amounts apply only to psychiatric hospitals and rehabilitation hospital units and LTCHs. Payments for these excluded hospitals and units are based on the lesser of a provider's cost per discharge or its hospital-specific cost per discharge, subject to this cap.
• Section 4415 of the BBA amended section 1886(b)(1) of the Act by revising the percentage factors used to determine the amount of bonus and relief payments and establishing continuous improvement bonus payments for excluded hospitals and units for cost reporting periods beginning on or after October 1, 1997. If a hospital is eligible for the continuous improvement bonus, the bonus payment is equal to the lesser of: (1) 50 percent of the amount by which operating costs are less than expected costs; or (2) 1 percent of the target amount.
• Sections 4416 and 4419 of the BBA amended sections 1886(b) of the Act to establish a new framework for payments for new excluded providers. Section 4416 of the BBA added a new section 1886(b)(7) to the Act that established a new statutory methodology for new psychiatric and rehabilitation hospitals and units, and LTCHs. Under section 4416 of the BBA, payment to these providers for their first two cost reporting periods is limited to the lesser of the operating costs per case, or 110 percent of the national median of target amounts. This is adjusted for differences in wage levels, for the same class of hospital for cost reporting periods ending during FY 1996, updated to the applicable period.

3. BBRA Amendments to TEFRA

The BBRA of 1999 refined some of the policies mandated by the BBA for hospitals and units paid under the TEFRA provisions. The provisions of the BBRA, amending section 1886(b)(3)(H) of the Act, were explained in detail and implemented in the IPPS interim final rule published in the Federal Register on August 1, 2000 (65 FR 47026) and in the IPPS final rule also published on August 1, 2000 (65 FR 47054).

With respect to the TEFRA payment methodology, section 4414 of the BBA had provided for caps on target amounts for excluded hospitals and units for cost reporting periods beginning on or after October 1, 1997. Section 121 of the BBRA amended section 1886(b)(3)(H) of the Act to provide for an appropriate wage adjustment to these caps on the target amounts for certain hospitals and units paid under the TEFRA provisions, effective for cost reporting periods beginning on or after October 1, 1999 through September 30, 2002.

4. BIPA Amendments to TEFRA

Section 306 of BIPA amended section 1886 of the Act by increasing the incentive payments for psychiatric hospitals and psychiatric units to 3 percent for cost reporting periods beginning on or after October 1, 2000 and before October 1, 2001.

II. Provisions of the Proposed Regulations

On November 28, 2003, we published a proposed rule in the Federal Register (68 FR 66920) as required by section 124 of the BBRA that proposed a PPS for Medicare payment of inpatient hospital services furnished in IPFs. The IPF PPS would replace the current reasonable cost-based payment system under the TEFRA provisions.

We proposed to base the IPF PPS on data from the fiscal year (FY) 1999 Medicare Provider Analysis and Review (MedPAR) file, which includes patient characteristics (for example, patients' diagnoses and age), and data from the FY 1999 Hospital Cost Report Information System (HCRIS), which includes facility characteristics (for example, location and teaching status). We proposed the following policies and methodology for the IPF PPS. We proposed to:

• Add a new subpart N in 42 CFR 412 for the IPF PPS, and make conforming changes to parts 412 and 413 regarding the implementation of the IPF PPS.
• Compute a standardized Federal per diem payment to be paid to all IPFs based on the sum of the national average routine operating, ancillary, and capital costs for each patient day of psychiatric care in an IPF adjusted for budget neutrality.
• Adjust the Federal per diem payment to reflect certain patient and facility characteristics that were found in the regression analysis to be associated with statistically significant cost differences.
• Provide patient-level adjustments for age, specified diagnosis-related groups (DRGs), and selected comorbidity categories.
• Provide facility adjustments that include a wage index adjustment, rural location adjustment, and a teaching status adjustment.
• Recognize variable per diem adjustments to account for the higher costs incurred in the early days of a psychiatric stay.
• Adopt an outlier policy to target greater payment to the high cost cases.
• Provide an interrupted stay policy for the purpose of applying the variable per diem adjustment and the outlier policy.
• Implement the IPF PPS for IPF cost reporting periods beginning on or after April 1, 2004, with a 3-year transition period. We proposed that the first update would occur on July 1, 2005.
• Include a coding policy that would require IPFs to report patient diagnoses using the International Classification of Diseases-9th Revision, Clinical Modification (ICD-9-CM) code set.
• Update a regulatory reference to the Diagnostic and Statistical Manual of Mental Disorders (DSM) from the Third Edition to the Fourth Edition, Text Revision (DSM-IV-TR).
• Use the 1997-based excluded hospital with capital market basket to establish the labor-related share of the Federal per diem base rate, to calculate the budget neutrality adjustment, and to update the Federal per diem base rate.
• Provide the annual update strategy for the IPF PPS.
• Include research information for future refinement of the patient classification system.

III. Analysis of and Responses to Public Comments

In the November 28, 2003 Federal Register (68 FR 66920), we published the proposed IPF PPS and provided for a 60-day comment period. On January 30, 2004, we published a notice in the Federal Register (68 FR 4464) extending the comment period for an additional 30 days in response to public requests. The comment period that would have closed on January 27, 2004, was extended 30 days. Thus, the comment period for the proposed rule closed on February 26, 2004.

We received 273 comments from hospital associations, psychiatric hospitals, providers, acute care hospitals, health research organizations, patient advocacy organizations, State associations, and physicians. We reviewed each commenter's letter and grouped related comments. Some comments were identical. After associating like comments, we placed them in categories based on subject matter or based on the section(s) of the regulation affected. Summaries of the public comments received and our responses to those comments are set forth below.

IV. Overview of the IPF PPS Proposed Payment Methodology

In the November 2003 proposed rule, we proposed to establish a Federal payment for each patient day in an IPF derived from the national average daily routine operating, ancillary, and capital costs in IPFs. The Federal per diem payment would comprise a Federal per diem base rate adjusted by factors for patient and facility characteristics that account for variation in patient resource use. The Federal per diem base rate would be updated to the midpoint of the first year under the IPF PPS, standardized to account for the overall positive effects of the IPF PPS payment adjustments, and adjusted for budget neutrality.

We proposed that psychiatric hospitals and psychiatric units paid under section 1886(b) of the Act would be paid under the IPF PPS for cost reporting periods beginning on or after April 1, 2004. We proposed that the IPF PPS would apply to inpatient hospital services furnished by Medicare participating entities in the United States that are classified as psychiatric hospitals or psychiatric units as specified in § 412.22, § 412.23, § 412.25, and § 412.27. As specified in § 400.200, the United States means the 50 States, the District of Columbia, the Commonwealth of Puerto Rico, the Virgin Islands, Guam, American Samoa, and the Northern Mariana Islands.

However, the following hospitals are paid under special payment provisions specified in § 412.22(c) and, therefore, would not be paid under the IPF PPS:

• Veterans Administration hospitals.
• Hospitals that are reimbursed under State cost control systems approved under 42 CFR part 403.
• Hospitals that are reimbursed in accordance with demonstration projects specified in section 402(a) of Public Law 90-248 (42 U.S.C. 1395b-1) or section 222(a) of Public Law 92-603 (42 U.S.C. 1395b-1(note)).
• Non-participating hospitals furnishing emergency services to Medicare beneficiaries.

We received a variety of comments on the proposed applicability requirements of the IPF PPS. In this final rule, we are adopting the proposed policies regarding applicability of the IPF PPS.

Comment: One commenter recommended that CMS develop a separate payment system for government-operated IPFs. The commenter believes that these hospitals provide a different service than other psychiatric hospitals and psychiatric units.

Several commenters requested that psychiatric units be excluded from the IPF PPS until a more equitable system can be created.

Response: Section 124 of Public Law 106-113 requires the Secretary to implement a prospective payment system for psychiatric hospitals described in clause (i) of section 1886(d)(1)(B) of the Act and psychiatric units described in clause (v) of this section. Government-operated psychiatric hospitals and psychiatric units fall within the definition of a psychiatric hospital and unit outlined in section 124 of the BBRA to which this IPF PPS applies. Consequently, these entities, like all other psychiatric hospitals and units, must be paid under this system effective with the start of the implementation of the IPF PPS.

With regard to the equity of the payment system, we believe that we are implementing an equitable prospective payment system based on the best data available.

We also believe it is important to note that a per diem approach explains a significant percentage of the cost variation among inpatient psychiatric patients. We estimate that the final IPF PPS explains the 33 percent variation in per diem cost among IPF cases. A commenter indicated that the combination of the explanatory power of a per diem system and the proposed adjustments on case level costs is approximately 80 percent. Our analysis confirmed the commenter's findings, however, we found the explanatory power of a per diem system and the final adjustment factors to be approximately 85 percent, solidifying our belief that the payment model combination we are using, a per diem system with adjustments based on case level costs, is equitable.

Comment: One commenter questioned whether psychiatric units that are currently paid under the IPPS and do not meet the requirements of § 412.22, § 412.25, and § 412.27 would be excluded from the IPF PPS. The commenter also asked whether these providers would be paid under the IPF PPS if they would meet the requirements of § 412.22, § 412.25, and § 412.27. A few commenters asked if “DRG-exempt status” for psychiatric units would continue to be an option after the effective date of the IPF PPS.

Response: If a hospital has a psychiatric unit that meets the requirements specified in § 412.22, § 412.25, and § 412.27, the psychiatric unit is excluded from the IPPS (that is, DRG-exempt). The IPF PPS will replace the reasonable cost-based payments currently paid to excluded psychiatric hospitals and units for cost reporting periods beginning on or after January 1, 2005. Once the IPF PPS is implemented, hospitals will be paid under the IPF PPS for all patients admitted to the excluded psychiatric unit.

Comment: One commenter recommended that critical access hospitals (CAHs) be allowed cost-based reimbursement for services in their psychiatric units. If a hospital or unit treats psychiatric patients but it does not meet the statutory definition of a psychiatric hospital or unit, then the IPF PPS would not apply.

Response: Section 405(g)(2) of the MMA specifies that the amount of payment for services in psychiatric units of a CAH described in section 1820(c)(2)(E) of the Act shall be equal to the amount that would otherwise be made if the services were inpatient hospital services provided in a distinct part psychiatric unit. Therefore, we have amended § 413.70(e) to clarify that, effective for cost reporting periods beginning on or after January 1, 2005, certified psychiatric units in CAHs will be paid under the IPF PPS. We believe the statute is very clear concerning methodology.

Comment: Several commenters requested an exceptions process through which an IPF could seek additional payment.

Response: We believe that the final IPF PPS explains a sufficient amount of the cost variation among IPF patients and that an exceptions process is not necessary.

More importantly, when we become aware of patient or facility characteristics that lead to higher per diem costs, we would propose to establish an adjustment factor to the IPF PPS so that all IPFs that qualify could benefit from the adjustment as part of routine claims processing rather than through an exceptions process through which an individual IPF could request additional payment. Therefore, we will be accounting for their differences in costs.

V. Development of the Budget-Neutral Federal Per Diem Base Rate

In the proposed rule, we proposed that the IPF PPS be based on a standardized Federal per diem base rate calculated from IPF average per diem costs and adjusted for budget-neutrality. We proposed that the Federal per diem base rate would be used as the standard payment per day for the IPF PPS. In addition, the Federal per diem base rate would be adjusted by the applicable wage index factor and the patient-level and facility-level adjustments that are applicable to the stay.

A. Calculation of the Average Per Diem Cost

To calculate the proposed Federal per diem base rate, we estimated the cost per day for—(1) routine services from FY 1999 cost reports (supplemented with FY 1998 cost reports if the FY 1999 cost report is missing); and (2) ancillary costs per day using data from the FY 1999 Medicare claims and corresponding data from facility cost reports.

For routine services, the per diem operating and capital costs were used to develop the base for the psychiatric per diem amount. The per diem routine costs were obtained from each facility's Medicare cost report. To estimate the costs for routine services included in the proposed Federal per diem base rate calculation, we added the total routine costs (including costs for capital) submitted on the cost report for each provider and divided it by the total Medicare days.

Some average routine costs per day were determined to be aberrant, that is, the costs were extraordinarily high or low and most likely contained data errors. The following method was used to trim extraordinarily high or low cost values in order to improve the accuracy of our results.

First, the average and standard deviations of the total per diem cost (routine and ancillary costs) were computed separately for cases from psychiatric hospitals and psychiatric units. Separate statistics were computed because we did not want to systematically exclude a larger proportion of cases from the higher cost psychiatric units. Before calculating the means, we trimmed cases from the file when covered days were zero or routine costs were less than $100 or greater than $3,000. We selected these amounts because we believe this range captured the grossly aberrant cases. Elimination of the grossly aberrant cases would prevent the means from being distorted.

Second, we trimmed cases when the provider's total cost per day was outside the generally-accepted statistical trim points of plus or minus 3.00 standard deviations from the respective means for each facility type (psychiatric hospitals and psychiatric units). If the total cost per day was outside the trim value, we deleted the data for that provider from the per diem rate development file because it helped eliminate skewing of the data. After trimming the data, the average routine cost per day in FY 1999 was calculated to be $495.

For ancillary services, we calculated the costs by converting charges from the FY 1999 Medicare claims into costs using facility-specific, cost-center specific cost-to-charge ratios obtained from each provider's applicable cost reports. We matched each provider's departmental cost-to-charge ratios from their Medicare cost report to each charge on their claims reported in the MedPAR file. Multiplying the total charges for each type of ancillary service by the corresponding cost-to-charge ratio provided an estimate of the costs for all ancillary services received by the patient during the stay.

For those departmental cost-to-charge ratios that we considered to be aberrant because they were outside the generally-accepted statistical trim points of plus or minus 3.00 standard deviations from the facility-type mean, we replaced the individual cost-to-charge ratios for each department with the median department cost-to-charge ratio by facility type (psychiatric hospital or psychiatric unit). We considered using the mean of the cost to-charge ratio as the substitution value, but because the distribution of ratios of cost-to-charges is not normally distributed and there is no limit to the upper ceiling of the ratio, the mean ratio would be overstated due to the higher values on the upper tail of the bell curve. Therefore, we chose the median by facility type as a better measure for the substitution value when the facility's actual cost-to-charge ratio was outside the trim values.

After computing the estimated costs of applying the applicable cost-to-charge ratios, and, when appropriate, the median cost-to charge ratio, to the total ancillary charges for each patient stay, we determined the average ancillary amount per day by dividing the total ancillary costs for all stays by the total number of covered Medicare days. Using this methodology, the average ancillary cost per day in FY 1999 was calculated to be $67.

Adding the average ancillary costs per day ($67) and the average routine costs per day including capital costs ($495) provides the estimated average per diem cost for each patient day of inpatient psychiatric care in FY 1999 ($562). We used the above described procedures to calculate the average per diem cost in this final rule as well.

Comment: Several commenters recommended that CMS use more current data for the final IPF PPS. The commenters suggested that CMS use the FY 2002 MedPAR data and the FY 2002 HCRIS data, supplemented with FY 2001 cost report data when necessary.

A few commenters indicated it would be preferable to use the most current cost report data, with an appropriate audit adjustment factor, if necessary.

Response: We used the best available data when we developed the proposed rule. We are continuing to use the best data available for this final rule. Specifically, we calculated the average cost per day using FY 2002 claims and cost report data supplemented with FY 2001 cost report data if the FY 2002 cost report was missing. Using FY 2002 data and the methodology described above, we calculated the per diem cost for each patient day of inpatient psychiatric care in an IPF in FY 2002. We note that currently, less than 50 percent of the hospitals have filed their FY 2003 cost reports. Therefore, we believe that FY 2002 cost report data provides the best available information for this final rule.

B. Determining the Update Factors for the Budget-Neutrality Calculation

Section 124(a)(1) of the BBRA requires that the IPF PPS be budget neutral. In other words, the amount of total payments under the IPF PPS, including any payment adjustments, must be projected to be equal to the amount of total payments that would have been made if the IPF PPS were not implemented. Therefore, in the proposed rule as well as in this final rule, we have calculated the budget-neutrality factor by setting the total estimated PPS payments to be equal to the total estimated payments that would have been made under the TEFRA methodology had the IPF PPS not been implemented.

In the proposed rule, we based the rate setting calculations and estimated impacts on an April 1, 2004 implementation date. However, in order to create a more efficient process of updates for the various Medicare payment systems, we proposed to establish a July 1 annual update cycle for the IPF PPS. We also indicated we would not update the rates on July 1, 2004 because we believed there would be an insufficient time under the new IPF PPS to generate data that would be useful in updating the IPF PPS. As a result, we calculated the proposed Federal per diem base rate to be budget neutral for the 15-month period April 1, 2004 through June 30, 2005.

In this final rule, we calculated the final Federal per diem base rate to be budget neutral during the implementation period under the IPF PPS. As in the proposed rule, we will use a July 1 update cycle. Similar to the proposed rule, we will not update the IPF PPS during the first year of implementation because we believe there would be an insufficient amount of time under the IPF PPS to generate data useful in updating the system. Thus, the implementation period for the final IPF PPS is the 18-month period January 1, 2005 through June 30, 2006. As a result, we updated the Federal per diem base rate to the midpoint of the January 1, 2005 through June 30, 2006, implementation period (that is, October 1, 2005).

1. The 1997-Based Excluded Hospital with Capital Market Basket

Since FY 2003, the 1997-based excluded hospital with capital market basket has been used to establish the rates-of-increase for excluded hospitals and units paid under TEFRA. As a result, in the proposed rule, we proposed to use the 1997-based excluded hospital capital market basket to update the Federal per diem base rate to the midpoint of the implementation period under the IPF PPS, to establish the labor-related share for applying the wage index (see section V. of this final rule), and to update the Federal per diem base rate after the implementation period (see section V. of this final rule).

In the proposed rule, we explained that we periodically rebase (moving the base year for the structure of costs), and revise (changing data sources, cost categories, or price proxies used) the market basket to reflect more current cost data. We provided a detailed comparison of the 1992-based excluded hospital with capital market basket that had been in effect prior to October 1, 2002 to the rebased and revised 1997-based excluded hospital with capital market basket.

In the proposed rule, we explained that the operating portion of the 1997-based excluded hospital with capital market basket is derived from the 1997-based excluded hospital market basket. The methodology used to develop the operating portion was described in the IPPS final rule published in the Federal Register on August 1, 2002 (67 FR 50042 through 50044). In brief, the operating cost category weights in the 1997-based excluded hospital market basket were determined from the 1997 Medicare cost reports, the 1997 Business Expenditure Survey from the Bureau of the Census and the 1997 Annual Input-Output data from the Bureau of Economic Analysis. As was discussed in the IPPS final rule, we made two methodological revisions in developing the 1997-based excluded hospital market basket: (1) Changing the wage and benefit price proxies to use the Employment Cost Index (ECI) wage and benefit data for hospital workers; and (2) adding a cost category for blood and blood products.

As we indicated in the proposed rule (68 FR 66926), when we add the weight for capital costs to the excluded hospital market basket, the sum of the operating and capital weights must still equal 100.0. Because capital costs account for 8.968 percent of total costs for excluded hospitals in 1997, operating costs must account for 91.032 percent. Each operating cost category weight in the 1997-based excluded hospital market basket was multiplied by 0.91032 to determine its weight in the 1997-based excluded hospital with capital market basket.

The aggregate capital component of the 1997-based excluded hospital market basket (8.968 percent) was determined from the same set of Medicare cost reports used to derive the operating component. The detailed capital cost categories of depreciation, interest, and other capital expenses were also determined using the Medicare cost reports. There are two sets of weights for the capital portion of the market basket. The first set of weights identifies the proportion of capital expenditures attributable to each capital cost category, while the second set represents relative vintage weights for depreciation and interest. The vintage weights identify the proportion of capital expenditures that is attributable to each year over the useful life of capital assets within a cost category (see the IPPS final rule on August 1, 2002 (67 FR 50045 through 50047), for a discussion on how vintage weights are determined).

The cost categories, price proxies, and base-year FY 1997 weights for the excluded hospital with capital market basket are presented in Table 1 below. The vintage weights for the 1997-based excluded hospital with capital market basket are presented in Table 1(A) below.

In the proposed rule (68 FR 66928) we described an analysis we conducted to ensure that the excluded hospital with capital market basket provides a reasonable measure of the price changes facing IPFs. We conducted an analysis of annual percent changes in the market basket when the weights for wages, pharmaceuticals, and capital in IPFs were substituted into the 1997-based excluded hospital with capital market basket. Other cost categories were recalibrated using ratios available from the IPPS market basket. Our analysis found that on average between 1995 and 2002, the excluded hospital with capital market basket increased at nearly the same average annual rate (3.4 percent) as the market basket with IPF weights for wages, pharmaceuticals, and capital (3.5 percent). This difference is less than the 0.25 percentage point criterion that determines whether a forecast error adjustment is warranted under the IPPS update framework.

Based on this analysis, we believe that the excluded hospital with capital market basket is doing an adequate job of reflecting the price changes facing IPFs. For this reason, in this final rule we are adopting the 1997-based excluded hospital with capital market basket to update the Federal per diem base rate to the midpoint of the IPF PPS implementation period, to establish the labor-related share of the Federal per diem base rate, and to update the IPF PPS after the implementation period.

2. Calculating the Budget-Neutrality Adjustment Factor

Many commenters stated that they were concerned that the data used in the proposed rule were not current and did not reflect an accurate view of the services provided to Medicare psychiatric patients. The data sources we used to calculate the proposed budget-neutrality factor were the best data available for IPFs at that time and included FY 1999 cost report data and FY 1999 Medicare claims data from the June 2001 update of the MedPAR files. We updated the data for each IPF to the midpoint of the proposed 15-month implementation period (April 1, 2004 through June 30, 2005) and used the projected market basket update factors for each applicable year. For this final rule, we used FY 2002 data, the best data available.

a. Cost Report Data for January 1, 2005 Through June 30, 2006

In the proposed rule, we proposed to update each IPF's cost to the midpoint of the proposed implementation period April 1, 2004 through June 30, 2005. We explained that to calculate the operating costs, we would use the applicable percentage increases to the TEFRA target amounts for FY 1999 through FY 2002 in accordance with § 413.40(c)(3)(vii) and the full excluded hospital market-basket percentage increase for FY 2003 and later in accordance with § 413.40(c)(3)(viii).

In this final rule, in order to determine each provider's projected operating cost for the IPF PPS implementation period adopted in this final rule, we updated each IPF's per diem cost in FY 2002 to the midpoint of the implementation period January 1, 2005 through June 30, 2006. We used the most recent projection of the full percentage increase in the 1997-based excluded hospital market basket index for FY 2003 and later in accordance with § 413.40(c)(3)(viii).

Comment: A few commenters recommended that CMS project IPF operating and capital costs using the full TEFRA market basket indexes.

Response: We used FY 1999 data in the proposed rule. In order to update the data to the midpoint of the proposed implementation period, we applied the cap imposed by section 4414 of the BBA in accordance with § 413.40(c)(3)(vii). The BBA caps sunset after FY 2002. Since we used the FY 2002 cost reports to project TEFRA costs and payments in this final rule, we used the full excluded hospital market basket indexes to project the costs and payments to the midpoint of the IPF PPS implementation period in accordance with § 413.40(c)(3)(viii).

Since the IPF PPS includes both the operating and capital-related costs, we projected the capital-related cost under the TEFRA system as well. We used the excluded capital market basket to project the capital-related costs under the TEFRA system. Table 2 below summarizes the excluded hospital market basket (without capital) and the excluded capital market basket indexes.

b. Estimate of Total Payments Under the TEFRA Payment System

Consistent with the proposed rule, in this final rule, we estimated payments for inpatient operating and capital costs under the current TEFRA system using the following methodology:

Step 1: IPF's Facility-Specific Target Amount

The facility-specific target amount for an IPF was calculated based on the IPF's allowable inpatient operating cost per discharge for the base period, excluding capital-related, non-physician anesthetist, and graduate medical education costs. We updated the target amount using the rate-of-increase percentages specified in § 413.40(c)(3)(viii).

Step 2: Calculating Each IPF's TEFRA Payments for Inpatient Operating Services

Under the TEFRA system, an IPF's payment amount for inpatient operating services is the lower of—

• The hospital-specific target amount multiplied by the number of Medicare discharges (the ceiling); or
• The hospital's average inpatient operating cost per case multiplied by the number of Medicare discharges.

In addition, under the TEFRA system, payments may include a bonus or relief payment, as follows:

• IPFs whose net inpatient operating costs are lower than or equal to the ceiling would receive the lower payment of—(1) the net inpatient operating costs plus 15 percent of the difference between the inpatient operating costs and the ceiling; or (2) the net inpatient operating costs plus 2 percent of the ceiling.
• IPFs whose net inpatient operating costs are greater than the ceiling, but less than 110 percent of the ceiling, would receive the ceiling payment.
• IPFs whose net inpatient operating costs are greater than 110 percent of the ceiling would receive the ceiling payment plus the lower of—(1) 50 percent of the difference between the 110 percent of the ceiling and the net inpatient operating costs; or (2) 10 percent of the ceiling payment.

Step 3: IPF Payments for Capital-Related Costs

Under the TEFRA system, in accordance with section 1886(g) of the Act, Medicare allowable capital-related costs are paid on a reasonable cost basis. Each IPF's payment for capital-related costs is taken directly from the cost report and updated for inflation using the excluded capital market basket.

Step 4: IPF Total Operating and Capital-Related Costs Under the TEFRA Payment System

Once estimated payments for inpatient operating costs were determined (including bonus and relief payments, as appropriate), we added the TEFRA adjusted operating payments and capital-related cost payments together to determine each IPF's total payments under the TEFRA payment system.

c. Payments Under the IPF PPS Without a Budget-Neutrality Adjustment

Consistent with the proposed rule, in this final rule, we used the 1997-based excluded hospital with capital market basket to trend the FY 2002 base year data to the midpoint of the IPF PPS implementation period and, for the purpose of applying a wage index adjustment, to establish the labor-related portion of the Federal per diem base rate.

In this final rule, by trending the cost using the applicable market basket increase factors, we updated the average per diem cost to the midpoint of the January 1, 2005 through June 30, 2006 implementation period. The updated average cost per day of $724.43 was then used in the payment model to project future payments under the IPF PPS.

The next step is to apply the associated wage index and all applicable patient-level and facility-level adjustments to determine the appropriate IPF PPS payment amount for each stay in the final payment model file.

C. Standardization of the Federal Per Diem Base Rate

We must standardize the IPF PPS payments in order to account for the overall positive effects of the final IPF PPS payment adjustment factors. The proposed standardization factor was calculated to be 17 percent. However, in the proposed rule, we included a 19-percent budget-neutrality adjustment and a 2-percent outlier adjustment, and did not identify the percentage of the overall budget-neutrality adjustment that was attributable to standardization.

As was done in the proposed rule and in this final rule, to standardize the IPF PPS payments, we compared the IPF PPS payment amounts calculated from the psychiatric stays in the FY 2002 MedPAR file to the projected TEFRA payments from the FY 2002 cost report file updated to the midpoint of the IPF PPS implementation period. The standardization factor was calculated by dividing total estimated payments under the TEFRA payment system by estimated payments under the IPF PPS. The standardization factor was calculated to be 0.8367. As a result, the $724.43 Federal per diem base rate was reduced by 16.33 percent.

D. Calculation of the Budget Neutrality Adjustment

As we noted above, in the proposed rule we identified a 19-percent budget-neutrality factor, but did not break it out into separate components. In this final rule, we are identifying each component of the budget neutrality adjustment, that is, the outlier adjustment, stop-loss adjustment, and behavioral offset.

1. Outlier Adjustment

Since the IPF PPS payment amount for each IPF includes applicable outlier amounts, using an approach consistent with the proposed rule, we reduced the standardized Federal per diem base rate to account for aggregate IPF PPS payments estimated to be made as outlier payments. The appropriate outlier amount was determined by comparing the adjusted prospective payment for the entire stay to the computed cost per case. If costs were above the prospective payment plus the adjusted fixed dollar loss threshold, an outlier payment was computed using the applicable risk-sharing percentages, as explained in greater detail in section VI.D.1. of this final rule. The outlier amount was computed for all stays, and the total outlier amount was added to the final IPF PPS payment. The outlier adjustment was calculated to be 2 percent. As a result, the Federal per diem base rate includes a reduction of 2 percent.

2. Stop-Loss Provision Adjustment

As explained in detail in section VI.D.3. of this final rule, we will provide stop-loss payments to ensure that an IPF's total PPS payments are no less than a minimum percentage of their TEFRA payment, had the IPF PPS not been implemented. As with outlier payments, in this final rule, we reduced the standardized Federal per diem base rate by the percentage of aggregate IPF PPS payments estimated to be made for stop-loss payments.

The stop-loss payment amount was determined by comparing aggregate prospective payments that the provider would receive under the IPF PPS to aggregate TEFRA payments that the provider would have otherwise received without implementation of the IPF PPS. If an IPF's aggregate IPF PPS payments are less than 70 percent of its aggregate payments under TEFRA, a stop-loss payment was computed for that IPF. The stop-loss payment amounts were computed for those IPFs that were projected to receive the payments, and the total amount was added to the final IPF PPS payment amount. In our calculation, we needed to include a reduction of 0.39 percent in the standardized Federal per diem base rate to maintain budget neutrality in the final IPF PPS.

We note that the 0.39 percent adjustment due to the stop-loss provision is temporary in nature. This adjustment will be removed after the transition because, as explained in section IV.D.3. of this final rule, the stop-loss provision is applicable only during the transition period.

3. Behavioral Offset

As explained in the proposed rule, we expect that once the IPF PPS is implemented, IPFs may experience usage patterns that are significantly different from those they currently experience. For example, since the IPF PPS is a per diem system, IPFs might have an incentive to keep patients in the facility longer to maximize their use of beds or to receive outlier payments. In addition, the current TEFRA payment system does not depend on coding a principal diagnosis; however, payment will depend on properly coding the principal diagnosis under the IPF PPS. Therefore, we expect that IPFs will have an incentive to comprehensively code for the presence of comorbidities and ultimately the coding practice of IPFs should improve once the IPF PPS is implemented.

As a result of these behavioral changes, Medicare may incur higher payments than assumed in our calculations. These effects were taken into account when we calculated the proposed budget-neutral Federal per diem base rate. Accounting for these effects through an adjustment is commonly known as a behavioral offset.

Based on accepted actuarial practices and consistent with the assumptions made under the Inpatient Rehabilitation Facility PPS, we assumed in determining the behavioral offset, that IPFs would regain 15 percent of potential “losses” and augment payment increases by 5 percent. We applied this actuarial assumption, which is based on our historical experience with new payment systems, to the estimated “losses” and “gains” among the IPFs.

Comment: A few commenters disagreed with CMS's concern that the IPF PPS would provide an incentive for IPFs to increase length of stay. They stated that the incentive to increase length of stay already exists under the current TEFRA payment system. The commenters stated that under TEFRA, the longer the stay, the higher the payment as long as the hospital stays under its TEFRA limit.

Commenters stated that despite this incentive, length of stay has continuously declined over the last decade. One commenter mentioned that IPFs use clinical practice guidelines used by Quality Improvement Organizations, rather than Medicare reimbursement standards, to determine when a patient is ready for discharge.

Several commenters stated that they do not foresee any significant increase in length of stay for psychiatric admissions and recommended that CMS adopt a smaller behavioral offset initially. They suggested that the length of stay could easily be monitored by CMS and adjusted in the future, if necessary.

Response: Since per diem payment systems pay on a per day basis rather than a per discharge basis, there is an incentive to keep patients more days. Therefore, we believe that including a behavioral offset will make our calculations and impact analysis more accurate. We will monitor the extent to which current practice in IPFs changes such as how the average length of stay is affected by implementation of a per diem payment system and may propose adjustments to the behavioral assumptions, accordingly.

In addition to the length of stay, the final IPF PPS payment model depends on the accurate coding of diagnoses for the DRG and comorbidity adjustments. We expect that IPFs will try to code diagnoses for each stay more accurately after the implementation of the IPF PPS in order to receive payment adjustments. This behavior change could result in significantly higher Medicare payments to IPFs than we assumed when we calculated the final Federal per diem base amount.

The behavioral offset for the final IPF PPS was calculated to be 2.66 percent. As a result, we reduced the standardized Federal per diem base rate by 2.66 percent to maintain budget neutrality.

To summarize, the proposed Federal per diem base rate with an outlier adjustment and budget neutrality with a behavioral offset was calculated to be $530. This amount included a 2-percent reduction to account for proposed outlier payments and a 19 percent reduction to account for budget neutrality and the behavioral offset to the Federal per diem base rate otherwise calculated under the methodology as described above. Of that 19-percent reduction, 17 percent is attributable to standardization, and 2 percent is attributable to the behavioral offset (see section V.C. of this final rule for an explanation of standardization).

Using the FY 2002 data for this final rule, the final budget-neutral Federal per diem base rate with an outlier adjustment, a stop loss provision with a behavioral offset is calculated to be $575.95. This amount includes a 16.33-percent reduction from $724.43 to account for standardization to the projected TEFRA per diem payment for the implementation period, a 2-percent reduction to account for outlier payments, a 0.39-percent reduction to account for stop-loss payments and a 2.66-percent reduction to account for the behavioral offset.

VI. Cost Regression Used To Develop Payment Adjustment Factors

In the proposed rule, we provided a detailed description of the data file used for the regression analysis, our trimming methods, and the limitations associated with IPFs reporting routine per diem costs as an average. As a result of the regression analysis, we proposed patient-level payment adjustments for age, DRG assignment based on patients' principal diagnoses, selected comorbidities, and a day of stay adjustment (the variable per diem adjustments) to reflect higher resource use in the early days of an IPF stay. We also proposed facility-level payment adjustments for wage area and rural location, and a teaching status adjustment.

Comment: One commenter stated that the regression models used in the proposed rule may not have appropriately modeled the data. The commenter believes that data entered into the regression model(s) are of a hierarchical nature, namely patients within facilities. Therefore, within a facility they cannot be considered independent observations, a requirement of simple regression models. To account for the fact that patients are nested within hospitals, hierarchical linear models need to be used. This will allow the covariance structure to be modeled. The commenter also believes that this will allow facility level variables to be modeled in the appropriate place. The commenter stated that although this would have to be explored, a model might estimate average facility costs while individual variability attributable to the patients and their covariates would be estimated separately.

Response: There are two parts to our response to this comment. The first part addresses why our data are not well-suited for the use of hierarchical linear models. The second part addresses the potential consequences for the payment adjustment factors of using ordinary least squares to estimate the cost regression instead of a method applicable for hierarchical linear models. We use ordinary least squares in the proposed rule as well as in this final rule.

First, the commenter is correct that, in principle, multi-level or hierarchical linear models would be appropriate for cost data that varied among patients within psychiatric facilities (commonly referred to as within group variation) and among psychiatric facilities (commonly called between group variation). However, in our cost data, each facility assigns the same per diem routine cost to all of its patients. As a result, there is no per diem routine cost variation among patients within the same facility, and, since routine costs are a large proportion of total cost, our measure of routine cost contains relatively little within group variation. In our data, ancillary cost differences are the only source of within group variation in per diem cost. This constraint substantially limits our ability to model patient effects within facilities. We concluded that under these circumstances, we are not able to meaningfully estimate a hierarchical linear model and that the data could be appropriately modeled using ordinary least squares.

Second, there are two potential consequences of using ordinary least squares to estimate the cost regression rather than a statistical method applicable for hierarchical models. According to statistical theory, the first consequence is that the standard errors of the regression coefficients may differ in the 2 cases. These differences could influence the conclusions drawn from tests of statistical inference about the role of the regression's independent variables (for example, patient age and length of stay) in explaining variation in per diem costs. The significance of this problem is that, potentially, we might develop a payment adjustment based on a variable that we believe to be a significant determinant of per diem cost, when we would not have developed a payment adjustment for that variable if we had estimated the cost regression using a statistical technique that would yield more accurate standard errors. To test whether this problem applies to our cost regression, we estimated the regression using a method applicable to hierarchical models.

As noted by the commenter, the advantage of hierarchical linear models is that they allow modeling of the covariance structure. The method we used (the SAS procedure named Proc Mixed) allows the user to select among alternative models of the data's covariance structure. Among the options in Proc Mixed, we used a random effects model with “compound symmetry” as a compromise between the assumptions of ordinary least squares and the completely unstructured case, which imposes no assumptions on the covariance structure. The results of this test were, as predicted by statistical theory, that the standard errors from Proc Mixed often differed from those estimated using ordinary least squares. However, there was no change in the conclusions drawn from statistical inference tests because the variables that were significant using ordinary least squares remained highly significant using Proc Mixed. As a result, both statistical techniques imply that the same variables are important determinants of per diem cost and, hence, potential candidates for payment adjustment factors.

The second potential statistical consequence of using ordinary least squares rather than a hierarchical model method to estimate the cost regression is that the size of the regression coefficients of the independent variables may be different. In turn, differences in regression coefficients will produce differences in sizes of the payment adjustment factors. However, statistical theory does not predict that the ordinary least squares estimates are subject to statistical bias. Furthermore, statistical theory implies that very large sample sizes such as ours will improve the accuracy of ordinary least squares estimates. Therefore, statistical theory does not imply that the regression coefficients estimated using ordinary least squares are necessarily less accurate than those estimated with Proc Mixed or a similar method.

Based on the three considerations just described, we believe that the statistical methods we used in the proposed and final rule enabled us to model the data appropriately. That is, although in principle our data is hierarchical, in practice, it does not contain the full extent of variation at the patient and facility levels that would yield meaningful hierarchical modeling. In addition, our conclusions about which variables are important in explaining cost variation are not affected by our use of ordinary least squares. Finally, statistical theory of hierarchical modeling does not imply that there is necessarily a problem with the size of the regression coefficients obtained from ordinary least squares.

Comment: A commenter stated that CMS estimated a “structural model” rather than a “payment model” by including variables in the regression that were not used as payment adjustors (size and the occupancy rate). The commenter acknowledged that there is some debate about which type of model is most appropriate in constructing payment systems, but expressed the opinion that the “research and policy community” believes that payment models are preferred to structural models.

Response: This commenter is referring to two different approaches in using cost regressions to develop payment adjustments. In the “payment model” approach, the only independent variables included in the cost regression are those variables that are used as payment adjustments. In the “structural model” approach, all variables that are hypothesized to be important determinants of cost are included in the cost regression, whether or not they are going to be used as payment adjustments. Omitting “structural” variables from the cost regression will affect the sizes of the regression coefficients for “payment” variables if the omitted variables are correlated with some or all of the payment variables, which will in turn affect the magnitude of the payment adjustment factors. If omitted structural variables are completely uncorrelated with any of the payment variables, omission of the structural variables from the cost regression will lower the overall explanatory power of the regression, but will not affect the sizes of the regression coefficients for the payment variables. Debate over whether the payment or the structural approach is preferred generally centers on the case when one or more structural variables are positively correlated with one or more payment variables. In this case, the payment approach will result in paying for some of the effects of the omitted structural variable(s) via the payment adjustments of some of the payment variables. That is, the payment adjustment factors for some payment variables will be greater than they would have been had the structural model been used. The structural approach will result in smaller payment adjustment factors for some payment variables because the effects of the omitted structural variables are not reflected in the regression coefficients of those payment variables, but rather are captured by the regression coefficients of the structural variables included in the cost regression.

We believe the commenter is questioning whether CMS included variables in the cost regression that were not used as payment adjustors. The two variables cited in the comment are measures of facility size and occupancy. In fact, in neither the proposed nor the final rule did we include facility size in our cost regression. We followed the payment model approach with respect to the size variable because facility size has never been regarded as an acceptable payment variable in any of our prospective payment systems since it is a variable over which a facility has a substantial degree of control. However, in adopting the payment model approach for the size variable, we are allowing the effects of size to increase payment adjustment factors to the extent that facility size is positively correlated with acceptable payment variables. For example, small facilities that are small because of other factors such as rural location will be compensated for their higher costs due to those factors. Therefore, adopting a structural payment model approach would have adversely penalized small facilities and we recognize that small facilities may be important providers of psychiatric services in many circumstances. In the case of the occupancy rate, we adopted the structural approach and included the variable in the regression. Whether a facility is large or small, we think that it is appropriate to control for variations in the occupancy rate in estimating the effects of the payment variables on per diem cost to avoid compensating facilities for inefficiency associated with underutilized fixed costs.

Comment: A commenter asked whether the age and comorbidity variables identified the same groups of patients, and as a result, whether by including both variables in our regression, we were making the same adjustment twice.

Response: Although the presence of comorbidities is more common among the elderly, the age and comorbidity variables do not identify exactly the same groups of patients. In the proposed rule, the age variable grouped all patients over age 65 in the same category and the comorbidity variables identified 17 different conditions. Comorbidities were present for patients under age 65 as well as those over age 65. Further, since we identified 17 separate comorbid categories, some elderly patients have no comorbidities, others have a single comorbidity, and still others may have multiple comorbidities. Including the age and comorbidity variables in the regression does not measure the same adjustment twice, but rather utilizes the fact that the variables are not perfectly correlated to measure separate effects for age and comorbidities.

Comment: One commenter recommended that CMS compare the relationship between costs per day among the various types of IPFs to the same relationship among types of SNFs.

The commenter stated that hospital-based SNFs have higher per diem costs than freestanding SNFs, but the shorter lengths of stay for hospital-based SNFs result in approximately equal per case costs for freestanding and hospital-based SNFs.

Response: The government-operated psychiatric hospitals have relatively low per diem costs, relatively long lengths of stay, and relatively high per case costs. However, among the other main types of psychiatric facilities (non-profit hospitals, for-profit hospitals, and psychiatric units), there is a direct relationship between per diem and per case costs because lengths of stay are very similar for these types of facilities. Psychiatric units have the highest per diem and per case costs, followed by non-profit hospitals, and last by for-profit hospitals.

Comment: A few commenters suggested that CMS adopt the DRG methodology used under the IPPS instead of utilizing adjustment factors for age, comorbidities, and DRG assignment. The commenters believe that by using this method, the DRGs would be established for cases with and without the presence of comorbidities and for various age categories.

Response: As we discussed in the proposed rule, adopting a patient classification system based on diagnosis alone may not explain the wide variation in resource use among IPF patients. There is no indication that regrouping the psychiatric DRGs as the commenter suggests will explain more of the variation in per diem cost than the methodology we are adopting.

Since the DRGs are also used to pay inpatient psychiatric cases treated outside the distinct part psychiatric unit, we believe that before any basic changes to the DRG structure could be proposed, we would first need to conduct a thorough examination of the potential effects on both the IPPS and the IPF PPS. We have not conducted such an approach because there was insufficient time, and we did not want to delay implementing the IPF PPS.

Comment: Several commenters described a recent study in which the researchers regrouped psychiatric diagnoses and comorbidities and included variables for certain activity of daily living deficits (toileting, transferring, and personal hygiene), patient dangerousness (strong suicide or assaultive tendencies), and patients who undergo electroconvulsive therapy (ECT). The commenters recommended that we adopt the study findings in the final IPF PPS.

Response: Although the commenters did not explicitly identify the study, we believe that they are referring to the CMS funded RTI International^® (trade name of Research Triangle Institute) study of inpatient psychiatric care that was designed to complement the development of the IPF PPS. RTI International^® addressed two major limitations of the administrative claims and cost report data available to CMS for the IPF PPS.

First, the administrative data only captures the uniform routine daily cost assigned to each patient treated in the same facility, so that no variation in routine daily cost can be observed for patients in the same facility, but who have different resource requirements. This artificial reduction in cost variation may impede efforts to accurately identify and measure the effects of certain patient characteristics. Second, the patient characteristics collected on the claims are limited to demographic and diagnostic information and do not include other characteristics that may be more important in explaining resource use.

The RTI International^® study is noteworthy for its success in dealing with these two issues. First, RTI International^® developed a measure of cost per patient day that captured variations in patients' daily resource use both within and across facilities. This task was accomplished by collecting information on the time spent in various activities by patients and facility staff over the course of a 3-shift day for a period of 7 days. After converting the staff time data to daily patient costs, RTI International^® was able to go beyond the potential constraints of administrative data to study differences among patients across days of the stay.

Second, RTI International^® collected a small set of patient characteristics that are not in CMS administrative data. They were able to test the importance of these variables in explaining cost variation. Most important among these factors were certain activities of daily living (toileting, transferring, and personal hygiene) and patient dangerousness (strong suicidal or assaultive tendencies).

Like virtually all studies that collect primary data for a sample population, RTI International^® faced choices about how to obtain the most useful information possible with the limited funds available. RTI International^® collected information for 4,149 Medicare patient days of care delivered to 834 unique Medicare patients in 40 facilities. We believe that RTI's sample is large enough to provide reliable information about the types of patients treated in all psychiatric facilities. However, the sample is small compared to even the typical 10 or 20 percent samples of the MedPAR data, and data collection costs made it uneconomical to sample all types of IPFs. In particular, rural facilities and small and government-operated hospitals could not be represented as robustly as other types of IPF providers.

In addition, although they collected data for 7 days in each facility, it was uneconomical to collect information for entire stays in a large number of cases. Also, in order to limit the costs of data collection, RTI International^® did not collect ancillary service use, but instead relied on claims data for this information.

The findings of the RTI International^® study have played an important role in the development of the IPF PPS in several ways. First, RTI International^® analysis of its daily cost variable supports the use of the administrative data in developing the IPF PPS without being seriously misled about the relative importance of different variables. For example, both sets of analysis found age to be very important in explaining per diem cost variation. Although RTI International^® elected to group diagnoses differently than using DRGs, both analyses supported prior findings that diagnosis plays a limited role in explaining cost variation. RTI International^® also found ECT to be an important cost factor.

However, many other variables commonly thought to affect cost either produced inconsistent results or were found to have a minor effect, once more important factors were taken into account. Among these variables were cognitive impairment, risk of falls, Global Assessment of Function (GAF) score, gender, dual diagnosis, and number of medications.

Second, RTI International^®'s analysis of cost variation by day of stay proved a very useful point of comparison for the variable per diem adjustment factors that we present in this rule. Third, the RTI International^® study provides us with a starting point for future refinements of the IPF PPS. As noted above, RTI International^®'s identification of certain patient characteristics not currently collected in the administrative data is very helpful for starting the process of considering whether we might want to collect some or all of these data items in the future. As a result of this research, we did not choose to adopt adjustment variables for activity of daily living deficits or patient dangerousness. We discuss the adjustment for patients who undergo ECT in section VI.B.6.of this final rule.

Comment: One commenter expressed the opinion that the regression results for the age and diagnosis variables would not be skewed by the inability of CMS routine cost variable to capture cost variations among patients within the same facility. The commenter further predicted that the research conducted by RTI International^® would find that elderly psychiatric patients use fewer resources than younger patients.

Response: The commenter's prediction that RTI International^® would find that elderly psychiatric patients use fewer resources than younger patients was not supported. RTI International^® found, as we did in our cost regressions, that elderly patients are more costly than younger patients. There is no way to directly test the commenter's assertion that our regression results are not affected by the limitations of our routine cost variable. In addition, since the RTI International^® data was able to capture cost variations among patients within the same facility and RTI International^® had results similar to ours about the effects of diagnosis and age on per diem costs, this consistency in results leads us to believe our regression were accurate.

A. Final Regression Analysis

In this final rule, in order to ensure that the IPF PPS would be able to account adequately for each IPF's case-mix, we performed an extensive regression analysis of the relationship between the per diem costs and both patient and facility characteristics to determine those characteristics associated with statistically significant cost differences. For characteristics with statistically significant cost differences, we used the regression coefficients of those variables to determine the size of the corresponding payment adjustments.

The final IPF PPS payment adjustments were derived from a regression analysis of 100 percent of the FY 2002 MedPAR data file because this was the best data available. The MedPAR data file used for the final regression analysis contains 483,038 cases that have a LOS of 1 day or more. We deleted 8,012 (1.66 percent) from this file because cost report or reasonable routine cost data for certain IPFs were not available. In order to include as many IPFs as possible in the regression, we substituted the FY 2001 Medicare cost report data for routine cost and ancillary cost-to-charge ratios (using the FY 2001 Medicare cost report data).

For the remaining 475,026 cases, we used the same method to trim extraordinarily high or low cost values that we used for the per diem rate development file and in the proposed regression analysis (see section V.A. of this final rule).

The trimming criteria eliminated another 3,490 cases, leaving 471,536 cases that were used in the final regression.

We computed a per diem cost for each Medicare inpatient psychiatric stay, including routine operating, ancillary, and capital components using information from the FY 2002 MedPAR file and data from the FY 2002 Medicare cost reports.

To calculate the cost per day for each inpatient psychiatric stay, routine costs were estimated by multiplying the routine cost per day from the IPF's FY 2002 Medicare cost report by the number of Medicare covered days on the FY 2002 MedPAR stay record. Ancillary costs were estimated by multiplying each departmental cost-to-charge ratio by the corresponding ancillary charges on the MedPAR stay record. The total cost per day was calculated by summing routine and ancillary costs for the stay and dividing it by the number of Medicare covered days for each day of the stay.

Since we will pay for emergency department (ED) costs of IPFs with qualifying EDs and IPFs that are part of hospitals with qualifying EDs, as described in section VI.B.5.b. of this final rule, through a specific adjustment to the day one variable per diem adjustment factor, ED costs were excluded from the dependent variable used in the cost regression. ED costs were excluded in order to remove the effects of ED costs from other payment adjustment factors with which ED costs may be correlated. We need to remove the effects on other payment adjustments to avoid overpaying ED costs. Removing ED costs from the regression has no effect on the calculation of the Federal per diem base rate or on budget neutrality because ED costs were not excluded from those calculations.

The log of per diem cost, like most health care cost measures, appears to be normally distributed. Therefore, the natural logarithm of the per diem cost was the dependent variable in the regression analysis. We included variables in the regression to control for psychiatric hospitals that do not bill ancillary costs and for ECT costs that we will pay separately (see the section VI.A. of this final rule).

The per diem cost was adjusted for differences in labor cost across geographic areas using the FY 2005 hospital wage index unadjusted for geographic reclassifications, in order to be consistent with our use of the market basket labor share in applying the wage index adjustment.

We computed a wage adjustment factor for each case by multiplying the Medicare 2005 hospital wage index based on MSA definitions defined by OMB in 1993 for each facility by the labor-related share (.72528) and adding the non-labor share (.27472). We used the 1997-based excluded hospital with capital market basket to determine the labor-related share. The per diem cost for each case was divided by this factor before taking the natural logarithm (that is, a standard mathematical practice accepted by the scientific community). The payment adjustment for the wage index was computed consistently with the wage adjustment factor, which is equivalent to separating the per diem cost into a labor portion and a non-labor portion and adjusting the labor portion by the wage index.

With the exception of the teaching adjustment, the independent variables were specified as one or more categorical variables. Once the regression model was finalized based on the log normal variables, the regression coefficients for these variables were converted to payment adjustment factors by treating each coefficient as an exponent of the base e for natural logarithms, which is approximately equal to 2.718. The payment adjustment factors represent the proportional effect of each variable relative to a reference variable.

B. Patient-Level Adjustments

We proposed adjustments for the DRG assignment of the patient's principal diagnosis, selected comorbidities, and patient age. The proposed rule included a discussion regarding a gender variable, however, we did not propose a gender adjustment.

1. Adjustment for DRG Assignment

In the proposed rule, we proposed adjustment factors for 15 diagnosis-related groups (DRGs). The adjustment factors were expressed relative to the most frequently reported DRG (DRG 430) and were derived from the proposed regression analysis. We did not propose payments under the IPF PPS for all DRGs that contain a psychiatric ICD-9-CM code because for some DRGs, there were too few psychiatric cases to obtain a reliable adjustment factor.

In this final rule, we are providing payment under the IPF PPS for all DRGs that contain a psychiatric ICD-9-CM code. However, as discussed later in this section, we are not providing a DRG adjustment for these cases.

We proposed that IPFs would continue to report diagnoses using the ICD-9-CM coding system. In addition, we specified that current regulations at § 412.27 require that a psychiatric unit admit only those patients who have a principal diagnosis that is listed in the Diagnostic and Statistical Manual of Mental Disorders (DSM) or classified in Chapter Five (“Mental Disorders”) of the ICD-9-CM. We requested public comment on whether we should continue to reference the DSM. The DSM is currently in its fourth edition, text revision (DSM-IV-TR).

We received a significant number of public comments expressing support for the DSM, including several requesting that we permit IPFs to report diagnoses using DSM codes. Many comments asserted that the DSM provides a common language for psychiatrists and other health care professionals and sets forth diagnostic criteria for mental disorders and ways of measuring and reporting severity. Others agreed that the DSM established validity and provides standardized definitions.

Comment: One commenter indicated that Chapter Five of the ICD-9-CM is too limited to be the only diagnostic codes considered and that symptoms that are commonly treated in inpatient psychiatry include DSM codes that are not in the ICD-9-CM. Another commenter suggested that CMS use a combination or subset of diagnostic codes that includes codes that appear in both Chapter Five of the ICD-9-CM and the DSM-IV-TR.

One commenter expressed concern that misalignment between the DSM-IV-TR and the ICD-9-CM codes would cause underpayment of certain cases. The commenter recommended that CMS develop a modifier to the ICD-9-CM code to ensure that DSM codes crosswalk to the most appropriate case mix weight.

Response: We agree that the DSM serves an essential function in the diagnosis and treatment of mental illness. For this reason, we are retaining the reference to the DSM in § 412.27 and updating the reference of the DSM-III-TR to the DSM-IV-TR. As explained in the proposed rule, we acknowledge that the DSM is routinely used by clinical staff to diagnose patients and plan treatment, while the ICD-9-CM coding system is currently used for reporting diagnostic information for payment purposes. However, the Standards for Electronic Transaction final rule published in the Federal Register on August 17, 2000 (65 FR 50312), identifies the ICD-9-CM as the designated code set for reporting diseases, injuries, impairments, other health related problems, their manifestations, and causes of injury, disease, impairment, or other health-related problems. As a result, the DSM codes may not be reported on Medicare claims.

Several commenters included examples of ICD-9-CM codes that do not crosswalk to the DSM-IV-TR, as well as DSM-IV-TR definitions and codes that do not crosswalk to the ICD-9-CM. Preliminary analysis of the codes confirmed the commenters' findings. We considered the possibility of using a modifier to crosswalk certain ICD-9-CM codes to their respective DSM-IV-TR counterpart, but found this method to be too complex and cumbersome for the purposes of billing since each ICD-9-CM code would require a modifier.

More importantly, as we previously explained in section VI of this final rule, we believe it is essential to maintain the same diagnostic coding for IPFs that is used under the IPPS for providing the same psychiatric care. For these reasons, we are not limiting the Chapter Five ICD-9-CM diagnosis codes that may be reported by IPFs under the IPF PPS at this time. We intend to continue our analysis as we implement the IPF PPS to ensure that we identify the appropriate ICD-9-CM codes for coding of patients' principal diagnoses.

We will reconsider these coding issues as we develop the FY 2006 hospital IPPS proposed rule in order to maintain consistent coding rules for all psychiatric cases.

Comment: One commenter asked why CMS used the existing DRGs, rather than developing new groupings for the DRG classification system based on current data. This commenter also asked whether the DRGs would change if they were designed to explain differences in cost per day, rather than cost per case.

Response: We did not attempt to modify the DRG classifications. (see section VI of this final rule for a detailed explanation). Our rationale for proposing to use the existing DRGs to group IPF PPS cases is that the DRGs are currently used to pay inpatient psychiatric cases under the hospital IPPS.

Instead of explicitly attempting to adapt the DRGs to a per diem system by changing the DRG definitions, we analyzed whether there was empirical support for using the existing DRGs. Specifically, we tested whether the DRGs contributed explanatory power to the explanation of differences in per diem costs. Although previous research indicates that diagnosis plays a limited role in explaining cost variation for psychiatric care, existing DRGs provide an acceptable degree of explanatory power.

Additional research will be needed to determine how the DRG classification system or payment weights under the IPPS would change if they were redesigned to measure cost per day.

Comment: One commenter requested that CMS delay implementation of the IPF PPS until the ICD-10-CM is adopted for Medicare billing purposes.

Response: The National Committee on Vital and Health Statistics (NCVHS) has recommended that HHS, under its HIPAA responsibilities, prepare a proposed regulation to require that the ICD-10-CM be adopted as the HIPAA standard code set to replace the ICD-9-CM. HHS is assessing the NCVHS recommendation. We do not believe it is appropriate to tie implementation of the IPF PPS to another initiative that has not been developed.

Comment: Many commenters requested that CMS adopt the clinical structure of the DSM (the DSM diagnostic categories) to classify IPF cases rather than the DRG classification system. A few commenters suggested that CMS use a modified version of the DSM diagnostic categories.

Response: We tested various groupings of diagnoses. Our data analysis indicated that regrouping the ICD-9-CM codes into the DSM diagnostic categories or other similar categories raised the explanatory power of the payment model by less than one-half of one percent. Thus, the DRGs and the DSM diagnostic categories explain the same amount of per diem cost differences. Moreover, the research conducted by THEORI, a research component of the Greater New York Hospital Association, confirmed our results. Therefore, since we were unable to detect a measurable difference in the explanatory power of the DSM and DRGs with respect to the grouping of the ICD-9-CM codes, we are finalizing the DRG approach.

As mentioned earlier, we are concerned about establishing a different classification scheme for IPF PPS than is used for psychiatric discharges under IPPS. We are also concerned about the fiscal burden associated with establishing a separate classification system for the IPF PPS.

As a result, this final rule includes adjustment factors for the DRG assigned to the claim. The coefficient values and adjustment factors were derived from the final regression analysis. The adjustment factors are expressed relative to DRG 430. See Table 3 at the end of this section and Addendum A.

Comment: Commenters overwhelmingly disagreed with the proposed policy to only pay for a limited selection of psychiatric diagnoses under the IPF PPS. The commenters indicated that all DRGs containing psychiatric codes should be recognized in the final IPF PPS. Other commenters recommended that CMS add a new DRG “Other Psychiatric Diagnosis” to include the ICD-9-CM diagnosis codes that are excluded when crosswalked to the DSM-IV-TR.

Response: As we explained earlier in this section, we agree that the IPF PPS should recognize all ICD-9-CM psychiatric codes regardless of their DRG assignment. Therefore, we will provide the Federal per diem base rate payment under the IPF PPS for claims with a principal diagnosis included in Chapter Five of the ICD-9-CM or the DSM-IV-TR. However, only those claims with diagnoses that group to a psychiatric DRG will receive a DRG adjustment. Although the IPF will not receive a DRG adjustment for a principal diagnosis not found in one of our identified 15 psychiatric DRGs, the IPF will still receive the Federal per diem base rate and all other applicable adjustments. Since there are only a few non-psychiatric DRGs that contain one or two rarely used psychiatric codes, whose frequencies were so low that we were unable to calculate an adjustment, we believe this is an equitable way to pay for these cases.

We have not established a new DRG for these psychiatric ICD-9-CM codes that are assigned to non-psychiatric DRGs. Rather, we plan to monitor the data from these other codes and, if indicated through data analysis, may consider proposing revisions to this policy in the future.

Comment: One commenter requested that we revise the DRG adjustment factor to 1.00 for DRG 433 Alchohol/Drug Abuse or Dependence, Left Against Medical Advise. The commenter indicated that the 0.88 proposed adjustment factor would be insufficient to cover the extensive diagnostic procedures, complex treatment, and monitoring these patients often needed.

The commenter also indicated that since the total reimbursement for these patients is directly related to their length of stay, there should be no penalty attached to the DRG assignment.

Response: Our analysis did not indicate or reflect that a 1.00 adjustment was appropriate. The analysis, a cost regression analysis that used hospital claims data resulted in 0.88 adjustment factor for DRG 433 Alchohol/Drug Abuse or Dependence, Left Against Medical Advise. Unlike IPPS that uses DRG weights as the basis for payment, the IPF PPS payment is based on a Federal per diem base rate and numerous additional payment adjustments. In addition to DRG adjustments, the IPF PPS payment includes payment adjusters to accommodate differing lengths of stays (the variable per diem adjustment) that is intended to account for the increased cost in the early days of an inpatient stay. For more information on the variable per diem adjustments, see section VI.B.5 of this preamble.

Comment: A commenter asked for clarification as to the classification of substance abuse as a psychiatric condition.

Response: Substance abuse is not only included in Chapter Five (Mental Disorders) of the ICD-9-CM and defined in the DSM-IV-TR (Substance-Related Disorders) but is also included in the Psychiatric Boards, which physicians take to become Board Certified in the field of psychiatry. However, substance abuse is rarely the primary diagnosis for inpatient psychiatric treatment, and in those rare cases, there are generally mitigating factors to justify why the patient cannot be treated in an outpatient setting. To be covered as an inpatient hospital service, it must meet the criteria for being medically necessary.

2. Comorbidities

In the proposed rule, we proposed 17 comorbidity categories and identified specific ICD-9-CM codes that would generate a payment adjustment. Our intent was to identify conditions that would require comparatively more costly treatment during an IPF stay than other comorbid conditions.

We specifically solicited comments on other conditions that may be expected to increase the per diem cost of care in IPFs. In response, we received a number of comments regarding our proposed comobidity adjustments. A number of commenters expressed support that the proposed IPF PPS recognized the increased cost associated with comorbid medical conditions. Others identified what they believe to be flaws in the analysis used to develop the proposed comorbidity adjustments. A majority of the commenters indicated that hospitals design specialized programs with highly trained staff to treat Medicare beneficiaries who are disabled or geriatric psychiatric patients. The commenters stated that the proposed comorbidity adjustments are inadequate to capture these coexisting medical and psychiatric conditions requiring treatment during a hospital stay.

We also received comments offering suggestions on how we could improve the comorbidity list. The suggestions ranged from a request for addition of a single ICD-9-CM code to a request for comorbidity categories to account for every ICD-9-CM and DSM-IV-TR diagnosis.

Comment: Commenters expressed concern that payment for treating complex cases would decrease because the proposed comorbidity list does not include the conditions seen in their patient populations. Several comments stated that most psychiatric patients are treated for multiple common conditions and illnesses (for example, heart conditions, stroke), none of which would trigger a payment adjustment under the proposed IPF PPS.

Other commenters stated that the proposed comorbidity list includes mostly acute medical conditions that would require transfer to an acute care hospital. One commenter indicated that the adjustment proposed for renal failure should be much higher. Many commenters stated that the range of diagnostic codes proposed for adjustment often did not include all the ICD-9-CM codes within a diagnostic category. For example, the list of codes under diabetes did not include all the diabetes codes.

Response: We have reconsidered our approach to the comorbidity adjustments and have revised the comorbidity list. We analyzed the FY 2002 data to determine the prevalence of the diagnoses suggested most often in the public comments (for example, hypertension, chronic constructive pulmonary disease, and urinary tract infection). In an attempt to address the commenters concerns, we had CMS staff physicians and FI Medical Directors who are psychiatrists review the list of proposed comorbidities and cost and frequency data on all ICD-9-CM diagnoses codes that had been submitted on the FY 2002 claims.

We explained to the CMS staff physicians and FI Medical Directors that the data used in calculating the Federal per diem base rate for both the proposed rule and the final rule included all the costs for comorbid diagnoses submitted in the FY 2002 claims. Therefore, the cost for providing patient care (for example, medications, and routine nursing care required for the common conditions seen in the psychiatric population and recommended for comorbidity adjustment by the commenters (that is, heart conditions or strokes) are included already in the Federal per diem base rate and a comorbidity adjustment for their presence was unnecessary.

One significant issue raised by the CMS physician and FI Medical Director panel was the extent of medical treatment permitted in a psychiatric unit. In the secure environment of a psychiatric unit, common treatments such as IV antibiotics therapy would not be permitted as they could compromise patient safety. The prohibition of items that present a potential risk as a mechanism to inflict injury on oneself or others is strictly enforced. Thus, for many medical treatments for the more complex and costly comorbid, medical, or surgical conditions the psychiatric patient would be required to be moved to a medical floor for treatment with one-on-one staff observation. Consequently, since the patient would no longer be a patient of the IPF, it would be unnecessary to give the IPF an adjustment for such a case.

The intent of the comorbidity adjustments is to provide additional payments for a concurrent medical or psychiatric condition that is expensive to treat. The physicians determined that the high cost of certain diagnoses is related to the cost of the therapy to treat the diagnoses. For example, the cost to treat a patient with a malignant neoplasm is related primarily to the cost of the therapy to treat the tumor, whether it is chemotherapy or radiation therapy, or both. As a result, we have added two ICD-9-CM V codes, one for chemotherapy (V58.0) and for one radiation treatment (V58.1). We are also requiring that, in order to receive the comorbidity adjustment for malignant neoplasm, IPFs will need to code the ICD-9-CM code for the specific malignant neoplasm from the ICD-9-CM chapter 2 codes (140-239) and one of the two ICD-9-CM procedures codes (chemotherapy ((V58.0)) or radiation treatment ((V58.1)) to indicate the treatment modality the patient received.

Based on the clinical expertise of the CMS physicians and FI Medical Directors, we made numerous changes to the list of ICD-9-CM codes eligible for a comorbidity adjustment. These changes include adding one new category entitled, “Developmental Disabilities,” deleting the “HIV” category and moving it into the “Infectious Diseases” category, and changing the titles of two categories from “Malignant Neoplasms” to “Oncology Treatments” and for “Atherosclerosis of extremity with Gangrene” to “Gangrene.”

In response to comments requesting adjustment for Developmental Disabilities and the results of the regression analysis on the FY 2002 data, the higher cost of caring for patients with developmental disabilities indicated a comorbidity adjustment of 1.04 was appropriate. The regression analysis of FY 2002 data would have provided the same adjustment for the “HIV” category as for the “Infectious Disease” category. Therefore, we merged the two categories under the “Infectious Disease” category with an adjustment factor of 1.07. The “Malignant Neoplasm” category was modified to “Onocology Treatments” since the CMS staff physicians and FI Medical Directors believed the higher cost was related to the treatment of the neoplasms rather than the presence of the tumor. We are also requiring that the treatment code be included on the claim form to receive the 1.07 comorbidity adjustment. The last category change was in the title of “Atheroscleosis of Extremity with Gangrene to “Gangrene” to account for the higher cost of a patient with gangrene regardless of the cause.

The design of the IPF PPS with Federal per diem base rate, together with the numerous available adjustments, outlier policy, and stop loss policy during the 3-year transition should prevent the facility from being disadvantaged by decrease in payment for their more complex patients.

We are providing below a table that compares the proposed comorbidity categories to the categories we are adopting in this final rule.

Comment: Several commenters suggested that CMS include all psychiatric and non-psychiatric diagnoses submitted on the claim, whether they are designated as the primary or secondary.

Response: Billing instructions require hospitals to enter the ICD-9-CM code for the patient's principal diagnosis. The code must be the full ICD-9-CM diagnosis code, including all five digits when applicable. The principal diagnosis is the condition established after study to be chiefly responsible for this admission. Even though another diagnosis may be more severe than the principal diagnosis, the hospital enters the principal diagnosis. Entering any other diagnosis as principal on the claim form may result in incorrect DRG assignment and cause the hospital to be incorrectly paid. The hospital is also instructed to enter the full ICD-9-CM codes for up to 8 additional conditions if they co-existed at the time of admission or develope subsequently, and which had an effect upon the treatment or the length of stay. These codes may not duplicate the principal diagnosis.

The regression analysis established the DRG adjustment factors based on the principal diagnoses reported by hospitals and the comorbidity category adjustments based on the all the diagnoses reported by hospitals as other diagnoses. The principal diagnoses were used to establish the DRG adjustment and were not accounted for in establishing the comorbidity category adjustments, except where ICD-9-CM “code first” instructions apply. A description of the “code first” instructions appears in the next section of this final rule.

Comment: Several commenters indicated that the comorbidity adjustment factors did not take into account the extensive workup their patients require, such as the need for additional ancillary services (for example, specific medical or neurological examinations, specialized laboratory and radiological tests, supplies, medications, and consultations). In many instances, the commenter stated that these additional services are needed to identify the numerous physical conditions that exacerbate or first present as psychiatric symptoms.

Response: The adjustment factors for the proposed comorbidity categories were derived from the proposed regression analysis. Similarly, the final adjustment factors for the final comorbidity categories were derived from the final regression analysis. With regard to the additional ancillary services the commenters' patients require to establish their principal diagnoses, the variable per diem adjustments discussed in section VI.B. 5. of this final rule are intended to account for higher per diem costs early in an inpatient stay.

Comment: Commenters expressed concern that the comorbidity policy does not account for the costs associated with social issues (for example, poverty, lack of housing, poor nutrition, lack of primary medical care, and the cost of involuntary commitments and guardianship hearings). The commenters also expressed concern that the comorbidity policy does not account for the costs of patients with hearing, sight, and mobility disabilities or when English is not the patient's primary language.

Response: Most of the social issues identified by the commenters are not captured in the FY 2002 IPF claims data. As a result, we are not able to determine whether the psychiatric hospitalizations of patients with various social issues are more costly on a per diem basis than other psychiatric patients. Because we lack data that indicates IPFs that treat patients with various social issues are more costly on a per diem basis, we are not providing an adjustment in these cases.

We note that codes are currently available that describe some of the social issues that impact care delivery and management. For example, there are V codes to indicate that the patient has problems with sight (V41.0), problems with hearing (V41.2), or lack of housing (V60.0). Even though we have codes for problems with sight, hearing, or lack of housing, we had too few cases to be able to extrapolate any valuable empirical data that the presence of these codes correlated to higher per diem costs. We encourage IPFs to code all relevant diagnoses that impact the resources associated with their patient population for future analysis.

We note that one of the fields on the claim form indicates if patients were referred to the IPF by law enforcement or if the commitment were court ordered (FL 20 item 8, court/law enforcement). As a result, we were able to analyze the impact on per diem cost. The results of our analysis are included in section VI of this rule with other patient variables considered.

Comment: One commenter stated that diagnostic data alone may not be descriptive enough to supply the information CMS is seeking regarding comorbidities.

Response: Section 124 of the BBRA provides authority for CMS to require IPFs to submit additional data. We are not mandating new reporting requirements at this time, however, we may establish new reporting requirements based on results of the research underway to refine the IPF PPS.

Comment: One commenter asked how the comorbidity adjustment would be applied if a patient has multiple diagnoses within the same comorbidity category.

Response: IPFs may only receive one adjustment factor for each comorbidity category. However, if a patient has multiple diagnoses in several categories, the adjustment factors for each applicable category are multiplied by the Federal per diem base rate. The following is an example illustrating how payment would be made under the IPF PPS for a patient with multiple comorbidities

.

Example:

A 68 year old Female Caucasian presents at a qualified ED and is subsequently admitted to a non-teaching inpatient psychiatric facility within the “I'll Feel Better Hospital” in rural Smalltown, North Dakota. The ED is determined to be full-service and the patient had not been discharged from an IPPS stay. The patient had a primary diagnosis of Neurotic Depression (IDC-9-CM code 3004) DRG 426 Depressive Neuroses, and comorbid conditions of Obstructive Chronic Bronchitis without exacerbation 491.20, and mechanical complication of Tracheostomy ICD-9-CM code (ICD-9-CM code 519.02), Diabetes with ophthalmic manifestations (ICD-9-CM code 250.53), and Diabetes with peripheral circulatory manifestations (ICD-9-CM code 250.73). The patient length of stay was 10 days. In addition, the patient did not receive ECT during her inpatient stay.

Calculate Total Wage Adjusted Rate:

Step 1: Multiply the Wage Index Factor (for North Dakota) by the Labor Portion of the Federal base rate to get the Adjusted Labor Portion of the Federal per diem base rate = (0.7743 x 417.73 = $323.45).

Step 2: Add the Adjusted Labor Portion of the Federal Base Rate to the Non-Labor Portion of the Federal per diem base rate to get the Total Wage Adjusted Rate = (323.45 + 158.22 = $481.67).

Apply Facility- and Patient-Level Adjusters

Step 1: Using the information in Addendum A, determine which facility- and patient-level adjustment factors are applicable.

1. Teaching Adjustment: None.

2. Rural Adjustment: North Dakota—1.17.

3. COLA: None.

4. DRG Adjustment: DRG 426—Depressive Neuroses—0.99.

5. Age Adjustment: Age 68—1.10.

6. Comorbidity (All comorbidity codes are cited as presented in the ICD-9-CM text)

Comorbidity 491.20—Obstructive Chronic Bronchitis without exacerbation—None.

Comorbidity 519.02: Mechanical complication of Tracheostomy—1.06.

Comorbidity 250.53: Diabetes with ophthalmic—manifestations (Use additional code to identify manifestation as 362.02)—1.05.

Proliferative Diabetic Retinopathy [not allowed as principal Dx-“CODE FIRST” underlying disease as DIABETES 250.5) and Comorbidity—250.73—Diabetes with peripheral Circulatory—None 2nd in Category manifestations, (Use additional code to identify manifestation as 443.81—Diabetic Peripheral angiopathy [not allowed as principal Dx-“CODE FIRST” underlying disease as DIABETES MELLITUS 250.7).

7. ECT Treatments—None.

Step 2. Multiply the applicable adjustment factors to determine the PPS Adjustment Factor. = (1.17 x 0.99 x 1.10 x 1.06 x 1.05 = 1.4181).

Step 3. Calculate the Adjusted Per Diem.

Multiply the Total Wage Adjusted Rate by the PPS Adjustment Factor.

= ($481.67 x 1.4181 = 683.06).

Calculate the variable per diem adjustment.

Step 1. Determine the number of days in the stay.

Length of Stay: 10 days and the facility has a qualifying ED.

Day 1—1.31

Day 2—1.12

Day 3—1.08

Day 4—1.05

Day 5—1.04

Day 6—1.02

Day 7—1.01

Day 8—1.01

Day 9—1.00

Day 10—1.00

Step 2. Multiply the Variable Per Diem Adjustment Factors by the Total Wage and PPS-Adjusted Per Diem for each day of the stay to get the Total Variable Per Diem Amounts for each day of the stay. (See multiplication in step 3 below.)

Step 3. Add the Adjusted Variable Per Diem Amounts to get the Total Inpatient Psychiatric Facility PPS Payment.

Day 1 (adjustment factor 1.31) × 683.06 = $894.81

Day 2 (adjustment factor 1.12) × 683.06 = $765.03

Day 3 (adjustment factor 1.08) × 683.06 = $737.70

Day 4 (adjustment factor 1.05) × 683.06 = $717.21

Day 5 (adjustment factor 1.04) × 683.06 = $710.38

Day 6 (adjustment factor 1.02) × 683.06 = $696.72

Day 7 (adjustment factor 1.01) × 683.06 = $689.89

Day 8 (adjustment factor 1.01) × 683.06 = $689.89

Day 9 (adjustment factor 1.00) × 683.06 = $683.06

Day 10 (adjustment factor 1.00) × 683.06 = $683.06

Federal per diem payment amount $7,267.75

Comment: A commenter asked if the comorbidity adjustments would be applied to each day of the stay regardless of the patient's length of stay. For example, poisoning and arteriosclerosis of the extremity with gangrene may have higher cost only for the early days of a stay.

Response: The comorbidity adjustments are applied to each day of the stay. In estimating the cost impact of the comorbidity conditions, our dependent variable reflects the average cost per day over the entire stay. A significant effect on this cost variable for a comorbidity condition means that the average cost per day was higher for cases with the specific condition. Therefore, it is appropriate to apply the estimated effect to each day of the stay.

We would be especially concerned if data analysis began to show longer lengths of stay for DRG 424 stays or significantly more DRG 424 stays, with DRG 424 being the surgical DRG. We intend to monitor for changes in length of stay and the distribution of IPF cases across DRGs to ensure that the decision to pay all applicable adjustments throughout the stay does not lead to inappropriate increases in the length of stay or frequency of those cases.

Comment: One commenter indicated that the comorbidity policy does not distinguish between dormant serious medical conditions and labor-intensive procedures requiring additional behavioral and medical treatments during the IPF stay. Another commenter stated that when a non-psychiatric diagnosis exists in addition to a psychiatric diagnosis, the ICD-9-CM code for the non-psychiatric diagnosis should also be reported on the claim.

Response: In § 412.402 definitions, we proposed the following definition of comorbidity: “Comorbidity means all specific patient conditions that are secondary to the patient's primary diagnosis and that coexist at the time of admission, develop subsequently, or affect the treatment received or the length of stay or both. Diagnoses that relate to an earlier episode of care that have no bearing on the current hospital stay are excluded.” A serious medical condition that does not require treatment during the hospital stay must not be reported as a secondary or tertiary diagnosis and will not qualify for a comorbidity adjustment. We are retaining the proposed comorbidity definition in this final rule.

Comment: One commenter recommended that we provide an adjustment to reflect the increased staffing, greater frequency of comorbid conditions, and longer length of stay for developmentally disabled patients.

Response: We analyzed the frequency and costs in the FY 2002 claims data associated with developmentally disabled patients. We identified relevant claims by the presence of an ICD-9-CM code in the 317 through 319 range entered as a diagnosis in addition to a psychiatric principal diagnosis. We found that per diem costs associated with inpatient psychiatric stays of developmentally disabled mentally ill patients, are approximately 4 percent higher than stays for other patients. As a result of this analysis, we are establishing a new comorbidity category to reflect the higher per diem costs of developmentally disabled patients. The final IPF PPS comorbidity categories and adjustment factors are presented in the table below and Addendum A.

3. Other Coding Issues

We received several comments related to discrepancies with established coding conventions.

Comment: One commenter requested that CMS specify that hospitals must follow the ICD-9-CM Official Guidelines for Coding and Reporting and the Coding Clinic for ICD-9-CM. In addition, the commenter advocated the use of certified coding professionals to assign and validate codes and assist in the development of hospital coding policy.

Response: We agree with the commenter about the value of certified coding professionals. The ICD-9-CM Official Guidelines for Coding and Reporting was developed and approved by the Cooperating Parties for ICD-9-CM: The American Hospital Association, the American Health Information Management Association, the Centers for Medicare & Medicaid Services (formerly the Health Care Financing Administration or HCFA) and the National Center for Health Statistics to be used as a companion document to the official version of the ICD-9-CM as published by the Department of Health and Human Services and the Coding Clinic for ICD-9-CM, published by the American Hospital Association. In addition, this decision is consistent with the Standards for Electronic Transaction final rule (65 FR 50312). The ICD-9-CM Official Guidelines for Coding and Reporting can be found at www.cdc.gov/​nchs/​data/​ics9/​icdguide.pdf.

Comment: Several commenters requested that CMS provide detailed information about medical necessity requirements to support an IPF stay. The commenters expressed concern that IPFs are not experienced with medical review and the need to document medical necessity to support the stay. The commenters believe that in the absence of clear national standards for determining medical necessity, IPFs will be subject to various local coverage decisions promulgated by FIs.

Other commenters were concerned about the potential of differential access to inpatient psychiatric care depending on the geographic location of the IPF and how each FI interprets medical necessity. These commenters suggested that CMS incorporate safeguards against clinically unrealistic, inefficient, or inappropriate medical review practices by FIs. The commenters recommended that CMS include a mechanism for impartial appeal of FI decisions to ensure appropriate payment of IPF claims.

Response: Inpatient psychiatric services are intended for patients that require more intense services than can be provided in an outpatient setting. As a result, the patients admitted to an IPF must require intensive, comprehensive, multimodal treatment including 24 hours per day of medical supervision and coordination because of the mental disorder. The need for 24 hours of supervision may be due to the need for patient safety, psychiatric diagnostic evaluation, potential severe side effects of psychotropic medication associated with medical or psychiatric comorbidities, or evaluation of behaviors consistent with an acute psychiatric disorder for which a medical cause has not been ruled out.

The acute psychiatric condition being evaluated or treated by inpatient psychiatric hospitalization must require active treatment, including a combination of services (for example, intensive nursing and medical interventions, psychotherapy, occupational and patient education). Patients must require inpatient psychiatric hospitalization services at levels of intensity and frequency exceeding what may be rendered in an outpatient setting including partial hospitalization programs.

If a provider receives a medical necessity denial, they have the right to appeal the FI's determination that the inpatient hospital services were not reasonable and necessary. A request for reconsideration must be in writing and filed with the FI. The provider should contact their FI for additional information on the appeal process. The prescribed form to request an FI reconsideration “MCS-2649, Request for Reconsideration of Part A Health Insurance Benefits” is located on the CMS web site at www.cms.hhs.gov/​forms.

Comment: Several commenters indicated that the proposed rule included coding policies that were inconsistent with the ICD-9-CM Official Guidelines for Coding and Reporting with respect to the designation of primary and secondary diagnoses (the “code first” policy).

Response: In the proposed rule, we inadvertently failed to include the ICD-9-CM instructions pertaining to the code first diagnosis codes. The introduction of the ICD-9-CM text includes “Instructional Notations” in which “code first” underlying disease is explained. This instruction is for codes that are not intended to be used as a principal diagnosis or for those codes that are not to be sequenced before the underlying disease. The note requires that the underlying disease (etiology) be coded first (identified as the principal and diagnosis) with the code the note is applied to being coded second. This note appears only in the Tabular List (Volume 1).

The ICD-9-CM Official Guidelines for Coding and Reporting includes the following instructional guidance regarding the code first policy:

“(1) The guidelines identify codes that have both an underlying etiology and multiple body system manifestations due to the underlying etiology. The coding convention requires the underlying condition be sequenced first followed by the manifestation. Whenever a combination exists, there is a “use additional code” note at the etiology code, and a “code first” note at the manifestation code. These instructional notes indicate the proper sequencing order of the codes, that is, etiology followed by manifestation.

(2) “Code first” notes are also under certain codes that are not specifically manifestation codes but may be due to an underlying cause. When a “code first” note is present and an underlying condition is present, the underlying condition should be sequenced first.

(3) Code, if applicable any causal condition first, notes indicate that this code may be assigned as a principal diagnosis when the causal condition is unknown or not applicable. If a causal condition is known, then the code for that condition should be sequenced as the principal or first-listed diagnosis.

(4) Multiple codes may be needed for late effects, complications and obstetrics to more fully describe a condition. See the specific guidelines for these conditions for further instruction.”

For example, diagnosis code 294.1 Dementia in Conditions Classified Elsewhere is designated as a code first diagnosis and appears in the ICD-9-CM as follows:

294.1 Dementia in Conditions Classified Elsewhere

Code first any underlying physical condition, as:

Dementia in:

Alzheimer's disease (331.0)

Cerebral lipidosis (330.1)

Dementia with Lewy bodies (33.82)

Dementia with Parkinsonism (331.81)

Epilepsy (345.0-345.9)

Frontal dementia (331.19)

Frontotemporal dementia (331.19)

General paresis [syphilis] (094.1)

Hepatolenticular degeneration (275.1)

Huntington's chorea (333.4)

Jacob-Creutzfeldt disease (046.1)

Multiple sclerosis (340)

Pick's disease of the brain (331.11)

Polyarteritis nodosa (446.0)

Syphilis (094.1)

In accordance with the ICD-9-CM Official Guidelines for Coding and Reporting, when a primary (psychiatric) diagnosis code has a “code first” note, the provider would follow the instructions in the ICD-9-CM text. For example, 294.1, Dementia in conditions classified elsewhere states “code first any underlying physical condition as:” the provider would then code the appropriate physical condition, for example, 333.4 Huntington's chorea as the primary diagnosis and 294.1 as the secondary diagnosis. The submitted claim goes through the CMS processing system that will identify the primary diagnosis code as non-psychiatric and search the secondary codes for a psychiatric code to assign a DRG code for adjustment. The system will continue to search the secondary codes for those that are appropriate for comorbidity adjustment.

A list of ICD-9-CM codes identified as code first is provided in Addendum C.

Comment: A commenter questioned whether IPFs would be required to report ICD-9-CM procedure codes.

Response: IPFs will be required to report those ICD-9-CM codes indicated in the billing instructions. As mentioned above, the only unique coding will be for oncology treatment which requires the ICD for the specific neoplasm and the appropriate treatment V code V580 chemotherapy or V581 radiation. In addition, as discussed in section VI.B.5.C. of this final rule, we are providing additional payments for patients who undergo ECT treatments. In order to receive the additional payments, IPFs will have to report the ICD-9-CM procedure code for ECT (code 90870) and indicate the number of ECT treatments the patient received during the IPF stay. We encourage IPFs to provide as much information on the claim form to describe the services furnished to validate the principal diagnosis for payment purposes.

Comment: One commenter asked if delirium is considered a primary, secondary, or medical condition. The commenter also asked if delirium should be considered an adjustment disorder.

Response: Coding decisions are based on how the physician describes the diagnosis. The physician needs to indicate the type or cause of the delirium, which will determine whether the delirium is psychiatric diagnosis, a psychiatric secondary diagnosis (comorbidity), or a medical comorbid condition. According to the ICD-9-CM, delirium is listed as caused by medical conditions, substance or alcohol abuses, or with psychosis. Delirium is primarily located in the 290 series of ICD codes. If the physician indicates that the patient's diagnosis is “delirium, delirious” the ICD-9-CM index would refer to ICD-9-CM code 780.09—Alteration in consciouusness—Other. However, if the physician specifies that the delirium is acute, then the ICD-9-CM code is 293.0—Delirium Due to Condition Classified Elsewhere, and if the Delirium is caused by alcohol abuse, the ICD-9-CM code is 291.0—Alcohol withdrawal delirium. We recommend that the commenter review the ICD-9-CM index under the term delirium (to determine the different types of diagnosis).

We are not responsible for the determination of clinical definition and criteria. To establish how a condition is defined or identified, providers should review a text of psychiatric diagnoses. We are providing the definition for delirium and adjustment reaction or disorder as defined in the ICD-9-CM (2004) for the convenience of the reader.

Delirium is defined as “Transient organic psychotic condition with a short course in which there is a rapidly developing onset of disorganization of higher mental processes manifested by some degree of impairment of information processing, impaired or abnormal attention, perception, memory, and thinking. Clouded consciousness, confusion, disorientation, delusions, illusions, and often vivid hallucination predominate in the clinical picture.”

Adjustment reaction or disorder is defined as “Mild or transient disorders lasting longer than acute stress reactions which occur in individuals of any age without any apparent preexisting mental disorder. Such disorders are often relatively circumscribed or situation-specific, are generally reversible, and usually last only a few months. They are usually closely related in time and in content to stresses such as bereavement, migration, or other experiences. Reactions to major stress that last longer than a few days are also included. In children, such disorders are associated with no significant distortion of development.”

In review of the DSM diagnostic criteria, delirium is not included in the “Adjustment Disorder” category. Based on the ICD-9-CM definition and the DSM diagnostic criteria, we would not expect delirium to be identified as an adjustment disorder.

Comment: One commenter asked how to code multiple addictions, for example, drug and alcohol, or two drug diagnoses.

Response: We encourage IPFs to code all diagnoses requiring active treatment during the IPF stay. The ICD-9-CM index entry for addiction provides several sub-terms to direct the coder to the most appropriate ICD-9-CM code. The ICD-9-CM code for alcohol dependence is 303.9. However, the ICD-9-CM indicates under code 303.9 that a fifth digit is required based on whether the physician inidicates that the dependence is continuous, episodic, in remission, or there is no information, that is, unspecified.

Separate codes are listed for drug addiction. The index refers coders to “see dependence”. Under dependence, there are a variety of codes depending upon the specific addiction. The coder would enter as many codes as required to cover all the patient's dependencies (drug and alcohol). However, as noted above, only one comorbidity adjustment per comorbidity category will be paid under the IPF PPS.

Comment: Several commenters requested clarification of specific ICD-9-CM codes they suspected were erroneous.

Response: We agree with the commenters and acknowledge that we made the following typographical errors in the proposed rule:

• In Table 3 (68 FR 66931), in the Infectious Disease category, the correct range of codes is 07950 through 07959.
• In table 7 (68 FR 66941), the correct adjustment for Diabetes is 1.10 and the correct adjustment factor for Chronic Renal Failure is 1.14.

4. Patient Age

We proposed a 13 percent payment adjustment for patients 65 years of age and over to reflect the additional costs associated with treating elderly patients. We received a wide range of comments about the proposed age adjustment. In general, the comments favored the creation of additional age groups and payment adjustments.

Comment: Commenters requested clarification on how the proposed 13 percent differential between age groups was calculated. The commenters stated that the proposed adjustment factor is too low and does not reflect the current cost required to treat the elderly.

Several commenters recommended that CMS revise the age groupings to include a payment adjustment for patients under 14 years of age, under 40 years of age, 55 to 64 years of age, and 75 years of age and over. Other commenters suggested a payment adjustment for patients 65 years of age and over with increments added for each additional 5 years in age.

Response: As indicated in the proposed rule (68 FR 66931), the 13 percent differential was calculated using the same cost regression that was used to estimate the payment adjustments for the other variables included in the proposed payment system. The dependent variable was the natural logarithm of average cost per day for each inpatient stay. The regression included a single variable for persons 65 years of age and over to estimate the relative cost per day of persons 65 years of age and over compared to persons less than 65 years of age. Since the cost variable was in logarithms, the age coefficient in the cost regression was then raised to the power of the base e to convert it to the relative payment factor, 1.13.

In response to the public comments to create additional age payment adjustments (under 14 years of age and under 40 years of age, 55 to 64 years of age, and over 75 years of age), we updated our analysis of the impact of age on per diem cost by expanding the age variable (that is, the range of ages for payment adjustments). Since we have relatively few cases for persons under 40 years of age (and virtually no cases for persons under 14 years of age), we combined all persons under 40 years of age into a single category. Similarly, all persons over 80 years of age were placed in a single category. For patients in between 40 and 80 years of age, we categorized cases into 5-year intervals. As indicated in the proposed rule, the cost per day increases with increasing age. With the exception of the 40 through 44 age group, all the older age groups are more costly than the under 40 years of age group, the differences increase for each successive age group, the differences among the age groups increase for each successive age group, and the differences are statistically significant.

Based on these results, in this final rule we are expanding the relative adjustment factor for age from the single factor for patients 65 years of age and over to 8 adjustment factors beginning with age groupings 45 and under 50 years of age to patients 80 years of age and over. The magnitudes of these factors are shown in Table 6 below and in Addendum A. We are also adopting as final the same methodology we used in the proposed rule (that is, cost regression analysis) except we are using an updated and revised regression based on FY 2002 data and the age groupings described above (that is, 5 year intervals and 8 adjustment factors).

5. Variable Per Diem Adjustments

Cost regressions indicate that the per diem cost declines as the length of stay increases. Therefore, we proposed adjustments to account for ancillary and certain administrative costs that occur disproportionately in the first days after admission to an IPF. As we explained in the proposed rule, we examined the per diem cost over a range of 1 to 14 days. According to the FY 1999 MedPAR data file, the per diem costs were highest on day 1 and declined for days 2 through 8 as follows. Per diem costs for days 9 and thereafter remained relatively constant. The proposed cost regression analysis was used to determine the proposed payment adjustment factors. Relative to a stay of 9 or more days, we proposed a variable per diem adjustment of 26 percent for day 1, a 12-percent adjustment for days 2 through 4, and a 5-percent adjustment for days 4 through 8. No variable per diem adjustments would be made after the 8th day.

We received multiple comments on the proposed variable per diem adjustments, primarily dealing with the amount of the proposed payment adjustments and the breakpoints for the adjustments.

Comment: One commenter asked how CMS determined the cost per day for the different lengths of stay. Another commenter recommended more justification of the method used to control for length of stay. Specifically, this commenter asked whether CMS tested alternative breakpoints for the length of stay categories and whether CMS considered other approaches for estimating the relationship between per diem cost and length of stay. One commenter objected to the proposed length of stays blocks, in which days 2 through 4 and days 5 through 8 would be paid at the same rate rather than declining smoothly for each successive day. The commenter believes that the proposed approach creates incentives to terminate or unnecessarily extend the length of stay.

Response: As indicated in the proposed rule, the relationship between cost per day and length of stay was estimated within the same cost regression used to derive other payment adjustments. First, we defined variables for each stay's length of stay (from 1 to 14 days). The effects of the first 14 days on cost were measured relative to stays of more than 14 days. Based on the results of this regression, we considered payment breakpoints for each day up through 14 days. Based on the size and pattern of variation of the regression coefficients for the individual day coefficients (that is, the magnitude of decline), we decided to group the days into the categories presented in the proposed rule (that is, day 1, days 2 through 4, days 5 through 8, and days 9 and thereafter). We then re-estimated the cost regression including the first 3 of these groups and stays of more than 8 days as the reference group.

As a result of converting the regression coefficients to payment factors, we proposed to pay the first day of each stay 26 percent more than the Federal per diem base rate. Similarly, we proposed to pay days 2 through 4 of each stay 12 percent more than the Federal per diem base rate and days 5 through 8 about 5 percent more than the Federal per diem base rate. The Federal per diem base rate implicitly reflects the cost of stays with more than 8 days.

We used regression analysis to estimate the average differences in per diem cost among stays of different length. Regression analysis simultaneously controls for cost differences associated with the other variables (for example, age, DRG, and presence of specific comorbidities). The regression coefficients measure the relative average cost per day for stays of differing lengths compared to a reference group's length of stay. In the proposed rule, the variable per diem adjustment factors derived from the regression coefficients were applied to specific days within the stay. As indicated above, we proposed to pay all stays 26 percent more than the Federal per diem base rate for day 1, 12 percent more than the base payment amount for days 2 through 4, and 5 percent more than the base payment amount for days 5 through 8.

To accurately measure the relative cost of specific days within the stay, we need estimates of the additional or marginal (not average) cost of those days. Using the relative average cost differences as if they were marginal cost differences will result in overpayment for the days with payment factors greater than 1.00. The reason for the overpayment is that, using a 4-day stay as an example, the average cost per day over the 4 days already contains the higher marginal costs of the preceding 3 days. In paying more than the 4-day average cost per day for days 1 through 3, we would be paying more than the total cost of the stay.

In reconsidering the variable per diem adjustments for this final rule, we re-evaluated the length of stay breakpoints in the regression and the method of applying the regression results for payment. Using the FY 2002 MedPAR data, we re-estimated the cost regression, expanding the number of length of stay categorical variables from 1 through 14 to 1 through 30 days in order to potentially allow payments to decline in smaller, more increments over a wider range of days. From the regression, we derived factors indicating the average cost per day, for example, a 1-day stay, a 2-day stay, and a 3-day stay, relative to a stay of more than 22 days.

Since the variable per diem adjustments are applied to all IPFs stays, the adjustments should reflect daily cost differences experienced by all types of IPFs, and not cost differences among different types of IPFs with different lengths of stay. Therefore, we also tested the sensitivity of the regression coefficients to the inclusion of the government-operated IPF stays, which tend to have longer lengths of stay than the other types of IPFs. For example, about one-third of all government-operated IPF stays are longer than 22 days, compared to only 10 to 13 percent of stays in for-profit or non-profit hospitals or in psychiatric units. We found that our coefficients varied little depending on whether cases from government-operated IPFs were included or excluded.

CMS-funded research by RTI International^®, which was not available for the proposed rule, provides additional information about the variation in relative marginal costs by day of the stay. RTI International^® examined the variation in routine resource use across days within stays in its study of a sample of patients from 40 facilities. RTI International^® constructed a measure of a patient's routine cost for each of 7 days during which they were collecting data within a facility.

As a result, RTI International^® data has a significant advantage compared to the MedPAR data that was available at the time of the proposed rule for examining cost variation by day-of-stay. Specifically, RTI International^® data enabled them to estimate a relationship between per diem cost and the day-of-stay that is consistent with the way we used the variable per diem adjustment factors for payment. In addition, since RTI International^® did not average daily routine costs over the entire length of stay, its estimates should provide a better approximation of the relationship of marginal cost than we were able to construct. RTI International^® did not collect information on ancillary usage by day-of-stay. In constructing its measure of daily total cost, RTI International^® allocated 1 day of average ancillary costs from the matching MedPAR stay record. RTI International^® used the same breakpoints that we used for the proposed rule.

In the table below, we compare the revised CMS adjustment factors with the RTI International^® day-of-stay relative weights. Both sets of factors were scaled to set the day-9 (the median length of stay) factor equal to 1.00. The two series of factors are very similar, with the biggest differences occurring for days 2 to 4 and for day 19 and beyond. The differences for days 2 to 4 may be due to how the two methods handle ancillary costs, especially our exclusion of ED costs from the cost variable used in our regression analysis. The differences for day 19 and beyond probably are a result of the fact that RTI International^® only estimated specific day effects for the first 14 days.

Overall, the similarity of the adjustment factors gives us confidence that our variable per diem adjustment factors are reasonably accurate. The revised factors are also responsive to the comment that the variable per diem adjustments should decline more continuously than those presented in the proposed rule. Therefore, in this final rule we are using the updated variable per diem adjustment factors in adjusting per diem payments by day-of-stay. We note that the variable per diem adjustment are made in a budget-neutral manner.

Comment: Several commenters recommended that CMS re-evaluate the decision to have no variable per diem adjustment paid after the 8th day. The commenters requested that we re-examine the analysis supporting the conclusion that “per diem costs for days 9 and thereafter remain relatively consistent with the median length of stay.”

A few commenters expressed concern that averages were used in all analyses except for the proposed variable per diem adjustments that were based on the median length of stay. The commenters believe use of the median creates distortions and requested that CMS analyze the impact if the variable per diem adjustments were based on the average length of stay.

Response: We re-evaluated the decision to make no variable per diem adjustments to the Federal per diem base rate beyond the eighth day. We examined the per diem cost relationship for the first 30 days of the stay and found that beyond day 22, there was no consistent continuing pattern of decline. In addition, since the proportion of stays longer than 21 days is relatively small, there is relatively high statistical variability in the estimates of declining cost increases beyond day 22, which makes the estimates less reliable. As a result of that analysis, we found that the average per diem cost continued to decline until the twenty second day. Therefore, in this final rule we are extending the variable per diem adjustments through day 22. The adjustment for day 22 would be applied to any days after day 21.

We believe the commenter misunderstood the role of the median length of stay in the variable per diem adjustment factors. As indicated in the proposed rule, the median length of stay serves only as a point of reference for the variable per diem adjustment factors relative to the Federal per diem base rate (the day for which the factor equals the base amount). In addition, the actual magnitudes of the variable adjustment factors were not affected by using the median in this manner because the median had no impact on the cost regression from which the variable per diem adjustment factors are derived. The Federal per diem payment would be the same no matter which day of the stay (the median, the mean, or some other day) was used as the reference point. In this final rule, we are adopting as final the same methodology proposed to calculate the variable per diem adjustments.

Comment: A few commenters expressed concern that the lack of variability in average daily charges results in understating the effect of the length of stay variable.

Response: We disagree with the commenters. The RTI International^® research evaluated the variation of per diem cost by day of the stay using a measure of routine cost that varied according to the day of the stay. In addition, the comparison of RTI International^® results and our results did not support the commenters' concerns that the variable per diem adjustment factors are understated.

Comment: Many commenters recommended increasing the per diem adjustment factor for day 1, or for the first several days of care.

One commenter recommended that in order to avoid the significant impact the proposed rule would have on high cost per discharge-short length of stay providers, the variable per diem adjustments for the first days of the stay should be weighted higher. The commenter recommended that CMS double the adjustments to 52 percent for day 1, 24 percent for days 2 through 4, and 10 percent for days 5 through 8.

Other commenters recommended that days 2 and 3 receive the same adjustment factor as day 1. However, some commenters recommended that the per diem payment be uniform rather than variable throughout the patient's stay. They suggested that a higher per diem base payment amount for each day of stay would be preferable and more in line with the distribution of costs over an inpatient episode.

Response: These comments reflect a wide range of opinion about the appropriate range and magnitude of the variable per diem adjustment factors. We have updated and revised our variable per diem adjustment policy on the basis of our analysis of FY 2002 data and in response to public comments. In arriving at the final variable per diem adjustments, we have relied upon our empirical analysis, as previously described earlier in this section, to better approximate the additional costs of each successive day of the stay. We have also compared our results with the results of CMS-funded research by the RTI International^®. We believe that the outcome of the process we undertook to improve the variable per diem adjustment factors is a reasonably accurate, empirically-based set of adjustment factors.

Comment: Several commenters expressed concern that the length of stay assumptions in the proposed rule did not take into consideration that certain interventions necessitate longer stays. A particular commenter indicated that medical safety standards for ECT dictate stays of more than 9 days.

One commenter stated that the elderly and younger chronically mentally ill adults represent two groups with longer than average lengths of stay. Another commenter stated that length of stay might be increased by the inclusion of trainees in a patient's care.

Response: We are not sure that we understand these comments. As required by the BBRA, the IPF PPS is a per diem system. As a result, the IPF PPS recognizes differences in length of stay and will pay the Federal per diem base rate and applicable adjustments for each day of the inpatient stay. Therefore, the IPF PPS accounts for differences in length of stay regardless of cause (including providing ECT or other factors).

Comment: A few commenters recommended that CMS undertake a research inquiry into the added staffing costs for the first few days of a stay at an inpatient psychiatric unit or develop two per diems, one for routine patients and another for “clinically determined critical patients.”

Response: The RTI International^® study addressed the issue raised by this comment because it examined the variation in routine cost by day of the stay. RTI International^® studied this relationship for all the patients in its sample, which included the full range of patients treated in IPFs. In addition, we are not sure how we could define “clinically determined critical” patients, especially considering the common practice of admitting to psychiatric facilities only those patients whose medical needs have either been resolved or are sufficiently controlled as to require limited attention for the period of the psychiatric admission.

Comment: One commenter expressed concern that CMS would misinterpret increases in IPF admissions that result from the planned transition of inpatient psychiatric care from government-operated facilities to community-based resources such as private hospitals.

Response: Under the IPF PPS, both admissions referred to in the comment would be paid on a per diem basis, so that each facility (the government-operated facility and the private hospital) would be paid for the days of care it provides.

Comment: One commenter recommended that CMS more accurately reflect the MedPAR data by using a variable Patient Day adjustment equal to the median value of 9 days, rather than limit the adjustments to days 1 through 8.

Response: By extending our analysis through 30 days, we more fully modeled the shape of the relationship between average per diem costs and length of stay and did not truncate the adjustments at either the median or the mean length of stay. As a result, the revised variable per diem adjustment factors presented in this final rule more accurately reflect the cost-day relationship than those we presented in the proposed rule.

Comment: One commenter recommended that CMS provide more justification for the method used to control for length of stay.

A few commenters expressed concern that use of the median length of stay significantly understates the length of stay for an IPF that accepts chronic psychiatric patients (for example, a government-operated psychiatric hospital). The commenters believe that the proposed IPF PPS rewards acute psychiatric facilities for discharging patients quickly and provides an incentive for those facilities to discharge patients into government-operated IPFs.

Response: We believe the commenter misunderstood the intent of the variable per diem adjustment policy, which is not to control for length of stay, but to better align the payment of each day of the say with its corresponding cost. Therefore, the facilities would have no incentive to either shorten or extend a patient's length of stay beyond what is clinically needed.

We agree with the commenters that certain types of IPFs have lengths of stay greater than the median length of stay. The variable per diem adjustment factors are intended to track the relative costs an IPF needs to spend on a case throughout the days of a stay. Thus, a facility with a length of stay greater than the median, or the mean for that matter, should be adequately reimbursed for the cost of care provided to a Medicare beneficiary. As explained above, we do not believe that the final IPF PPS provides an incentive for early discharge from one type of IPF to a government-operated facility. In addition, our use of the median length of stay has no effect on the actual payment amounts for each day of the stay.

6. Other Patient-Level Adjustments

Although we proposed specific patient-level adjustments, we recognized that there were other variables not collected on the claim form. Therefore, we requested public comments on other patient-level adjustments for the IPF PPS. In response to our request for public comments, we received numerous comments recommending that we consider the following other types of adjustments:

a. Gender

We invited public comments on the appropriateness of including a gender variable as a payment adjustment.

Comment: Several commenters stated that elderly female patients represent 68 to 70 percent of the population they serve and recommended that CMS recognize the cost differential in treating female patients.

Response: We analyzed the FY 2002 data and found that the cost regression continues to imply that female patients are approximately 2 percent more costly than male patients. However, as we found in the proposed regression analysis, adding an adjustment for gender increases the explanatory power of the patient model by less than one half of 1 percent, which means that the addition of gender does very little to improve explanatory power of the overall model. In addition, we are unable to determine the extent to which the interaction of psychiatric unit status with age and gender indicates higher direct costs of treating the elderly and women, as opposed to other reasons for the higher costs of psychiatric units. However, to the extent that gender is correlated with age and DRGs, facilities will be partially reimbursed for gender-related costs, since gender was not included as a variable in the regression. Therefore, we are not adopting a patient-level adjustment for gender.

b. Patients Admitted Through the Hospital's ED

We received many comments recommending that we recognize the cost of ED services and provide a patient-level adjustment for patients who were admitted to a distinct part psychiatric unit through the hospital's ED.

Comment: Many commenters recommended that CMS add a patient-level adjustment for patients who are admitted through the ED of the same hospital for inpatient psychiatric care.

Response: Our analysis indicated these cases were more costly on a per diem basis than cases without an ED admission. However, we are not including an adjustment for patients admitted through the ED. We are concerned about creating an incentive for psychiatric units in acute care hospitals with EDs to ensure that all psychiatric patients are admitted through the ED. However, we are providing a facility-level adjustment for psychiatric hospitals, or psychiatric units of acute care hospitals, with qualifying ED. Additional information regarding the analysis of ED costs is included in section VI.B.5.b. of this final rule.

c. Patients Who Receive Electroconvulsive Therapy (ECT)

We received numerous comments recommending that we include ECT as a patient-level adjustment because furnishing ECT treatment adds significantly to the cost of these IPF stays.

Comment: Several commenters recommended that CMS include ECT (procedure code 90870) under DRG 424 (Operating room procedure with principal diagnosis of mental illness) that has an adjustment factor of 1.22. One commenter suggested that DRG 430, “Psychosis” be disaggregated into two DRGs, “Psychosis with ECT,” incorporating the added costs for ECT treatment and “Psychosis without ECT.”

Other commenters recommended that CMS provide as an alternative, an add-on payment to the DRG for those patients who receive ECT treatments.

Many commenters recommended modifying the payment structure to include a separate payment adjustment for ECT, which should be higher than the payment adjustment for DRG 424.

Response: After reviewing the public comments, we analyzed cases with ECT using the FY 2002 MedPAR data. We were able to identify ECT cases by the presence of procedure code 90870. Our analysis indicated that ECT cases comprised about 6 percent of all cases, and that almost 95 percent of ECT cases were treated in psychiatric units. Even among psychiatric units, ECT cases are concentrated among a relatively small number of facilities.

Overall, approximately 450 facilities had cases with ECT. Among these facilities, we estimate the mean number of ECT cases per facility to be approximately 25. In addition, approximately one-half of the IPFs providing ECT had no more than 15 cases in FY 2002.

Consistent with the comments we received about ECT, our analysis and review indicated that cases with ECT are substantially more costly than cases without ECT. On a per case basis, ECT cases are approximately twice as expensive as non-ECT cases ($16,287 vs. $7,684). Most of this difference is due to differences in length of stay (20.5 days for ECT cases vs. 11.6 days for non-ECT cases). The ancillary costs per case for ECT cases are $2,740 higher than those for non-ECT cases.

Based on this analysis, in this final rule we are providing an adjustment for each ECT treatment furnished during the IPF stay. In order to receive the payment adjustment, IPFs must indicate on their claims the revenue code and procedure code for ECT (Rev Code 901; procedure code 90870) and the number of units of ECT, that is, the number of ECT treatments the patient received during the IPF stay. Providing this data will ensure that facilities are appropriately reimbursed for the treatments they provided.

After careful review and analysis of IPF claims, we were unable to separate out the cost of a single ECT treatment. Therefore, we are using the pre-scaled and pre-adjusted median cost for procedure code 90870—developed for the hospital OPPS, based on hospital claims data.

We used unadjusted hospital claims data under the OPPS, that is, the pre-scaled and pre-adjusted median hospital cost per treatment, to establish the ECT payment because we did not want the ECT payment under the IPF PPS to be affected by factors that are relevant to OPPS but not specifically applicable to IPFs. The median cost is then standardized and adjusted for budget neutrality. We will adjust the ECT rate for wage differences in the same manner that we adjust the per diem rate. The median cost for all hospital OPPS services are posted after publication of the hospital OPPS proposed and final rules at the following address: http://www.cms.hhs.gov/​providers/​hopps.

As explained above, we decided to pay the median cost for an ECT treatment, posted as part of the calendar year (CY) 2005 OPPS update, which is based on CY 2003 outpatient hospital claims. The amount is $311.88. Using the same OPPS CY 2003 claims that were used to calculate the aforementioned ECT median, we were able to calculate the average number of ECT treatments for a given patient to be approximately 9. A rate of $311.88 per ECT treatment multiplied by 9 is very close to the $2740 difference in ancillary costs observed for ECT and non-ECT cases. Accordingly, we believe that the payment adjustments for ECT will appropriately and adequately provide payment for ETC services provided to IPF patients. After applying the standardization factor, behavioral offset, stop-loss adjustment, and outlier adjustment (as described in section V.C. of this final rule), the adjusted ECT payment is $247.96.

We have established the ECT adjustment as a distinct payment under the PPS methodology, our preferred approach would be to include a patient level adjustment as a component of the model (for example, determined through the regression analyses) to account for the higher costs associated with ECT. We believe the approach will better control incentives towards over-utilization and be more consistent with the approach used for other patient level adjustments under the PPS. During the transition period we expect to collect more data on the number of ECT treatments per stay, and associated costs. We will utilize these data to evaluate alternative approaches for incorporating an adjustment for ECT in the payment system. We expect to complete this analysis during the first year of the transition and potentially propose changes at the time of the first annual update of the payment system.

ECT is an intensive procedure. Therefore, we are concerned that including a payment adjustment for ECT treatments in the final IPF PPS could result in a rise in the use of ECT treatment. We will monitor this area to ensure that the increased payments do not lead to changes in the frequency of utilization.

d. Patients Involuntarily Committed to the IPF

We did not proposed to provide a payment adjustment for patients who are involuntarily committed to an IPF. However, we received multiple comments encouraging us to recognize the additional costs associated with these patients.

Comment: Several commenters indicated that patients involuntarily committed to an IPF often require costly court proceedings before treatment can begin and that the hospital my incur cost for caring for these patients while awaiting the court decision.

Other commenters identified patient management issues, for example, more frequent one-on-one staff attention and more complex discharge planning. A few commenters indicated that involuntarily committed patients are often uncooperative and difficult to treat. One commenter reported a 27 percent longer length of stay for involuntarily committed patients.

Response: One of the fields on the claim form indicates if patients were referred to the IPF by law enforcement or if the commitment were court ordered (FL 20, item 8, court/law enforcement). As a result, we were able to analyze the FY 2002 claims data to determine if the costs identified by the commenters are evident in the claims. The data did not indicate that patients involuntarily committed to the IPF are more costly on a per diem basis. We note that many of the costs associated with involuntary commitments (for example, legal fees, staff time to accompany the patient to court, and transportation costs) are part of the hospital's average routine per diem cost.

In addition, there are certain costs that are the responsibility of the court system or law enforcement, for example, where a court orders a 3-day psychiatric evaluation for a patient or where discharge is delayed pending court action. Thus, IPFs should be adequately reimbursed for patients involuntarily committed, even in the absence of a specific payment adjustment.

Therefore, at this time we are not providing an adjustment for involuntarily committed patients.

e. Administrative Necessary Days

We received several comments recommending that we recognize the cost of administrative necessary days for continued inpatient care when discharge is delayed due to a lack of community resources.

Comment: Commenters indicated that hospitals would be unable to discharge a patient without an appropriate discharge plan. The commenters requested that CMS provide reimbursement for this type of situation.

Response: Current hospital discharge planning requirements in § 482.43(a) and (b) require the discharge planning evaluation to include the likelihood of a patient needing post-hospitalization services and the availability of those services. Hospital personnel must complete the evaluation on a timely basis so that appropriate arrangements for post-hospital care are made before discharge, and to avoid unnecessary delays in discharge.

In addition, § 482.43(c)(4) requires that the hospital must reassess the patient's discharge plan if there are factors that may affect continuing care needs or the appropriateness of the discharge plan.

Moreover, § 412.27(c)(5) states, “the record of each patient who has been discharged must have a discharge summary that includes a recapitulation of the inpatient's hospitalization in the unit and recommendations from appropriate services concerning follow-up or aftercare as well as a brief summary of the patient's condition on discharge.”

Consequently, if an IPF determines that a patient needs post-hospitalization placement, then a statement to this effect is expected to be included in their discharge plan. Furthermore, if a patient cannot be safely discharged without this post-hospitalization placement and this placement is not available, then the patient has not met their discharge objectives and requires continued active treatment.

After careful review, we have decided not to provide additional payment for administrative necessary days for several reasons. Since claim data does not include coding or documentation for administrative data, we are unable to identify and discern the cost of these days. Therefore, we are unable to determine the extent to which the costs of administrative necessary days are included in the Federal per diem base payment amount.

Finally, since the IPF PPS is a per diem payment methodology, we are concerned about inadvertently creating an incentive to unnecessarily delay discharge in order to receive additional payment for administrative necessary days.

C. Facility-Level Adjustments

In the proposed rule, we proposed adjustments for the IPF's wage area, rural location, and teaching status.

1. Wage Index

Due to the variation in costs and because of the differences in geographic wage levels, we proposed that payment rates under the IPF PPS be adjusted by a geographic wage index. We proposed to use the unadjusted, pre-reclassified hospital wage index to account for geographic differences in labor costs. In the proposed rule, we proposed to use the inpatient acute care hospital wage data to compute the IPF wage since there is not an IPF-specific wage index available. We believe that IPFs generally compete in the same labor market as acute care hospitals since the inpatient acute care hospital wage data should be reflective of labor costs of IPFs. We believe this to be the best available data to use as proxy for an IPF specific wage index. We proposed to adjust the labor-related portion of the proposed Federal per diem base rate for area differences in wage levels by a factor reflecting the relative facility wage level in the geographic area of the IPF compared to the national average wage level for these hospitals. We believe that the actual location of the IPF as opposed to the location of affiliated providers is most appropriate for determining the wage adjustment because the data support the premise that the prevailing wages in the area in which the IPF is located influence the cost of a case. Thus, in the proposed rule and in this rule, we are using the inpatient acute care hospital wage data without regard to any approved geographic reclassification as specified in section 1886(d)(8) or 1886(d)(10) of the Act. Specifically, in this rule, we are using the FY 2005 hospital wage index (unadjusted, pre-reclassified) based on MSA definitions defined by OMB in 1993 (as opposed to the new MSA definitions that were used to define labor markets for the FY 2005 IPPS). Once we implement the IPF PPS, we will assess the implications of the new MSA definitions on IPFs. At the time of the proposed rule, the 2003 MSA definition had not been implemented for any medicare programs and consequently, were not proposed. We note that, after the publication of the IPF PPS proposed rule, new MSA definitions have been adopted for use in the IPPS. We, however, are not adopting those new definitions in this final rule. We expect that use of the new MSA (or labor market) definitions may have a significant impact on the wage index applied to IPFs and associated payments. Thus, before their use could be proposed, we would have to conduct a thorough analysis of their impact on the IPF PPS. Moreover, and most importantly, we believe it is appropriate to provide an opportunity for IPFs and other interested parties to comment on the use of the new definitions before proceeding with their possible application. We plan to publish in a proposed rule any changes that we consider for new labor market definitions, in order to provide the public with an opportunity to comment.

Comment: Several commenters recommended that CMS apply the hospital wage index with geographic reclassifications in the same way that other hospital PPS adjust payments to reflect wage differences. Commenters believe that the reclassification process ensures that areas that are geographically close to an MSA may compete to employ a sufficient amount of skilled healthcare workers. Other commenters believe that the pre-reclassified wage index may result in a potential decrease in payment, especially for psychiatric units within hospitals that draw from the same workforce as acute care hospitals.

Response: The statute does not require geographic reclassification of other hospitals paid under TEFRA (for example, freestanding psychiatric hospitals) or other hospitals paid under different prospective payment systems. Geographic reclassifications are not recognized under the IRF or LTCH payment systems, and are not recognized under the final IPF PPS.

Comment: A few commenters requested a modification to the portion of the payment that is adjusted by the wage index. The commenters stated that the proposed wage index should be applied to 72.8 percent of the Federal per diem base rate, as reflected in the proposed 1997-based excluded hospital with capital market basket. Generally, commenters in wage areas with a wage index above 1.0 indicated that the proposed labor portion of the payment was too low and commenters in wage areas with a proposed wage index less than 1.0 indicated that the labor portion was too high.

One commenter indicated that psychiatric care is more labor intensive than other modes of inpatient care, thus the commenter recommended that CMS research the costs of providing psychiatric care, and develop a labor adjustment that adequately compensates for the increased intensity of care for psychiatric patients.

Response: In both the proposed rule and in this final rule, to account for wage differences, we first identified the proportion of labor and non-labor components of costs. We used the 1997-based excluded hospital market basket with capital to determine the labor-related share of cost. We calculated the labor-related share as the sum of the weights for those cost categories contained in the 1997-based excluded hospital with capital market basket that are influenced by local labor markets. These cost categories include wages and salaries, employee benefits, professional fees, labor-intensive services, and a share of capital-related expenses.

The labor-related share for the implementation period of the final IPF PPS (January 1, 2005 through June 30, 2006) is the sum of the relative shares which measure the relative importance of each labor-related cost category for this period. It also reflects the different rates of price change for these cost categories between the base year (FY 1997) and this period. 0 labor-related components of operating costs (wages and salaries, employee benefits, professional fees, and labor-intensive services) is 68.818 percent, as shown below in Table 8. Since capital cost also contains a significant component of labor-related cost, the labor-related share of total cost will be greater than the labor-related share of operating costs alone. The portion of capital cost that is influenced by local labor markets is estimated to be 46 percent. Because the capital accounts for 7.323 percent of the 1997-based excluded hospital with capital market basket for the period January 1, 2005 through June 30, 2006, the labor-related share of capital cost is 46 percent of 7.323 percent. The result, 3.369 percent, is then added to the 68.818 percent calculated for operating costs to determine the labor-related share of total cost. The resulting labor-related share that we are using in this IPF PPS rule is 72.247 percent. The table below shows that the labor-related share would have been 72.571 percent if we had not rebased the excluded hospital with capital market basket using more recent 1997 data rather than using 1992 data. As shown in Table 8, rebasing results in a lowering of the labor-related share by 0.324 percentage points.

The base methodology used to calculate the labor-related share for IPFs is the same as that used for calculating the labor-rated share for IPPS, SNFs, HHAs, LTCH, and IRFs PPS. The difference is that except for the IPPS, we use the relative importance for the effective period in developing this share, which changes annually. For IPPS, the labor share remains constant until the market basket is rebased.

CMS agrees with the commenter that it is important to have a market basket and labor share appropriate for use under the IPF PPS. We believe that using the excluded hospital with capital market basket accomplishes this goal. However, we indicated in the proposed rule that we plan to continue to study the feasibility of developing a market basket specific to IPF services. We hope that we may eventually be able to develop a market basket and labor-related share based primarily on IPF data (see 68 FR 66928).

The labor-related relative share of total cost in this rule changed from that in the proposed rule for two reasons. First, the labor-related share of 72.247 in this rule comes from Global Insight's 2004: quarter 3 forecast, with historical data through 2004: quarter 2, while the proposed rule used data from the 2002: quarter 4 forecast, with historical data through 2002: quarter 3, to calculate the proposed labor share of 72.828. Second, in addition to using more historical data in a more recent forecast, there is a different implementation period in this final rule, meaning that different periods of data were used to calculate the labor-related relative importance in this rule.

Comment: Several commenters requested that CMS establish a floor for the urban wage index so that an urban wage index would not fall below the wage index in a rural area in the same state. Another commenter requested that CMS apply the section of the MMA to the IPF PPS, which would limit an IPF's wage index to a minimum of 1.

Response: We did not propose a wage index floor. We are unclear of what the commenter is referring to because there is no MMA provision that limits the hospital wage index to a minimum of 1.0. In order to be consistent with the wage area adjustments used in the PPS developed for other excluded hospitals, we did not apply a floor wage index under the IPF PPS.

Comment: Many commenters suggested that CMS use more recent hospital wage data for the final IPF PPS.

Response: We are also using the best available hospital wage index data in this final rule (that is, the wage data used to establish the FY 2005 IPPS wage index for the October 1, 2004). We will continue to use the best data available for future updates to the IPF PPS.

2. Rural Location

We proposed a 16 percent payment adjustment for those IPFs located in a rural area. This adjustment was based on the proposed regression analysis, which indicated that the per diem cost of rural facilities was 16 percent higher than that of urban facilities after accounting for the influence of the other variables included in the regression. Many rural IPFs are small psychiatric units within small general acute care hospitals. In the proposed rule, we stated that small-scale facilities are more costly on a per diem basis because there are minimum levels of fixed costs that cannot be avoided, and they do not have the economies of size advantage.

We received several comments regarding the proposed rural adjustment. Most commenters supported the rural adjustment and encouraged us to recognize the higher cost incurred in rural settings.

Comment: Commenters expressed concern that despite the 16 percent adjustment to the Federal per diem base rate for IPFs located in rural areas Medicare payment would decrease for rural psychiatric units.

Response: In implementing this rule, we updated our cost regression analysis using the most recent complete data available (that is, FY 2002 data). Based on the results of our regression analysis, we are now providing a payment adjustment for IPFs located in rural areas of 17 percent instead of the proposed 16 percent. The small change in the rural payment adjustment is largely the result of the adjustment we made to the cost data to account for the ED adjustment. A full description of the ED policy appears later in this section.

As is the case with implementing any prospective payment system, since the payment rates are not directly tied to the costs of each individual facility, relatively high cost facilities may experience reductions in Medicare payments. However, our analysis of the impact of this rule during the first year of implementation (see section VIII of this final rule) show that on average rural facilities are expected to have a payment to cost ratio of 1.00. This means that Medicare payments during the first year of the IPF PPS transition are expected to be the same as they would have been had the IPF PPS not been implemented and IPFs continued to be paid 100 percent.

Comment: Several commenters specifically expressed concern that the multipliers used for urban and rural facilities are inappropriate and do not adequately adjust for higher per bed cost in smaller facilities. In addition, several commenters encouraged CMS to add a reasonable payment adjustment for urban psychiatric units.

Other commenters stated that if the proposed rules are adopted, hospitals may choose to close their psychiatric units.

Response: We did not include an explicit payment adjustment for urban facilities in the proposed rule and we are not adopting one in this final rule. We are not including this type of adjustment factor since our adjustment for rural facilities is based on an explicit comparison of the relative per diem costs of rural and urban facilities after accounting for the effects of the other variables included in the regression as previously explained in the cost regression section of this final rule. The result of that comparison (as reflected in our cost regression) was that rural facilities are more costly than urban facilities, largely because rural facilities are smaller on average than urban facilities. In addition, because a variable reflecting facility size was not included in the cost regression, the rural payment adjustment factor may partially reflect the influence of size on per diem cost.

As previously stated, we have not included an explicit payment adjustment factor to account for the higher per diem costs of small facilities, because we think that to do so is counter to the basic principle of prospective payment systems that payment adjustments should be based on characteristics that are not under the control of the facility. Specifically in the case of psychiatric units where a facility can choose how much of its inpatient psychiatric care it wishes to include in its Medicare certified unit, we would be concerned that a facility could reduce the size of its Medicare-certified unit in order to increase Medicare payments.

We plan to monitor the impact of the IPF PPS on the financial status of psychiatric facilities. We are particularly concerned about potential effects of facility closures on beneficiaries' access to inpatient psychiatric care. As a result of this issue, we are adopting a stop-loss provision as part of the transition to assist all IPFs with revenue shortfalls during the transition period (see section V.C.3. of this final rule for a discussion of the stop-loss provision).

3. Teaching Adjustment

We proposed to establish a facility level adjustment to the Federal per diem base rate for IPFs that are teaching institutions. In the past, we have made direct graduate medical education (GME) payments (for direct costs such as resident and faculty physician salaries, and other direct teaching costs) to teaching hospitals including those paid under the IPPS and those paid under the TEFRA rate of increase limits. However, we did not make separate indirect medical education (IME) payments to teaching hospitals paid under the TEFRA rate-of-increase limits because payments to these hospitals are based on the hospitals' reasonable costs. IME payments are authorized under the IPPS statute to be paid as an add-on to the IPPS per case payment, and there are no per case payments under the TEFRA system. In this final rule, we are establishing a facility-level adjustment for IPFs that are, or are part of, teaching institutions. The facility-level adjustment we are providing for teaching hospitals under the new IPF PPS parallels the IME payments paid under the IPPS. Both payments are add-on adjustments to the amount per case (there is now a per case payment to which the IPF teaching adjustment will be added) and both are based in part on the number of full-time equivalent (FTE) residents training at the facility.

In the proposed rule, we proposed to calculate a teaching adjustment based on the IPF's “teaching variable,” which is one plus the ratio of the number of FTE residents training in the IPF divided by the IPF's average daily census (ADC). Based on our initial regression analysis, we proposed to raise the teaching variable to the .5215 power. We also requested suggestions from the public regarding how to estimate IPFs' indirect teaching costs and alternative methodologies to recognize the higher costs of teaching IPFs. However, we did not receive any suggestions on this issue.

Accordingly, we are adopting our proposed formula for calculating the adjustment in this final rule. Based on the final regression analysis using FY 2002 data, we are raising the teaching variable from .5215 power to the .5150 power.

We also indicated we were considering alternatives to limit the incentives for IPFs to add FTE residents for the purpose of increasing their teaching adjustment. We indicated that we were considering imposing a cap, similar to that established by sections 4621 and 4623 of the BBA for the IPPS, and noted that these caps already apply to teaching hospitals, including IPFs, for purposes of direct GME payments according to regulations at § 413.75 through § 413.83.

As indicated in the proposed rule (68 FR 66932), we were concerned about establishing an open-ended payment for the teaching adjustment because the BBA froze the number of residents that hospitals may count for both direct and indirect GME payments in order to reduce incentives for teaching institutions to add residents. We recognized that if we imposed no limits on the teaching adjustment under the IPF PPS, teaching programs in those facilities could grow and receive payments in a manner that is inconsistent with that in teaching hospitals paid under the IPPS. In addition, we were concerned that if a teaching hospital had a distinct part psychiatric unit and had a number of FTE residents above the amount recognized for reimbursement under the BBA limits, the hospital could potentially circumvent those limits by assigning residents to train in the IPF. For example, if a teaching hospital has 110 FTE residents of which only 100 are recognized for purposes of Medicare IME reimbursement under the BBA limits, the hospital could assign the excess 10 residents to its distinct part psychiatric unit where those FTE residents would be included for purposes of the teaching adjustment to the IPF PPS payments, which is similar in amount to IPPS IME payments. As a result, the hospital would be able to count all 110 FTE residents for purposes of calculating a teaching adjustment, in contradiction to the Congress' intent in establishing the BBA limits.

We considered imposing a cap that would operate in a substantially similar manner to the BBA limits on the number of FTE residents that may be counted for purposes of making IPPS IME payments. The BBA cap operates by limiting the number of allopathic and osteopathic FTE residents that Medicare will recognize for the purposes of calculating IPPS IME payments to no more than the number of FTE residents in a teaching hospital's most recent cost reporting period ending on or before December 31, 1996. In addition, the BBA placed a cap on the entire resident-to-bed ratio used to calculate the IPPS IME payment so that a hospital's ratio in its current cost reporting period could not exceed the ratio from its previous cost reporting period.

In response to public comments on the teaching adjustment, only one commenter agreed with the appropriateness of establishing a cap on the number of FTE residents that may be counted for purposes of the teaching adjustment under the IPF PPS. The majority of commenters was opposed to imposition of any resident cap and indicated that a cap would be arbitrary and burdensome.

After carefully reviewing the public comments, we have decided to adopt a cap on the number of FTE residents that may be counted under the IPF PPS for the teaching adjustment. We made this decision in order to—(1) exercise our statutory responsibility under the BBA to prevent any erosion of the resident caps established under the IPPS that could result from the perverse incentives created by the facility adjustment for teaching under the IPF PPS; and (2) avoid creating incentives to artificially expand residency training in IPFs, and ensure that the resident base used to determine payments is related to the care needs in IPF institutions.

In adopting the FTE resident cap for purposes of the IPF PPS teaching adjustment, we wish to emphasize that we are not limiting the number of residents teaching institutions can hire or train; we are limiting the number of residents that may be counted for purposes of calculating the IPF PPS teaching adjustment, and thus, the amount Medicare will pay for the teaching adjustment under the new IPF PPS.

The FTE resident cap we are establishing will work identically in freestanding teaching psychiatric hospitals and in distinct part psychiatric units with GME programs. In order to establish the cap on the number of residents used in calculating the IPF PPS teaching adjustment, the following policies will apply.

• Similar to the regulations for counting FTE residents under the IPPS as described in § 412.105(f), we will calculate the “base year” number of FTE residents that trained in the IPF based on the hospital's most recently filed cost report before November 15, 2004. Residents with less than full-time status and residents rotating through the psychiatric hospital or unit for less than a full year will be counted in proportion to the time they spend in their assignment with the IPF (for example, a resident on a full-time, 3-month rotation to the IPF will be counted as 0.25 FTEs for purposes of counting residents to calculate the ratio). Hospitals can file adjusted cost report data with their FIs until the cost report is settled if they believe the resident counts as submitted on that cost report are incorrect. For purposes of determining an IPF's teaching adjustment under the IPF PPS, the number of FTE residents in the numerator cannot exceed the number of FTE residents in the hospital's most recently filed cost report.
• The denominator used to calculate the teaching adjustment under the IPF PPS is the IPF's average daily census (ADC) from the current cost reporting period. As we indicated in the proposed rule, although a hospital's number of available beds is used in the denominator of the IPPS IME adjustment, the ADC is used in the denominator of the ratio used to compute the IME adjustment under the capital PPS as specified at § 412.322. We are using the ADC for the teaching adjustment under the IPF PPS rather than the number of beds because the ADC is more closely related to the IPF's patient load, and thus, its need for interns and residents. As we stated in the proposed rule, we also believe the ADC is easier to define precisely and less subject to manipulation.

Thus, under the IPF PPS, we are placing a cap on the number of FTE residents (that is, the numerator) used for purposes of computing the teaching adjustment, and not on the ADC (the denominator), or on the entire ratio. An IPF's FTE resident cap will ultimately be determined based on the final settlement of the hospital's cost report filed most recently before November 15, 2004. If a change is made to the base year cost report, the intermediary will reconcile any changes in IPF PPS teaching payments as appropriate.

If a psychiatric hospital or unit has fewer FTE residents in a given year than in the base year, payments in that year will be based on the lower number. This approach is consistent with the IME adjustment under the IPPS. The hospital will be free to add FTE residents and count them for purposes of calculating the teaching adjustment until it returns to its base year FTE resident count.

In this final rule, we are adopting the policy currently applied under the BBA for IPPS teaching hospitals that start new teaching programs as specified in § 413.79 (1) for new teaching IPFS and for teaching IPFs that start new programs. We note that under § 412.105(f)(1)(vi) concerning IME payments under the IPPS, hospitals that have shared residency rotational relationships may elect to apply their respective IME resident caps on an aggregate basis via a Medicare GME affiliation agreement. Our intent is not to affect affiliation agreements and rotational arrangements for hospitals that have residents that train in more than one hospital. We are not implementing a provision concerning affiliation agreements specifically pertaining to the FTE caps used in the teaching adjustment under the IPF PPS at this time. This is an area we expect to closely monitor, and we will consider allowing IPFs to aggregate and adjust their FTE caps through affiliation agreements in the future.

We believe these policies fairly balance our responsibilities under the statute to assure appropriate enforcement of the BBA and the overall limits on payment adjustments for teaching hospitals with the greater precision that can be achieved by adjusting payments for teaching IPFs. We also believe that we have designed a cap that balances the need for limits with the unique conditions of teaching programs in freestanding psychiatric hospitals and in distinct part psychiatric units. We will, however, monitor the impact of these policies closely and consider changes in the future when appropriate.

Comment: Several commenters indicated that a cap amounts to an absolute freeze on the number of residents that Medicare will recognize for payment purposes. In addition, the commenters stated that a cap allows only decreases and no increases in established resident counts at any time.

Response: We acknowledge that the number of FTE residents will be frozen under the IPF PPS. As discussed above, we are adopting a cap on the number of FTE residents that may be counted under the IPF PPS teaching adjustment. This policy is to exercise our statutory responsibility under the BBA to prevent any erosion of the resident caps established under the IPPS that could result from the perverse incentives created by the facility adjustment for teaching hospitals under the IPF PPS. In addition, we wish to avoid creating incentives to artificially expand residency training in IPFs, and ensure that the resident base used to determine payments is related to the care needs in IPF institutions. Again, we will monitor the impact of these policies closely and consider changes in the future when appropriate.

Comment: Several commenters were concerned that the administrative burden in reviewing resident counts back to 1996 cost reports would be excessive and recommended not imposing an FTE resident cap for the IPF PPS teaching adjustment for this reason.

Response: The resident cap under the IPPS is based on the hospital's 1996 cost report. However, the resident cap we are establishing under the IPF PPS relies on the number of residents training in the IPF for the most recently filed cost report before November 15, 2004. In addition, establishing the IPF PPS resident cap does not require the hospitals to submit information not currently included in their cost reports. As a result, we do not believe there is a significant burden associated with establishing the IPF PPS resident cap.

Comment: Several commenters asked if the teaching adjustment would be limited to those hospitals with a dedicated psychiatric teaching program. In addition, the commenters asked if the adjustment would also apply to hospitals that schedule rotations to the psychiatric unit from a non-psychiatric teaching program.

Response: Under the IPPS, Medicare makes IME payments only for costs associated with residents in approved graduate medical education (GME) programs as defined in § 412.105(f)(1)(i) that are approved by one of the organizations listed in § 415.152, not residents in other types of teaching programs. Thus, IPFs that have residents in approved GME programs will receive the IME adjustment. The GME program could be a psychiatric teaching program or scheduled rotations to the IPF unit from a non-psychiatric teaching program.

Comment: One commenter urged CMS to consider applying any cap on the number of interns and residents in a manner that is less sensitive to rapid declines in patient census. The commenter believes the use of the ratio of residents to ADC will negatively affect government-operated IPFs.

Response: Although we are unsure of the commenter's point, the commenter seems to be implying that the teaching adjustment would decline if there were a reduction in the IPF's ADC. However, a decrease in the ADC would result in an increase in the teaching adjustment.

Comment: One commenter requested that CMS provide an example to show how the calculation of the teaching adjustment would be computed. The commenter requested that the example use a hypothetical resident count and ADC and the final teaching adjustment factor.

Response: We were not able to present a single proportional factor that represents the payment adjustment for teaching as we did for most of the other payment variables (for example, age and rural location). The reason is because the teaching adjustment varies among teaching hospitals depending on the degree of their teaching intensity as measured by the ratio of interns and residents to the ADC.

The following example shows a step-by-step calculation of the teaching adjustment for 2 teaching hospitals. Hospital A has an interns and residents to ADC ratio of 0.10. Hospital B has an interns and residents to ADC ratio of 0.20.

Step 1: Add 1.0 to the interns and residents to ADC ratio:

Hospital A: 1.0 + 0.1 = 1.1

Hospital B: 1.0 + 0.2 = 1.2

Step 2: Raise the factors in Step 1 to the power given by the regression coefficient for the teaching variable (.5150).

Hospital A: 1.1 × exp (.5150) = 1.050

Hospital B: 1.2 × exp (.5150) = 1.098

The Step 2 results indicate that Hospital A's payment will be 5.1 percent higher than the comparable payment for a non-teaching hospital and the Hospital B's payment will be 9.9 percent higher than the comparable payment for a non-teaching hospital.

Step 3: Multiply the factors obtained in Step 2 by the appropriate per diem payment adjusted by all other relevant payment factors. For purpose of this example, the per diem payment is assumed to be $625 for both Hospital A and Hospital B.

Hospital A: $625 × 1.050 = $656.25

Hospital B: $625 × 1.098 = $686.25

The step 3 results indicate that Hospital A's per diem payment would be $656.25 compared to $686.25 for Hospital B.

Comment: A commenter questioned why CMS used the ratio of interns and residents to the ADC, rather than the ratio of interns and residents to the number of beds.

Response: Using the ADC rather than the number of beds as the denominator of the teaching variable has two main advantages: Whereas there are many different and frequently imprecise ways of counting beds (licensed beds, available beds, staffed beds), the ADC is a single standard measure that hospitals know how to calculate. It is just the total number of patients days of care divided by 365, the number of days in the year.

Average daily census, which reflects the number of occupied beds in a year, is a readily available, more consistent measure than the number of beds because patient days are more accurately measured than are beds. Because it is directly measured by patient days, ADC is also less subject to understatement in an effort to increase the value of the teaching variable and in turn, teaching payments.

4. Other Facility-Level Adjustments

In the proposed rule, we indicated that we considered facility-level adjustments for IPFs located in Alaska and Hawaii and an IPF's disproportionate share intensity. Other adjustment factors discussed in this section were requested in public comments.

a. Adjustment for Psychiatric Units

In the proposed rule, we did not propose an adjustment for psychiatric units. We received a significant number of public comments expressing concern that the proposed IPF PPS is biased towards psychiatric hospitals and detrimental to psychiatric units. Therefore, the commenters requested that we provide an adjustment specifically for psychiatric units. We are not adopting an adjustment for psychiatric units in this final rule.

Comment: Several commenters stated that the data analysis indicated that the average per diem cost in psychiatric units ($615) was 37 percent higher than the average per diem cost in psychiatric hospitals ($444). Although the proposed patient and facility adjustments account for 19 percent of the difference in average per diem costs, the commenters expressed concern that the proposed rule did not propose a specific adjustment for psychiatric units to account for the remaining 18 percent difference in average per diem costs.

Many commenters attribute the difference in average per diem cost to the types of patients admitted to psychiatric units and psychiatric hospitals. The commenters stated that patients admitted to psychiatric units generally present with multiple medical conditions in addition to severe or multiple psychiatric symptoms. In addition, EDs in acute care hospitals with psychiatric units serve as the portal for almost all psychiatric emergency patients, who usually are admitted to the psychiatric unit. As a result, psychiatric units have different patterns of care and staffing in order to treat patients with emergency psychiatric needs as well as comorbid medical conditions.

The commenters stated that freestanding psychiatric hospitals are not equipped or staffed to treat patients with complex comorbid medical conditions and generally do not admit patients who require treatment of chronic physical illnesses or who are not medically stable. As a result, freestanding psychiatric hospitals have lower average per diem costs than psychiatric units.

Many commenters recommended that we provide a Medicare-dependent IPF designation that would be applied to any IPF with at least an 80 percent Medicare share of admissions. An organization representing small, rural IPFs provided information describing rural psychiatric units and the patients generally treated in these units. The commenter indicated that rural psychiatric units usually have 12 or fewer beds and treat a high proportion (at least 80 percent of total patient days) of Medicare beneficiaries. The material furnished by the organization indicated that approximately 54 percent of these hospitals are located in areas not adjacent to a metropolitan area and 15 percent are in “completely rural” areas.

The organization indicated that these small rural Medicare-dependent units generally have average costs per day that are 27 percent higher than the national average due to the acuity of the patients they serve. In addition, an analysis conducted by the organization indicates an 11.9 percent negative impact between current TEFRA payments and estimated payments under the proposed IPF PPS.

Commenters also indicated that many of the psychiatric units are small, Medicare-dependent, and located in underserved rural and urban areas where they are the sole mental health provider. These commenters were concerned that inadequate Medicare payment would cause hospitals to close these units, resulting in diminished access to mental health services. The commenters stated that the proposed adjustments were insufficient and requested a specific adjustment for psychiatric units or, as an alternative, a temporary adjustment until we are able to refine the IPF PPS and account for more of the difference in average per diem cost.

Response: As we discussed in the November 2003 proposed rule, we do not believe it is appropriate to pay an adjustment to all psychiatric units for all cases, regardless of the unit's cost, efficiency, or case-mix.

With respect to providing an adjustment for psychiatric units, as explained previously in this final rule, the payment model we are adopting for IPFs explains approximately 33 percent of the variation in per diem cost among IPFs. As a result, we believe the IPF PPS will generate payments that are reasonably related to the per diem cost in psychiatric units. In addition, IPFs located in rural areas will receive an adjustment to account for higher per diem costs.

Commenters stated that IPFs have many patients with longer stays or multiple co-morbidities. The IPF PPS provides a base payment amount and adjustments for each day of the stay and multiple co-morbidity categories as well as a variety of other adjustments, we believe IPF PPS payments to psychiatric units will adequate meet their costs.

In addition, we are providing a stop-loss provision during the 3-year transition period during which a stop-loss policy will be in place to ensure that small rural, Medicare-dependent, and urban psychiatric units get an IPF PPS payment amount that is no less than 70 percent of what they would have otherwise been paid under TEFRA had the IPF PPS not been implemented. This “safety net” will prevent an IPF from sustaining a significant financial “loss” by converting to the IPF PPS. Simultaneously, these providers will learn how to adjust their business structures efficiently under the IPF PPS framework. See section V.C. of this final rule.

b. Cost of Living Adjustment

i. IPFs Located in Alaska and Hawaii

As indicated in the proposed rule, we did not propose a cost-of-living adjustment (COLA) for IPFs located in Alaska and Hawaii. Based on the FY 1999 data, there were two psychiatric hospitals and no psychiatric units in Alaska and one psychiatric hospital and one psychiatric unit in Hawaii. Our analysis indicated that some IPFs in Alaska and Hawaii would “profit” from the proposed IPF PPS and other IPFs would experience a “loss.” Based on the limited number of cases in the analysis, we determined that the results were inconclusive and therefore we did not propose a COLA for IPFs located in Alaska and Hawaii.

We received several comments requesting a COLA for IPFs located in Alaska and Hawaii. In response to the public comments, we analyzed the FY 2002 data. The FY 2002 data, unlike the FY 1999 data, demonstrated that IPFs in Alaska and Hawaii had costs disproportionately higher than IPFs across the nation. In the absence of a COLA, IPFs located in Alaska and Hawaii would receive payments under the IPF PPS that were far below their cost. Thus, the results of our analysis conclusively demonstrate that a COLA for IPFs located in Alaska and Hawaii would improve payment equity for these facilities. As a result of this analysis, we are providing a COLA adjustment in this final IPF PPS based on the higher costs found in Alaska and Hawaii IPFs.

Comment: A few commenters recommended that CMS provide a facility-specific adjustment to the per diem payment amount to reflect the higher cost-of-living in Alaska.

One commenter recommended using the 25 percent Alaska COLA used under hospital IPPS for non-labor costs as a proxy adjustment for IPFs located in Alaska. The commenter stated that, despite the lack of IPF cases to study, CMS recognizes the need for a COLA adjustment for hospitals in Alaska under the hospital IPPS. The commenter indicated that MedPAC recently recommended that CMS provide an adjustment to the non-labor costs of skilled SNFs located in Alaska and Hawaii.

Response: As indicated above, we analyzed the cases in the FY 2002 data and found that there are two IPFs in Alaska and four in Hawaii. Based on our analysis of the FY 2002 stays for these IPFs, we find that a COLA adjustment is warranted. However, the small number of cases from each IPF would make development of a facility-specific adjustment erroneous because, with few cases, a small number of extremely high-cost or low-cost cases could easily overstate or understate the IPF's per diem cost. In general, the COLA would account for the higher costs in the IPF and will eliminate the projected loss that IPFs in Alaska and Hawaii would experience absent the COLA. We will make a COLA adjustment for IPFs located in Alaska and Hawaii by multiplying the non-labor share of the Federal per diem base rate by the applicable COLA factor based on the county in which the IPF is located. The COLA factors were obtained from the U.S. Office of Personnel Management and used in other PPS system. For the convenience of the reader, Table 8 below lists the specific COLA for Alaska and Hawaii IPFs.

TABLE 9—COLA Factors for Alaska and Hawaii IPFS

ii. IPFs located in California

Although we did not propose a cost-of-living adjustment for a specific State, we received a comment requesting that we provide an adjustment for California. We are not making a COLA to IPFs located in California as detailed below.

Comment: One comment recommended that CMS establish a facility-specific adjustment for psychiatric units located in California to reflect the higher resource costs associated with mandatory staffing ratios.

Response: Although recently imposed State staffing ratios would not be evident in the FY 2002 data, we analyzed the FY 2002 MedPAR data to assess whether IPFs located in California have higher per diem cost than IPFs located in other States. We determined that after adjustment for facility mix, IPF per diem costs in California are slightly higher (1.6 percent). While we did not assess the variation for each State, we acknowledge that every State will have some variation from the average cost per day under the IPF PPS. We do not believe the slightly higher per diem cost in California warrants a special adjustment. There may be laws in other States that could create a cost difference greater or lower than California and it is not practical to account for all of the cost differences in every State resulting from State and local laws.

c. Disproportionate Share Intensity

As indicated in the proposed rule, we did not propose an adjustment for disproportionate share hospital (DSH) status because the proposed regression analysis did not support an increase in payments. If we had proposed a payment adjustment for DSH facilities based on our empirical analysis, we would have proposed a reduction to the Federal per diem base rate paid to DSH facilities. Based on our analysis, we found a statistically significant negative relationship between per diem cost and DSH status. We did not believe that negative payment adjustment would be consistent with the intent of a DSH adjustment, which is intended to provide additional payments to providers to account for the costs of treating low-income patients. Therefore, we proposed no DSH adjustment.

We received numerous comments regarding the DSH adjustments. Most of the commenters disagreed with the proposed rule and stated that our reason for not providing a DSH adjustment was inadequate. A significant number of comments recommended that we re-examine the regression analysis and include a favorable DSH adjustment in the IPF PPS final rule. Based on the analysis discussed below, we are not providing a DSH adjustment in this final rule.

Comment: Several commenters stated that hospitals providing large amounts of care to low-income individuals often serve as key access points for low-income Medicare beneficiaries and other low-income patients requiring psychiatric care.

Response: In the proposed rule, we indicated that we would continue to monitor whether we could find empirical evidence to indicate a relationship between disproportionate patient percentages and higher per diem costs to support the establishment of a DSH adjustments. We re-examined our regression analysis, as commenters requested, but did not find any relationship between DSH intensity and higher per diem costs. Our analysis of the FY 2002 data yielded the same results as our analysis of the FY 1999. Therefore in this final rule we are not making a DSH adjustment.

Comment: One commenter stated that since CMS provided for a DSH adjustment in both the hospital IPPS and IRF PPS, IPFs should also receive this additional payment.

Another commenter indicated that the reluctance to allow psychiatric hospitals to participate in DSH payments is related to the belief that the DSH hospitals are low cost providers.

Response: Consistent with the approach we have taken in the proposed rule and in this final rule, we believe that any IPF PPS DSH payment adjustment should be supported by data showing that DSH facilities experience higher per diem costs than other IPFs. Our data failed to demonstrate that the IPFs who serve a disproportionate number of low income patients have higher per diem costs. Therefore, we do not see a justification to make a DSH adjustment in the IPF PPS. Unlike IPFs, the IPPS and IRF PPS had data supporting the need for a DSH adjustment. IPPS and IRF PPS data showed that serving a disproportionate share of low income patients has a direct connection to higher facility costs.

Comment: A commenter suggested that if government-operated hospitals bias the result, the analysis should be redone excluding those hospitals.

Response: We believe the commenter misunderstood our statements in the proposed rule about the impact of government-operated hospitals in our analysis. Our intention was not that the government-operated hospitals might be responsible for the finding of a negative relationship between per diem cost and the DSH variable. Instead, we were emphasizing that many observers might think that the limitations of measuring DSH for government-operated hospitals (too low a value for their DSH variable) might explain why we found higher DSH intensity associated with lower cost. However, our finding was not attributable to the government-operated hospitals because we found the same negative relationship when we excluded them from the regression.

Comment: Some commenters indicated that because Medicaid does not pay for services to certain individuals in an institution for mental diseases (IMD), low-income beneficiaries in psychiatric hospitals cannot be identified as Medicaid beneficiaries. In addition, the commenters believe that the Medicaid proportion will be biased downwards smaller than it should be.

Response: In the proposed rule and in this rule, the basis for the decision not to provide a DSH adjustment is our inability to find a correlation between available measures of low-income patient percentages and higher per diem costs. As previously indicated, potential measurement error in the Medicaid proportion did not explain the lack of a positive correlation between per diem cost and DSH status. We recognize that inpatients in institutions for mental diseases may still be eligible for Medicaid for purposes of the calculation of the DSH percentage (although there might be little incentive for facilities to establish a patient's Medicaid eligibility when there is no Medicaid payment available). The fact remains that, with currently available data, we found no basis for a DSH adjustment.

Comment: Several commenters asked how section 402 of the MMA would impact payments under the IPF PPS.

One commenter recommended that CMS wait until after December 8, 2004, to develop the IPF DSH factors (when the MMA is implemented and CMS begins to furnish DSH data to all hospitals). The commenter indicated that they expect the data to be a viable source of information that could be used to establish an appropriate DSH adjustment factor for the IPF PPS.

Response: Section 402 of the MMA has no effect on the IPF PPS as it only applies to DSH under the IPPS. The commenter is apparently referring to section 951 of the MMA, which requires that the Secretary arrange to furnish subsection (d) hospitals (those hospitals subject to the hospital IPPS) with the data necessary to compute the number of patient days used in computing the disproportionate patient percentage. We acknowledge that it is possible for this requirement to improve the accuracy of the disproportionate patient percentages for hospitals at some future point in time. However, we are making our decision not to include a DSH adjustment based on the best available data. If better data becomes available that indicates a need for a DSH adjustment, and an appropriate methodology for such an adjustment, the issue can be addressed in a future rulemaking.

d. IPFs With Full-Service Emergency Departments (EDs)

We did not propose an adjustment for IPFs with a qualifying ED. However, we received many comments requesting a facility adjustment for hospitals that maintain an ED and provide crisis management services. Several commenters recommended that IPFs with an ED should receive a facility-level adjustment empirically determined through the regression model. One commenter recommended a 20 percent adjustment factor for IPFs in hospitals with an ED.

In this final rule, we are providing an adjustment to the Federal per diem base rate to account for the costs associated with maintaining a full-service ED. We conducted an analysis, as described below, to develop an appropriate payment adjustment to account for ED costs and to define the subset of IPFs that have, or are part of acute care hospitals that have, a full-service ED.

The overhead costs associated with maintaining an ED are included in each IPF's routine cost amount, but since routine costs are reported as a average, we are unable to determine the portion of the routine cost directly attributable to ED costs. As an alternative, we analyzed cases admitted through the ED using FY 2002 claims data. ED cases were identified by the presence of ED or ambulance charges on the MedPAR record. We found that about one-third of all cases were admitted through the ED, and that 98 percent of the cases were treated in psychiatric units. Among the psychiatric hospitals and units with at least one admission from an ED, the ED admissions comprise about 43 percent of all admissions.

In analyzing the relative cost of ED and other admissions, we limited the comparison to IPFs with ED admissions to avoid attributing cost differences to ED admissions that are due to other unrelated factors. On a per case basis, ED admissions are actually slightly less expensive than other admissions ($7,672 versus $8,036). Most of the difference results from the fact that ED stays are about one day shorter than other psychiatric stays (10.6 days versus 11.5 days). The ED costs average about $198 per case, and the mean difference in ancillary costs per case (which includes ED costs) is about $196. Thus, the ED costs effectively account for all of the difference in ancillary costs per case between the ED and other admissions. On average, admissions through the ED do not appear to require any more ancillary services than other admissions except for the ED costs themselves.

Although this analysis indicated that patients admitted through the ED were more costly on a per diem basis than cases without an ED admission, we are not including an adjustment for patients admitted through the ED. As explained previously, we are concerned about creating an incentive for psychiatric units in acute care hospitals with EDs to inappropriately admit all psychiatric patients through the ED of the acute care hospital in which it is located in order to receive a patient-level ED adjustment. An ED adjustment at the patient level would be approximately $200. To the extent a psychiatric unit ensured that all of its patients were admitted for inpatient psychiatric care through the ED of the acute care hospital in which it is located, even though admission through the ED was unnecessary and inappropriate, Medicare would be substantially overpaying for these cases.

As an alternative, we have decided to provide a facility-level adjustment for IPFs, for both psychiatric hospitals and acute care hospitals with a distinct part psychiatric unit, that maintain a qualifying ED. We are providing the adjustment to psychiatric units in acute care hospitals because the costs of the ED are allocated to all hospital departments, including the psychiatric units. We intend that the adjustment only be provided to hospitals with EDs that are staffed and equipped to furnish a comprehensive array of emergency services and that meet the definition of a “dedicated emergency department” in § 489.24 and the definition of “provider-based entity” in § 413.65. We are defining a full-service ED in order to avoid providing an ED adjustment to an intake unit that is not comparable to a full-service ED with respect to the array of emergency services available or cost.

However, where a psychiatric unit would otherwise qualify for the ED adjustment, but an individual patient is discharged from that acute care hospital, we would not apply the ED adjustment. The reason we would not give an ED adjustment in this case is that the costs associated with maintaining the ED would have already been paid through the DRG payment paid to the acute care hospital. Thus, if we provided an ED adjustment in this case, the hospital would be paid twice for the overhead costs of the ED.

The ED adjustment will be incorporated into the variable per diem adjustment for the first day of each stay. That is, IPFs with qualifying EDs, will receive a higher variable per diem adjustment for the first day of each stay than will other IPFs.

Three steps were involved in the calculation of the ED adjustment factor. First, we estimated of the proportion by which the ED costs of a case would increase the cost of the first day of the stay. Using the IPFs with ED admissions in 2002, we divided their average ED cost per stay admitted through the ED ($198) by their average cost per day ($715), which equals 0.28. Second, we adjusted the factor estimated in step 1 to account for the fact that we will pay the higher first day adjustment for all cases in the qualifying IPFs, not just the cases admitted through the ED. Since on average, 44 percent of the cases in IPFs with ED admissions are admitted through the ED, we multiplied 0.28 by 0.44, which equals 0.12. Third, we added the adjusted factor calculated in the previous 2 steps to the variable per diem adjustment derived from the regression equation that we used to derive our other payment adjustment factors. The first day payment factor from this regression is 1.19. Adding the 0.12, we obtained a first day variable per diem adjustment for IPFs with a qualifying ED equal to 1.31.

D. Other Proposed Adjustments and Policy Changes

1. Outlier Policy

We proposed a 2 percent outlier policy to promote access to IPFs for those patients who require expensive care and to limit the financial risk of IPFs treating unusually costly cases. As explained in the proposed rule, we believe that it is appropriate to include an outlier policy in order to ensure that IPFs treating unusually costly cases do not incur substantial “losses” and promote access to care for patients requiring expensive care. Providing these additional payments to IPFs for costs that are beyond the IPF's control will also improve the accuracy of the payment system. Similar to the proposed rule, our payment simulations continue to support establishment of the outlier policy at 2 percent of total payments because it affords protection for vulnerable IPFs (and patients) while providing appropriate levels of payment for all other cases that are not outlier cases. The 2 percent target continues to provide an appropriate balance between patient access, IPF financial risk, and the payment rate reduction required for all cases to offset the cost of the policy.

We proposed to make outlier payments on a per case basis rather than on a per diem basis because it is the overall financial “gain” or “loss” of the case, and not of individual days, that determines an IPF's financial risk and, as a result, access for unusually costly cases. In addition, because patient level charges (from which costs are estimated) are typically aggregated for the entire IPF stay, they are not reported in a manner that would permit accurate accounting on a daily basis.

Thus, we proposed to make outlier payment for discharges in which estimated costs exceed an adjusted threshold amount ($4,200 multiplied by the IPF's facility adjustments, that is, wage area, rural location, teaching, and cost of living adjustment for IPFs located in Alaska and Hawaii) plus the total IPF adjusted payment amount for the stay. Where the case qualifies for an outlier payment, we proposed to pay 80 percent of the difference between the estimated IPF's cost for the case and the adjusted threshold amount for days 1 through 8 of the stay, and 60 percent of the difference for day 9 and thereafter. We established 80 percent and 60 percent to lost sharing ratios because we were concerned that a single ratio established at 80 percent (like other Medicare hospital prospective payment systems) might provide an incentive under the IPF per diem system to increase length of stay in order to receive additional payments. After establishing the ratios, we determined the threshold amount of $4,200 through payment simulations designed to compute a dollar loss beyond which payments are estimated to meet the 2 percent outlier spending target. In this final rule, we adopted this proposed outlier policy methodology, with an adjusted threshold amount of $5700. The revised amount is based on updated simulations using more recent data (from FY 2002) and the modified policy for the loss sharing ratios (see below).

In this final rule, we modified application of the loss-sharing provision of the outlier policy to pay 80 percent of the difference between the IPF's estimated cost for the case and the adjusted threshold amount for days 1 through 9 of the stay ( including median length of stay instead of days 1 through 8 up to the median length of stay) and 60 percent thereafter. As we explain above, we decided to reduce the 80 percent loss-sharing ratio by an additional 20 percent, resulting in a 60 percent loss sharing ratio for day 10 and thereafter. With this modification, we will pay 80 percent of the costs eligible for outlier payments for all cases whose length of stay is no greater than the median length of stay (9 days) of all Medicare inpatient psychiatric cases.

In the proposed rule, we proposed a number of policies to ensure the accuracy and integrity of our outlier payments. We are adopting these policies in this final rule, as decribed below.

Referring back to the payment calculation example in Section VI.B.2 of this final rule, the total estimated payment for the case is $7267.75. The adjusted threshold amount is calculated below:

Step 1: Multiply threshold by labor share and the wage area.

$5700 × 0.72528 (labor share) × 0.7743 (area wage index) = $3201.03

Step 2: Add this number to the non-labor share threshold amount.

$5700 × 0.27472 (non-labor share) = $1565.90

$1565.90 + $3201.03 = $4766.93

Step 3: Apply the other facility-level adjustments.

$4766.96 × 1.17 (rural adjustment) × 1.0 (teaching adjustment) = $5577.31

Step 4: Calculate the adjusted threshold amount by adding the estimated payment amount to the amount above.

$5577.31 + $7267.75 = $12,845.06

If estimated costs exceed the adjusted threshold amount ($12,845.06), then the case will qualify for an outlier payment. If the IPF in the example reports charges of $21,000 and they have a cost-to-charge ratio of 0.8, then the estimated cost of the case would be $16,800. The outlier amount is calculated below:

Step 1: Calculate the difference between the estimated cost and the adjusted threshold amount.

$16,800—$12.845.06 = $3954.94

Step 2: Divide by the length of stay (in our example, 10 days).

$3594.94 / 10 = $395.49

Step 3: For days 1 through 9 of the stay, the IPF receives 80% of this difference.

$395.49 × 0.80 = $316.40

$316.40 × 9 days = $2847.60

Step 4: For days 10 and beyond, the IPF receives 60% of the difference.

395 × 0.60 = $237.30 (in the example, the patient stays for 10 days, so the IPF receives the above amount for day 10 only).

Therefore, the IPF in the example would receive a total outlier payment of $3084.90.

($2847.60 + $237.30).

a. Statistical Accuracy of Cost-to-Charge Ratios

We believe that there is a need to ensure that the cost-to-charge ratio used to compute an IPF's estimated costs should be subject to a statistical measure of accuracy. Removing aberrant data from the calculation of outlier payments will allow us to enhance the extent to which outlier payments are equitably distributed and continue to reduce incentives for IPFs to under serve patients who require more costly care. Further, using a statistical measure of accuracy to address aberrant cost-to-charge ratios would also allow us to be consistent with the outlier policy under the hospital inpatient prospective payment system. Therefore, we are making the following two proposals:

• We will calculate two national ceilings, one for IPFs located in rural areas and one for facilities located in urban areas. We will compute the ceiling by first calculating the national average and the standard deviation of the cost-to-charge ratios for both urban and rural IPFs.

To determine the rural and urban ceilings, we will multiply each of the standard deviations by 3 and add the result to the appropriate national cost-to-charge ratio average (either rural or urban). We believe that the method explained above results in statistically valid ceilings. If an IPF's cost-to-charge ratio is above the applicable ceiling, the ratio is considered to be statistically inaccurate. Therefore, we will assign the national (either rural or urban) median cost-to-charge ratio to the IPF. Due to the small number of IPFs compared to the number of acute care hospitals, we believe that statewide averages used in the hospital inpatient prospective payment system, would not be statistically valid in the IPF context.

In addition, the distribution of cost-to-charge ratios for IPFs is not normally distributed and there is no limit to the upper ceiling of the ratio. For these reasons, the average value tends to be overstated due to the higher values on the upper tail of the distribution of cost-to-charge ratios. Therefore, we will use the national median by urban and rural type as the substitution value when the facility's actual cost-to-charge ratio is outside the trim values. Cost-to-charge ratios above this ceiling are probably due to faulty data reporting or entry, and, therefore, should not be used to identify and make payments for outlier cases because these data are clearly erroneous and should not be relied upon. In addition, we will update and announce the ceiling and averages using this methodology every year.

• We will not apply the applicable national median cost-to-charge ratio when an IPF's cost-to-charge ratio falls below a floor. We are adopting this policy because we believe IPFs could arbitrarily increase their charges in order to maximize outlier payments.

Even though this arbitrary increase in charges should result in a lower cost-to-charge ratio in the future (due to the lag time in cost report settlement), if we propose a floor on cost-to-charge ratios, we will apply the applicable national median for the IPFs actual cost-to-charge ratio. Using the national median cost-to-charge ratio in place of the provider's actual cost-to-charge ratio would estimate the IPF's costs higher than they actually are and may allow the IPF to inappropriately qualify for outlier payments.

Accordingly, we will apply the IPF's actual cost-to-charge ratio to determine the cost of the case rather than creating and applying a floor. In such cases as described above, applying an IPF's actual cost-to-charge ratio to charges in the future to determine the cost of the case will result in more appropriate outlier payments.

Consistent with the policy change under the hospital inpatient prospective payment system, IPFs will receive their actual cost-to-charge ratios no matter how low their ratios fall. We are still assessing the procedural changes that would be necessary to implement this change. For this final rule, we are finalizing the above described policies.

b. Adjustment of IPF Outlier Payments

As discussed in the hospital inpatient prospective payment system final rule for outliers, we have implemented changes to the IPPS outlier policy used to determine cost-to-charge ratios for acute care hospitals, because we became aware that payment vulnerabilities exist in the current outlier policy. Because we believe the IPF outlier payment methodology is likewise susceptible to the same payment vulnerabilities, we are adopting the following changes:

• Include in § 412.424(c)(2)(v) a cross-reference to § 412.84(i) that was included in the final rule published in the Federal Register on June 9, 2003 (68 FR 34515). Through this cross-reference, FIs will use more recent data when determining an IPF's cost-to-charge ratio. Specifically, as provided in § 412.84(i), FIs will use either the most recent settled IPF cost report or the most recent tentatively settled IPF cost report, whichever is later to obtain the applicable IPF cost-to-charge ratio. In addition, as provided under § 412.84(i), any reconciliation of outlier payments will be based on a ratio of costs to charges computed from the relevant cost report and charge data determined at the time the cost report coinciding with the discharge is settled.

Include in proposed § 412.424(c)(2)(v) a cross reference to § 412.84(m) (that was included in the final rule published in the Federal Register on June 9, 2003 (68 FR 34415) to revise the outlier policy under the hospital inpatient prospective payment system). Through this cross-reference, IPF outlier payments may be adjusted to account for the time value of money during the time period it was inappropriately held by the IPF as an “overpayment.” We also may adjust outlier payments for the time value of money for cases that are “underpaid” to the IPF. In these cases, the adjustment will result in additional payments to the IPF. Any adjustment will be based upon a widely available index to be established in advance by the Secretary, and will be applied from the midpoint of the cost reporting period to the date of reconciliation.

We received several comments on the proposed outlier policy. Most of the comments expressed support for the proposed outlier policy.

Comment: Many commenters indicated that the outlier level is too low and that there should be a mechanism to appeal an outlier payment. The commenters recommended establishing the outlier policy at 5 percent of the total IPF PPS.

Response: We are maintaining a 2 percent outlier policy in the final IPF PPS. The 2 percent outlier target percentage is lower than the target outlier percentage of other prospective payment systems that contain outlier polices, which range from 3 percent in the inpatient rehabilitation PPS to 8 percent in the LTCH PPS. The target outlier percentage in IPPS is about 5 percent. However, these other systems are per case or per episode payment systems in which Medicare's payment does not automatically account for the higher costs associated with longer lengths of stay. In a per diem system, such as the IPF PPS, there is less of a need for outlier payments because it automatically adjusts payments for length of stay. Therefore, we believe that 2 percent of total IPF PPS payment is appropriate. We estimate that approximately 5 percent of IPF cases would meet the fixed dollar loss threshold amount and qualify for an average outlier payment of $3,248.

If the provider is dissatisfied with the amount of payment, they can invoke existing appeal rights.

Comment: Several commenters recommended modifying the outlier calculation so that the proposed risk sharing percentage of 60 percent for the ninth and subsequent days is increased to 80 percent.

Response: We proposed to reduce the risk sharing percentage from 80 percent to 60 percent after the 8th day of the stay. The choice of the 8th day was based on the fact that a single variable per diem adjustment was proposed for days 5 through 8, and we thought it appropriate to make the change in the risk sharing percentage change coincide with the change in the variable per diem adjustment factor. After analyzing new data and based on public comments, we have revised the variable per diem adjustment factors so that they vary continuously over the first 22 days of the stay. As a result, there is no longer any reason to make the change in the risk sharing percentage coincide with the variable per diem adjustment factors. In this final rule, we are changing the risk sharing percentage from 80 percent to 60 percent after the 9th day of the stay. We chose to include the 9th day in the 80 percent risk sharing category because 9 days is the median length of stay. The median implies that one-half of the cases have a length of stay greater than 9 days, and the other half have a length of stay less than 9 days, which also can be interpreted as implying that the “typical” case has a length of stay of 9 days. We will pay the 80 percent risk sharing percentage for all cases whose length of stay is less than or equal to the length of stay of the typical case. We are reducing the risk sharing percentage for cases whose length of stay exceeds that of the typical case, because as we noted in the proposed rule (68 FR 66934), we are concerned that a single risk sharing percentage at 80 percent might provide an incentive to increase length of stay in order to received additional outlier payments. Reducing the amount Medicare shares in the loss of high cost cases provides an incentive for an IPF to contain costs once a case qualifies for outlier payments. The reduction from 80 percent to 60 percent is adequate to provide such an incentive, while maintaining a significant degree of risk sharing.

Comment: Many commenters requested that CMS provide additional information to the sample calculation presented in the proposed rule. The commenters also recommended that CMS explain the circumstances under which an outlier would be paid (interim billing or at the time of discharge).

Response: Since outlier payments will be made on a per-case basis, a determination as to whether a case qualifies for an outlier payment cannot be made until discharge. We are concerned about the potential for overpayments associated with IPF stays that may appear to qualify for outlier payments early in the stay, but do not meet the fixed dollar loss threshold once all costs and IPF PPS payments are considered. To avoid this situation, we proposed in § 412.432(d), that additional payments for outliers are not made on an interim basis. Rather, outlier payments are made based on the submission of a discharge bill. We are adopting this provision in this final rule.

Comment: Several commenters recommended clarification on the methodology for determining the cost-to-charge ratio, a clear definition of the numerator and denominator in the ratio, identifying the applicable worksheet location for data on costs and charges, as well as the appeal or comments that might be available when the national cost-to-charge ratios are published.

Response: We intend to follow similar procedures as outlined in the IPPS final rule published in the Federal Register on June 9, 2003 (68 FR 34498). IPF PPS outlier methodology requires the FI to calculate the provider's overall Medicare cost-to-charge ratio using the facility's latest settled cost report or tentatively settled cost report (whichever is from the later period), and associated data. Cost-to-charge ratios will be updated each time a subsequent cost report is settled or tentatively settled. Total Medicare charges will consist of the sum of inpatient routine charges and the sum of inpatient ancillary charges including capital. Total Medicare costs will consist of the sum of inpatient routine costs (net of private room differential and swing bed cost) plus the sum of ancillary costs plus capital-related pass-through cost only. Based on current Medicare cost reports and worksheet, specific FI instructions are described below.

For freestanding IPFs, Medicare charges will be obtained from Worksheet D-4, column 2, lines 25 through 30, plus line 103 from the cost report. For freestanding IPFS, total Medicare costs will be obtained from worksheet D-1, Part II, line 49 minus (Worksheet D, Part III, column 8, lines 25 through 30, plus Worksheet D, Part IV, column 7, line 101). Divide the Medicare costs by the Medicare charges to compute the cost-to-charge ratio.

For IPFs that are distinct part psychiatric units, total Medicare inpatient routine charges will be estimated by dividing Medicare routine costs on Worksheet D-1, Part II, line 41, by the result of Worksheet C, Part I, line 31, column 3 divided by line 31, column 6. Add this amount to Medicare ancillary charges on Worksheet D-4, column 2, line 103 to arrive at total Medicare charges. To calculate the total Medicare costs for distinct part units, data will be obtained from Worksheet D-1, Part II, line 49 minus (Worksheet D, part III, column 8, line 31 plus Worksheet D, Part IV, column 7, line 101). All references to Worksheet and specific line numbers should correspond with the subprovider identified as the IPF unit, that is the letter “S” is the third position of the Medicare provider number. Divide the total Medicare costs by the total Medicare charges to compute the cost-to-charge ratio.

If the provider is dissatisfied with the FI's cost-to-charge ratio determination, they can invoke their applicable appeal rights.

2. Interrupted Stays

In the proposed rule, we proposed an interrupted stay policy based on our concern that IPFs could maximize inappropriate Medicare payment by prematurely discharging patients after they receive the higher variable per diem adjustments and then readmitting the same patient. Under the proposed policy, if a patient is discharged from an IPF and returns to the same IPF before midnight on the fifth consecutive day following discharge, the case is considered to be continuous for applying the variable per diem adjustments and determining whether the case qualifies for outlier payments. Therefore, we would not apply the variable per diem adjustments for the second admission and would combine the costs of both admissions for the purpose of outlier payments. We proposed this policy in order to lower the incentive for a hospital to move patients among Medicare-covered sites in order to maximize Medicare payments. We received many public comments regarding the proposed interrupted stay policy. Most of the commenters requested that we delete the interrupted stay policy, provide an exception for discharges to an acute care hospital in order to receive medical or surgical services, for readmissions due to psychiatric decompensation, or shorten the duration of the interrupted stay policy. In this final rule, we are retaining the interrupted stay policy, but we are shortening the duration to 3 days.

Therefore, if a patient is discharged from an IPF and admitted to any IPF within 3 consecutive days of the discharge from the original IPF stay, the stay would be treated as continuous for purposes of the variable per diem adjustment and any applicable outlier payment.

For example a patient is discharged from an IPF on March 10 after an initial stay of 7 days and is admitted to another IPF on March 12 (before midnight of the 3rd consecutive day). The “readmission” is considered a continuation of the initial stay. Therefore day 1 of the readmission will be considered day 8 of the combined stay for purposes of the variable per diem stay and any applicable outlier payment.

Comment: A few commenters stated that after a 5-day interruption, the patient would need a full workup similar to the admission process on the first day. One commenter stated that the proposed 5-day interrupted stay policy financially penalizes IPFs for ensuring that their patients receive necessary emergency medical care.

Most commenters requested that we shorten the duration of the interrupted stay policy. Other commenters stated that a 5-day interrupted stay policy would require IPFs to hold claims and not bill Medicare until after the fifth day of discharge and that a 5-day interrupted stay policy could cause IPFs to delay readmissions to avoid the policy.

Several commenters recommended that we reduce the duration of the interrupted stay policy to 3 days to coincide with the 72-hour rule for bundling of outpatient charges under IPPS. Other commenters suggested a 3-day interrupted stay policy in order to be consistent with the interrupted stay policy in the IRF prospective payment system. However, a few commenters suggested that we extend the interrupted stay policy to readmissions to the IPF within 15 or 30 days of the patients discharge that would prompt a readmission review by the hospital's Quality Improvement Organization.

Response: In the proposed rule, we indicated that an absence from the IPF of less than 5 days would not necessitate repeating many of the admission-related services such as psychiatric evaluations and the patient's medical history. After receiving public comments we reanalyzed the duration of the interrupted stay policy. We now agree that after a 5-day absence from the IPF there are psychiatric and laboratory tests that would need to be repeated. As a result, we have revised the duration of the interrupted stay policy in this final rule from 5 days to 3 days.

Comment: Several commenters did not believe an interrupted stay policy was necessary to avoid inappropriate transfers and readmissions to the IPF. One commenter stated that adequate safeguards already exist, such as the physician certification and recertification requirements, significant medical malpractice risk of premature discharge, periodic review of practice patterns by local licensing and national accreditation bodies, and FI audits.

Response: Despite the safeguards identified by the commenters, inappropriate transfers and readmissions of psychiatric patients continue to occur. For this reason, we continue to believe an interrupted stay policy is necessary to discourage inappropriate discharges and readmissions to IPFs.

Comment: The majority of commenters requested that we provide an exception to the interrupted stay policy when a patient is discharged to an acute care hospital for medical care. The commenters maintain that the resources required to treat the patient at the time of readmission are of similar intensity to those required at the point of first admission. All assessments (including history and physical and psychiatric assessment) as well as the comprehensive treatment plan need to be reviewed and revised. In addition, the medical condition that required treatment must be addressed and incorporated into the ongoing treatment. One commenter suggested that discharges and subsequent readmissions to the IPF due to psychiatric decompensation should not be subject to the interrupted stay policy as well.

Response: Although we agree that some additional resources will be expended by IPFs when a patient is readmitted, we believe the resources required to reassess a patient upon readmission would be greatly reduced after a 3-day interrupted stay compared to the proposed 5-day interrupted stay policy. In addition, since almost three fourths of IPFs are distinct part psychiatric units in acute care hospitals, we remain concerned about hospitals inappropriately shifting patients between the psychiatric unit and the medical unit, thus receiving both the full DRG payment for the admission to the acute care hospital, and IPF payment for the admission to the excluded psychiatric unit.

Comment: One commenter asked if the interrupted stay policy applies if a patient is discharged to receive acute care and is readmitted to a different IPF than the IPF that originally discharged and transferred the patient. The commenter indicated that the shuffling of psychiatric patients from hospital to hospital is an abusive practice that the interrupted stay policy should address.

Response: We share the commenter's concern about the “shuffling” of psychiatric patients from hospital to hospital. We believe adopting an interrupted stay policy will address this concern from the viewpoint of the IPF PPS.

One example is when a patient is discharged from a psychiatric unit to receive acute care and discharged at the completion of the hospital IPPS stay, then transferred to a freestanding psychiatric hospital rather than returning to the psychiatric unit. Under the interrupted stay policy, if the readmission to the psychiatric hospital occurs within the 3-day interrupted stay timeframe, of the initial psychiatric unit stay, we would not pay the psychiatric hospital the variable per diem adjustments for the initial days of the original psychiatric unit stay otherwise applicable to the stay. The transferring hospital would send the psychiatric hospital the patient's medical record that will include information regarding the prior psychiatric stay in accordance with the hospital condition of participation for discharge planning (§ 482.43).

As a result, we have revised § 412.424(d) to clarify that if a patient is discharged from an IPF and is readmitted to the same or another IPF before midnight on the third consecutive day following the discharge from the original IPF stay, the case is considered to be continuous for applying the variable per diem adjustments and determining whether the case qualifies for outlier payments.

Comment: Several commenters asked if the interrupted stay policy would apply if a patient is transferred from a distinct part psychiatric unit to the hospital's medical unit and is readmitted to the IPF within the 5-day interrupted stay timeframe, but with a different principal diagnosis.

Response: In the situation described by the commenter, the interrupted stay policy would apply. A psychiatric patient whose illness is severe enough to require inpatient psychiatric treatment, should be receiving care for all of their psychiatric conditions. Therefore, if this psychiatric patient was discharged for acute medical care, and upon discharge from the acute medical hospital the patient still required inpatient psychiatric treatment, that treatment should be considered a continuation of the original stay. Thus, the principal diagnosis upon readmission is not relevant to the interrupted stay policy.

Comment: One commenter asked if the interrupted stay policy would apply when a patient is discharged to a partial hospitalization program, decompensates while in that program, necessitating a readmission to the IPF within 5 days of the discharge from the IPF.

Response: Under this final rule, if a patient was in an IPF and was discharged to a partial hospitalization program but then required readmission to an IPF within the 3-day timeframe, the stay is considered an interrupted stay. The interrupted stay policy applies to all discharges and subsequent readmissions to an IPF within 3 consecutive days.

3. Stop-Loss Provision

Many commenters who believed that they would be disadvantaged by implementation of the IPF PPS, requested that we provide additional payments through a risk sharing arrangement. We considered alternatives that would reduce financial risk to facilities expected to experience substantial reductions in Medicare payments during the period of transition to the IPF PPS.

Specifically, we considered stop-loss policies that would guarantee each facility, total IPF PPS payments no less than a minimum percent of its TEFRA payments, had the IPF PPS not been implemented. The two values for the minimum percent of TEFRA payments we examined were 70 percent and 80 percent. The 80 percent option was considered because 80 percent is a commonly used rate of risk-sharing in Medicare programs. We pay 80 percent of the estimated costs of outlier cases beyond the outlier threshold, and 80 percent is similarly used in other Medicare PPS's, as well as in many other insurance arrangements. The 70 percent option was assessed as an alternative, because it more narrowly targets stop-loss payments to facilities with greater financial risk.

Each of these policies was applied to the IPF PPS portion of Medicare payments during the transition. Hence, during year 1, three-quarters of the payment would be based on TEFRA and one-quarter on the IPF PPS. In year 2, one-half of the payment would be based on TEFRA and one-half on the IPF PPS. In year 3, one-quarter of the payment would be based on TEFRA and three-quarters on the IPF PPS. In year 4 of the IPF PPS, Medicare payments are based 100 percent on the IPF PPS.

The combined effects of the transition and the stop-loss policies would be to ensure that the total estimated IPF PPS payments would be no less than 92.5 or 95 percent in year 1, 85 or 90 percent in year 2, and 77.5 or 85 percent in year 3, depending upon whether the 70 percent or the 80 percent stop-loss option were implemented. Under the 70 percent policy, 75 percent of total payment would be TEFRA payments, and the 25 percent would be IPF PPS payments, which would be guaranteed to be at least 70 percent of the TEFRA payments. The resulting 92.5 percent of TEFRA payments is the sum of 75 percent and 25 percent times 70 percent (which equals 17.5 percent).

The 70 percent of TEFRA payment stop-loss policy would require a reduction in the Federal per diem and ECT base rates of 0.39 percent in order to make the stop-loss payments budget neutral. We estimate that about 10 percent of IPFs would receive stop-loss payments under the 70 percent policy.

The 80 percent of TEFRA stop-loss policy would require a reduction in the Federal per diem rate of almost 2 percent in order to make the stop-loss policy budget neutral. We estimate that almost 27 percent of all facilities would receive additional payments under the 80 percent stop-loss policy.

We also considered a risk-sharing policy modeled on the same principles as the case-level outlier policy, but applied at the facility level. Under this approach, we considered the case in which an IPF would have to incur a 12 percent loss in IPF PPS payments relative to TEFRA and then we would pay 80 percent of additional losses. This approach was estimated to require a reduction in the Federal per diem and ECT base rates of about 12 percent.

In order to target the stop-loss policy to the IPFs that may experience the greatest impact relative to current payments and to limit the size of the reductions to the Federal per diem and ECT base rates required to maintain budget neutrality, we are adopting the 70 percent stop-loss provision. We have added a new paragraph (d) to § 412.426 to include the 70 percent stop-loss provision as part of the 3-year transition to the IPF PPS. We will monitor expenditures under this policy to evaluate its effectiveness in targeting stop-loss payments to IPFs facing the greatest financial risk.

4. Physician Recertification Requirements

In the proposed rule, we proposed to modify the timing of the first physician recertification after admission to the IPF. We proposed to revise § 424.14(d) to require that a physician recertify a patient's continued need for inpatient psychiatric care on the tenth day following admission to the IPF rather than the 18th day following admission to the IPF.

Also, we proposed to amend § 424.14 by adding a new paragraph (c)(3) to require that, in recertifying a patient's need for continued inpatient care, a physician must indicate that the patient continues to need, on a daily basis, inpatient psychiatric care (furnished directly by or requiring the supervision of IPF personnel) or other professional services that, as a practical matter, can be provided only on an inpatient basis. We received a few comments supporting the proposed change. However, most of the commenters did not support the proposed changes and indicated inconsistencies in the timeframes currently required for IPFs that warrant additional analysis. As a result, we are not including the proposed physician re-certification requirements in this final rule. We will continue to require that a physician recertify a patient's continued need for inpatient psychiatric care on the 18th day following admission to the IPF.

VII. Implementation of the IPF PPS

A. Transition Period

1. Existing Providers

We proposed a 3-year transition period during which IPFs would receive a blended payment of the Federal per diem payment amount and the facility-specific payment amount the IPF would receive under the TEFRA payment methodology. We proposed that the first year of the transition would be 15 months. Thus the first year of transition is for cost reporting periods beginning on or after April 1, 2004 and before July 1, 2005. The proposed total payment for this period would consist of 75 percent based on the TEFRA payment system and 25 percent based on the proposed IPF prospective payment amount.

We also proposed that for cost reporting periods beginning on or after July 1, 2005 and before July 1, 2006, the total payment would consist of 50 percent based on the TEFRA payment system, and 50 percent based on the proposed IPF prospective payment amount. In addition, we also proposed that for cost reporting periods beginning on or after July 1, 2006 and before July 1, 2007, the total payment would consist of 25 percent based on the TEFRA payment system and 75 percent based on the proposed IPF prospective payment amount. Thus, we proposed that payments to IPFs would be at 100 percent of the proposed IPF prospective payment amount for cost reporting periods beginning on or after July 1, 2007.

We proposed this transition period so existing IPFs would have time to adjust their cost structures and integrate the effects of changing to the IPF PPS payment system. We specified that we would not allow IPFs the option to be paid at 100 percent of the IPF PPS payment amount in the first year of the transition, but would require all IPFs to receive the blended IPF payments during the 3-year transition period.

However, new IPFs would be paid the full Federal per diem payment amount rather than a blended payment amount. This is because the transition period is intended to provide currently existing IPFs time to adjust to payment under the new system. A new IPF would not have received payment under TEFRA for delivery of IPF services before the effective date of the IPF PPS. Therefore, we believe new IPFs do not need a transition to adjust their operating or capital financing that IPFs that have been paid under the TEFRA payment methodology would need.

In the proposed rule (68 FR 66920), we defined new IPFs as those IPFs that, under current or previous ownership or both, have their first cost reporting period as an IPF beginning on or after April 1, 2004. In this final rule, we define a new provider as those IPFs that, under current or previous ownership or both, have their first cost reporting period as an IPF beginning on or after January 1, 2005 to coincide with the effective date of the final IPF PPS.

Comment: The majority of commenters requested that we provide an option for IPFs to forego the transition and be paid at 100 percent of the IPF PPS payment amount in the first year of the transition. The commenters stated that other PPSs, specifically IRF PPS and LTCH PPS, included that option.

The commenters also stated that a mandatory transition period causes IPFs to continue to be paid under the outdated TEFRA payment system. The commenters requested that IPFs that are substantially underpaid under TEFRA or those that would be last to begin the transition to the IPF PPS because of the timing of their cost reporting year should be permitted to receive 100 percent of the Federal per diem payment amount.

One commenter stated that failure to provide for a 100 percent IPF PPS payment option disadvantages efficient providers. The commenter indicated IPFs that choose this option would strive to become more cost efficient more quickly. In addition, the blended payment methodology during the transition period could lead to payments that are less than current cost-based payments and would penalize IPFs that have a low TEFRA rate. Several commenters indicated that a 100 percent IPF PPS payment option would avoid the complications and financial burden of a blended payment process due to accounting difficulties caused by being paid under two payment systems.

One commenter indicated that the protection offered by the transition is short-lived and that psychiatric units suffering the greatest losses will experience significant financial hardship until the IPF PPS is refined to account for more of the variation in the per diem costs of psychiatric units and psychiatric hospitals.

Another commenter indicated that hospitals would be unable to offset Medicare “losses” under the IPF PPS with gains in other services. The commenter indicated that it would be very difficult for many of these hospitals to support “losses” in their psychiatric units for the long term and that some hospitals may decide to close their psychiatric units, which would result in diminished access for beneficiaries.

However, several commenters specifically requested that CMS retain the proposed 3-year transition period. The commenters stated that the IPF PPS could have unexpected financial consequences for IPFs and the full transition period is needed to enable IPFs to adapt to the new payment system. The commenters are concerned that allowing immediate implementation of the IPF PPS would dilute the Federal per diem base rate and exacerbate the redistributive effect of the new payment system. Several commenters indicated that the availability of new funding, a 100 percent of the Federal per diem payment amount option would result in further reductions to the Federal per diem base rate. As a result, these commenters would support a 100 percent option, but only if there is new funding available.

Other commenters requested that CMS phase-in the new IPF PPS more slowly, to allow corrections to any serious errors in the IPF PPS before full implementaion. Commenters recommended that CMS lengthen the transition to 5 or 6 years and perhaps for as long as 10 years to enable CMS to refine the IPF PPS before the full implementation.

Response: We have retained the transition period in the final IPF PPS. We believe this approach strikes an appropriate balance between IPFs that are prepared immediately to move to full implementation of the IPF PPS and those IPFs that need time to make the changes before the full implementation of the new PPS.

Section 305(b)(10)(c) of BIPA allowed IRFs to elect to be paid 100 percent of the adjusted facility Federal prospective payment for each cost reporting period to which the blended payment methodology would other wise have been applied. In implementing LTCHs 5-year transition period of the PPS, one of the goals was to transition hospitals to full prospective payments as soon as appropriate. Due to the longer length of the transition period, under the LTCH PPS, we allowed LTCHs to elect payment based on 100 percent of the Federal rate at the start of any of its cost reporting periods during the 5-year transition period. Once the election to be paid 100 percent of the Federal per diem base rate was made, the LTCH was not able to revert to the transition blend.

The IPF statute does not mandate that IPFs be given the option to elect to be paid 100 percent of IPF PPS payment amount immediately Federal rate. The shorter timeframe of a 3-year transition period was to provide all IPFs adequate time to make the most prudent adjustments to their operations and capital financing to secure the maximum benefits of the new PPS.

Absent the availability of additional funds, the reallocation of existing funds in budget neutral payment systems cause shifts in facility payments. The aim of having an IPF PPS payment amount that is a blend of an ever-decreasing TEFRA portion and ever increasing IPF PPS portion is to mitigate dramatic negative effects of converting too quickly to a new payment system. Every budget neutral payment system will impact different provider groups differently. Some providers believe that they will “gain” under the new IPF PPS while others believe they will do less well compared to the payments they have received under TEFRA.

To provide the impartial treatment to all IPFs, in the final IPF PPS, we have required all IPFs to participate in the 3-year transition period. Therefore, prolonging the transitional period to 5 or 10 years would not help providers who believe they have been disadvantaged under TEFRA as well as those who feel they are not being helped under IPF PPS for a an even longer period of time.

However, we share the commenter's concern about the ability of IPFs to adjust to the IPF PPS so that access to inpatient mental health care is maintained. Thus, we have tried to ensure continued access to mental health care by accounting for the complexity of patients with concurrent psychiatric and medical health conditions. We have created a PPS with numerous patient and facility level adjustments, an outlier policy, as well as a stop-loss policy that when used in combination with the transition period should ensure that an IPF PPS payment adequately reflects the costs of furnishing inpatient psychiatric care to Medicare beneficiaries.

2. New Providers

We proposed a definition of a new IPF because new IPFs will not participate in the 3-year transition from cost-based reimbursement under TEFRA to the IPF PPS. The transition period is intended to provide existing IPFs time to adjust to payment under the IPF PPS. A new IPF would not have received payment under TEFRA for the delivery of IPF services before the effective date of the IPF PPS. Therefore, we do not believe that new IPFs require a transition period in order to make adjustments to their operating and capital financing, as will IPFs that have been paid under TEFRA, or need to otherwise integrate the effects of changing from one payment system to another payment system.

For purposes of applying the IPF PPS 3-year transition period, we proposed to define a new IPF as a provider of inpatient hospital psychiatric services that otherwise meets the qualifying criteria for IPFs, set forth in § 412.22, § 412.23, § 412.25, and § 412.27 under present or previous ownership (or both), and its first cost reporting period as an IPF begins on or after April 1, 2004, the effective date of the proposed IPF PPS. In this final rule, we are finalizing the definition, except we are replacing April 1, 2004 with January 1, 2005 in order to account for the revised effective date of the final IPF PPS. In other words, we are finalizing the definition of a new IPF as a provider of inpatient hospital psychiatric services that otherwise meets the qualifying criteria for IPFs, set forth in § 412.22, § 412.23, § 412.25, and § 412.27 under present or previous ownership (or both), and its first cost reporting period as an IPF begins on or after January 1, 2005.

B. Claims Processing

We proposed to continue processing claims in a manner similar to the current claims processing system. Hospitals would continue to report diagnostic information on the claim form and the FIs would continue to enter clinical and demographic information in their claims processing systems for review by the Medicare Code Editor (MCE).

Comment: We received a variety of comments from all-inclusive rate and nominal cost hospitals regarding specific billing issues.

Response: We are issuing operational instructions to address the specific billing issues raised by the commenters.

C. Annual Update

In the proposed rule, we indicated that section 124 of Public Law 106-113 does not specify an update strategy for the IPF PPS and is broadly written to give the Secretary discretion in proposing an update methodology. Therefore, we reviewed the update approach used in other hospital prospective payment systems (specifically, the IRF and LTCH PPS update methodologies).

As a result of this analysis, we proposed the following strategy for updating the IPF PPS: (1) use the FY 2000 bills and cost report data and the most current ICD-9-CM codes and DRGs when we issue the IPF prospective payment system final rule; (2) implement the system effective for cost reporting periods beginning on or after April 1, 2004; and (3) update the Federal per diem base rate on July 1, 2005, since a July 1 update coincides with more hospital cost reporting cycles and would be administratively easier to manage. As a result, the implementation period for the proposed IPF PPS was the 15-month period April 1, 2004 to June 30, 2005.

In this final rule, we calculated the final Federal per diem base rate to be budget neutral during the implementation period of the final IPF PPS. As in the proposed rule, for future updates, we will use a July 1 through June 30 annual update cycle. Similar to the proposed rule, we will not update the IPF PPS during the first year of implementation because we believe there would be an insufficient amount of time under the IPF PPS to generate data useful in updating the system. Thus, the implementation period for the final IPF PPS is the 18-month period January 1, 2005 through June 30, 2006. As a result, the first update to the IPF PPS will occur on July 1, 2006, and updated for each subsequent 12-month period thereafter.

As we noted in the proposed rule, we believe it is important to delay updating the adjustment factors derived from the regression analysis until we have IPF PPS data that includes as much information as possible regarding the patient-level characteristics of the population that each IPF serves. For this reason, we do not intend to update the regression and recalculate the Federal per diem base rate until we have analyzed one complete year of data under the IPF PPS. Until that analysis is complete, we proposed to publish a notice in the Federal Register each spring to update the IPF PPS and identified the various elements of the IPF PPS that we would update.

In this final rule, we are adopting the proposed annual update with minor modifications to reflect the policies contained in this final rule. For example, we did not include an adjustment for ECT in the proposed rule and as a result, the proposed update strategy did not address how we would update that payment amount.

We will publish a notice in the spring of CY 2006 to update the IPF PPS effective July 1, 2006 and will publish a update notice for each 12-month period thereafter. In the notice, we will:

• Update the Federal per diem base rate using the excluded hospital with capital market basket increase in order to reflect the price of goods and services used by IPFs.
• Apply the best available hospital wage index with an adjustment factor to the Federal per diem base rate to ensure that aggregate payments to IPFs are not affected by an updated wage index.
• Update the fixed dollar loss threshold to maintain an outlier policy that is 2 percent of total estimated IPF PPS payments.
• Describe relevant ICD-9-CM coding and DRG classification changes discussed in the IPPS that would affect IPF PPS coding and payment.
• Update the payment amount for ECT based on the best available OPPS data.

Finally, as we indicated in the proposed rule, we may propose an update methodology for the IPF PPS in the future. We anticipate that the update methodology would be based on the excluded hospital with capital market basket index along with other appropriate factors relevant to psychiatric service delivery such as productivity, intensity, new technology, and changes in practice patterns.

Comment: Several commenters requested that we delay the proposed April 1, 2004 implementation date until October 1, 2004 in order to be consistent with the October 1 update cycle for the IPPS. The commenters believe that an October 1 update cycle for the IPF PPS would avoid confusion and coding errors that would occur because of the introduction of ICD-9-CM and DRG changes mid-cycle. In addition, the commenters believe adopting an update cycle consistent with the IPPS would facilitate cost efficiency by also allowing educational efforts for coding and DRG changes to occur once per year.

Response: While we appreciate the commenter's concerns, it is important that CMS retain the flexibility to develop administratively feasible update schedules for the various prospective payment systems that must be updated annually. Therefore, we are retaining a July 1 through June 30 cycle for annual updating of the IPF PPS.

Comment: A few commenters requested clarification regarding the timing of implementation since hospitals have different cost reporting year start dates.

Response: IPFs will begin the first transition year of the IPF PPS at the beginning of their next cost reporting period after January 1, 2005. For example, if an IPF's cost reporting year begins on March 1, the IPF would begin to receive a blended payment amount consisting of 75 percent based on TEFRA payments and 25 percent based on IPF PPS payments for all discharges that occur after March 1, 2005.

VIII. Future Refinements

In the proposed rule, we described research efforts by RTI International^® and the University of Michigan that were underway at the time the proposed rule was published. Section VI. of this final rule describes the outcome of the RTI International^® project to study modes of practice and patient characteristics to analyze the components of the routine cost category of the Medicare cost report.

The University of Michigan project would assist us in developing a patient classification system based on a standard assessment tool, the Case Mix Assessment Tool (CMAT). We attached a draft of the assessment tool and explained that it had not been submitted to the Office of Management and Budget (OMB) for review in order to obtain approval to pilot test the draft assessment tool. We indicated that a public comment period would be available as part of the OMB review process.

We received multiple comments on the CMAT instrument.

Most of the comments received focused on the overall content of the instrument. There were several commenters that opposed the potential implements of the instrument.

Comment: One commenter indicated that CMAT appeared to address the primary diagnostic needs of the mentally ill, but fell short on the collection of information on functional status. The commenters recommended that variables be added to CMAT instrument to collect information on social integration and the recreational use of time. The commenter also indicated that it was not clear how the functionality section would affect payment. Other commenters recommended that the instrument be revised to capture better information on patient conditions and resources needed to provide care. One commenter indicated that while the CMAT, as proposed, was an excellent tool for describing psychiatric signs and symptoms, it fails to assess active comorbid medical conditions. Another commenter recommended that the CMAT instrument be expanded to collect information on the use of seclusion and restraints. Another commenter also indicated that the CMAT should contain sections that specifically address the assessment reference date, common observational periods, and multiaxial assessments.

Response: We are aware that the current draft CMAT instrument would not collect extensive information on patient conditions and comorbid conditions. However, if the instrument is pilot tested, and ultimately fielded for refinement purposes, we are planning to match the CMAT with CMS administrative files. This comparison will augment the collection capacity of the CMAT and provide detailed information of medical conditions. The draft CMAT instrument, which has not been proposed, is currently undergoing OMB review. Following this review, the instrument is to be pilot tested. The variables suggested in these comments (for example, seclusion and restraints, assessment dates, observational periods, and multi-axial assessments) are being evaluated for potential inclusion in the pilot test.

Comment: One commenter recommended that because the CMAT is controversial, any pilot test findings should be made available to the public.

Response: The results of the pilot test will be made available to the public. We plan to test the feasibility of administration, reliability and validity of the instrument, and recommendations regarding potential modifications to the draft CMAT. A report from the pilot test will be available, and CMS will use this report and experience garnered from the pilot test to determine next steps for the instrument. We will then decide whether to propose the use of the CMAT instrument to assist us in developing a patient classification system.

Comment: Several commenters expressed support for development of a standardized instrument to collect patient level information to augment CMS administrative data. One commenter stated that the costs for an instrument would be outweighed by the benefits of creating a tool that collects information on patient conditions and necessary resources, so long as the tool is easy to use and complete.

Another commenter was pleased with the development of the CMAT and indicated that only when information from the refined variables in CMAT are available would it be appropriate to implement the IPF PPS.

Response: We will implement the IPF PPS before the CMAT is pilot tested because once the instrument has been pilot tested and the instrument reflects changes resulting from the testing, the instrument will have to be cleared by the Office of Management and Budget (OMB). We do not want to further delay implementation of the IPF PPS while the CMAT is tested and approved. However, a detailed OMB information collection package will be prepared and made available to the public.

In addition, there are a number of steps that are necessary to insure that assessment instruments collect the most useful information. Pending the pilot test results and a national fielding of the CMAT instrument following the pilot test, and OMB clearance of a final instrument, we would potentially use these variables to propose future refinements to the IPF PPS.

Comment: Many of the comments focused on the burden associated with completion of the CMAT instrument. Commenters stated that completion of the CMAT instrument for each discharged patient would require additional staff. The commenters recommended that CMS consider providing an adjustment to the Federal per diem base rate payment amount for the additional staff resources that would be required to complete the CMAT instrument.

One commenter indicated that IPFs are already faced with funding and management challenges and should not be asked to allocate resources away from direct patient care to fulfill a reporting requirement.

Response: The CMAT instrument and supporting materials is currently undergoing OMB review for potential fielding of the pilot test. One of the considerations of OMB review is to assess the potential burden on providers to complete the pilot test. One of the areas that will be assessed in administering the pilot test is the direct burden on the facilities to complete the instrument. CMS will assess the results of the pilot test to determine the feasibility of administering this instrument on a national basis, and the overall resources required to complete the instrument.

If the pilot test is implemented, we have proposed approaches that could lessen the burden for administration, such as, automation of the instrument. In addition, we would allow the treatment team members providing patient care to complete the form, rather than to request that only nurses complete the form. CMS will monitor the experience in administering the form throughout the pilot test. Finally, the report on the pilot test will address the burden on staff of completing the CMAT instrument.

Comment: One commenter indicated that the CMAT instrument, as currently drafted, would collect excessive and duplicative (to the medical record) information. Other commenters stressed that the instrument was time-consuming to complete and the potential use of the information proposed for collection was not clear. These commenters indicated that the relationship of the proposed data collection to case mix and reimbursement was not described.

Some commenters referred to their experiences in implementing the assessment instruments currently in use for SNFs and IRFs, and indicated that the instruments used in those payment systems do not adequately collect information on the resources needed to provide patient care.

One commenter recommended that all research regarding the development of the CMAT instrument cease. Another commenter indicated that the tool, as currently drafted, requested superfluous data with too many gameable variables. Commenters also indicated that collection of the information contained on the CMAT instrument was not necessary for refinement purposes. Instead, they recommended expanding the variables that are collected as part of either the cost reports or the claims.

Response: We are aware that some of the variables proposed to be pilot tested in the draft CMAT instrument (which we did not propose to use in the proposed IPF PPS) may appear to be duplicative of the medical record. The availability in the medical record of the potential variables to be collected by the CMAT instrument are expected to facilitate the completion of the instrument and reduce completion time.

The number of steps to pilot test and implement an instrument on a national basis are many. When data is available on a national basis, we will be in a better position to test the predictability and usefulness of the variables and determine whether its use should be proposed as a refinement to the IPF PPS.

We are aware of the option of adding variables to the cost reports or claims. We have explored this option in developing other payment systems. Pending decisions on the implementation of the pilot test, we will explore either supplementing material from the CMAT or collecting stand alone variables using the cost reports or claims. In addition, we disagree with the commenters that suggest research for the development of the CMAT cease. Not only might continued development of the CMAT provide possible new useful information on patient resource needs and staffing utilization, it might ascertain whether our case mix is correct or need refinements. Furthermore, we believe the best way to ensure that our IPF PPS continues to be an adequate payment system is to continue research on all fronts so that we have the best available information to us when we must make policy decisions.

Comment: Commenters raised concerns regarding the limitation of the draft CMAT instrument for collecting staffing information.

Response: We note that other CMS research studies are currently working towards providing information on staffing resources needed to provide patient care. We will review the findings from the studies and consider incorporating them in any proposed refinements to the IPF PPS.

Comment: A few commenters recommended that CMS engage in additional research to acquire a greater understanding of the payment dynamics between comorbidities and resource utilization before implementing the IPF PPS.

Many commenters suggested that further analysis is needed to explain the difference in average per diem costs between psychiatric units and freestanding psychiatric hospitals. One commenter suggested an approach that would mirror a swing-bed methodology for patients needing both psychiatric and non-psychiatric inpatient services.

Response: Additional research is planned that will address many outstanding questions regarding differences among IPFs, unit characteristics, patient characteristics, discharge and transfer criteria, and economic incentives.

The current research agenda includes a project to assess the relationship between facilities that have scatter bed and organized DRG units and the IPF PPS. In addition, this research project will examine the role played by smaller psychiatric inpatient units and facilities, the continued use of partial hospitalizations and outpatient programs and their role in complementing and substituting for inpatient care. This project will further monitor the relationship between the IPF PPS, the OPPS, and IPPS payment systems over time.

Comment: One commenter indicated that if there was any future research in support of the IPF PPS it should focus only on costs and payment, and build off existing facility and payment variables. The commenter did not support the creation of a new set of variables requiring additional data collection unless there was evidence that it would dramatically increase the predictability of the models. The commenter recommended research that focused on mode of practice and staffing patterns across different types of inpatient psychiatric facilities.

Another commenter specifically questioned the need for the CMAT instrument in collecting new variables. The commenter also recommended that CMS consolidate all research efforts regarding payment for inpatient psychiatric services.

Response: In general, the majority of the prospective payment systems focus on data that predict the cost and/or payment for the provision of services. While this is the current focus, it is our position that costs and payments may be influenced by a number of variables that are beyond those currently used for payment. We anticipate that in the future, quality and outcome measures may be useful in determining payments. In addition, in most of the prospective payment systems that rely on patient assessment data, additional variables are collected that may not be directly or significantly related, at that time, to the payment system, but could nonetheless be useful at some future time.

We believe that relying only on those variables that are currently perceived as directly or significantly influencing payment, may preclude potential refinements to the IPF PPS, limit research in the area, and prohibit the future inclusion of variables that could significantly predict payment, outcome, and quality. Therefore, we are reluctant to restrict further research and scientific excellence by building only on existing and available facility and payment variables.

Comment: For patient characteristics, a commenter recommended adding two statistical parameters to the RTI International^® study, length of the IPF stay and length of time since their last psychiatric hospitalization.

Response: We agree that it would be useful to investigate the potential relationship between the frequency of an individual's hospitalizations, their length of stay, and the per diem cost of their care. In addition, we believe that the issue is relevant as a topic for our monitoring and evaluation activities.

IX. Comments Beyond the Scope of the Final Rule

In response to the proposed rule, many commenters chose to raise issues that are beyond the scope of our proposals. In this final rule, we are not summarizing or responding to those comments in this document. However, we will review the comments and consider whether to take other actions, such as revising or clarifying CMS program operating instructions or procedures, based on the information or recommendations in the comments.

X. Provisions of the Final Rule

We are making a number of revisions to the regulations in order to implement the IPF PPS. Specifically, we are making conforming changes in 42 CFR parts 412 and 413. We are establishing a new subpart N in part 412, “Prospective Payment System for Hospital Inpatient Services of Inpatient Psychiatric Facilities.” We have reorganized the regulations text to make it easier to follow.

This subpart implements section 124 of the BBRA, which requires the implementation of a per diem prospective payment system for IPFs. Subpart N sets forth the framework for the IPF PPS, including the methodology used for the development of the Federal per diem base payment amount and related rules. These revisions and others are discussed in detail below.

Section 412.1 Scope of Part

We are revising the authority citation to include “Section 124 of Public Law 106-113” and “Section 405 of Public Law 108-173.”

We are revising § 412.1 by redesignating paragraphs (a)(2) and (a)(3) as paragraphs (a)(3) and (a)(4).

We are adding a new paragraph (a)(2) that specifies that this part implements section 124 of Public Law 106-113 by establishing a per diem based prospective payment system for inpatient operating and capital costs of hospital inpatient services furnished to Medicare beneficiaries by an inpatient psychiatric facility that meets the conditions of subpart N.

We are revising § 412.1 by redesignating paragraphs (b)(12) and (b)(13) as paragraphs (b)(13) and (b)(14).

We are revising newly redesignated paragraph (b)(13) by removing reference “paragraph (a)(3)” and adding the reference “paragraph (a)(4)” in its place.

We are revising newly redesignated paragraph (b)(14) by removing reference “paragraph (a)(2)” and adding the reference “paragraph (a)(3)” in its place.

We are adding a new paragraph (b)(12) that summarizes the content of the new subpart N and sets forth the general methodology for paying operating and capital costs for inpatient psychiatric facilities effective with cost reporting periods beginning on or after January 1, 2005.

Section 412.20 Hospital Services Subject to the Prospective Payment Systems

We are amending § 412.20(a) by adding a reference to IPFs.

We are revising § 412.20 by redesignating paragraphs (b), (c), and (d), as paragraphs (c), (d), and (e).

We are adding a new paragraph (b) that indicates that effective for cost reporting periods beginning on or after January 1, 2005, covered inpatient hospital inpatient services furnished by an IPF as specified in § 412.404 of subpart N are paid under the IPF PPS.

Section 412.22 Excluded Hospitals and Hospital Units: General Rules

We are amending § 412.22(b) by revising paragraph (b) to state that except for those hospitals specified in paragraph (c) of this section, and § 412.20(b), (c), and (d), all excluded hospitals (and excluded hospital units, as described in § 412.23 through § 412.29) are reimbursed under the cost reimbursement rules set forth in part 413 of this chapter, and are subject to the ceiling on the rate of hospital cost increases as specified in § 413.40.

Section 412.23 Excluded Hospitals: Classifications

We are revising § 412.23 by redesignating paragraphs (a)(1) and (a)(2) as paragraphs (a)(2) and (a)(3).

We are adding a new paragraph (a)(1) that specifies the requirements a psychiatric hospital must meet in order to be excluded from reimbursement under the hospital IPPS as specified in § 412.1(a)(1) and to be paid under the IPF PPS as specified in § 412.1(a)(2).

We are revising paragraph (b) by removing the reference “§ 412.1(a)(2)” and adding the reference to “412.1(a)(3).”

We are revising paragraph (b)(9) by removing the reference to “§ 412.2(a)(2)” and adding the reference to “412.1(a)(3)” in its place.

We are revising paragraph (e) by removing the reference to “§ 412.1(a)(3)” and adding “§ 412.1(a)(4)” in its place.

Section 412.25 Excluded Hospital Units: Common Requirements

We are amending § 412.25(a) by adding a reference to § 412.1(a)(2).

Section 412.27 Excluded Psychiatric Units: Additional Requirements

We are amending the introductory text of § 412.27 by adding reference to § 412.1(a)(1) and (a)(2).

We are amending § 412.27(a) by removing the words the “Third Edition,” and adding in its place, “Fourth Edition, Text Revision.”

Section 412.429 Excluded Rehabilitation Units: Additional Requirements

We are revising the introductory text by removing the reference “§ 412.1(a)(2)” and adding “§ 412.1(a)(3)” in its place.

Section 412.116 Method of Payment

We are revising § 412.116 by redesignating paragraphs (a)(3) and (a)(4) as paragraphs (a)(4) and (a)(5).

We are adding a new paragraph (a)(3) that specifies the cost-reporting period to which the IPF PPS applies and how payments for inpatient psychiatric services are made to a qualified IPF.

Section 412.130 Exclusion of New Rehabilitation Units and Expansion of Units Already Excluded

Subpart N—Prospective Payment System for Hospital Inpatient Services of Inpatient Psychiatric Facilities

We are revising paragraph (a)(1) and paragraph (a)(2) by removing reference to “§ 412.1(a)(2)” and adding reference “§ 412.1(a)(3)” in its place.

We are adding a new subpart N as follows:

Section 412.400 Basis and Scope of Subpart

We are adding a new § 412.400. In § 412.400(a), we provide the requirements for the implementation of a PPS for IPFs.

In § 412.400(b), we specify that this subpart sets forth the framework for the IPF PPS, including the methodology used for the development of payment rates and associated adjustments, the application of a transition period, and related rules for IPFs for cost reporting periods beginning on or after January 1, 2005.

Section 412.402 Definitions

In § 412.402, we are defining the following terms for purposes of this new subpart:

• Comorbidity
• Federal per diem base rate
• Federal per diem payment amount
• Federal per diem
• Fixed dollar loss threshold
• Inpatient psychiatric facilities
• Interrupted stay
• Outlier payment
• Principal diagnosis
• Rural area
• Urban area

Section 412.404 Conditions for Payment Under the Prospective Payment System for Hospital Inpatient Services of Psychiatric Facilities

In § 412.404(a), we specify that IPFs must meet the following general requirements to receive payment under the IPF PPS:

• The IPF must meet the conditions as specified in this subpart.
• If the IPF fails to comply fully with the provisions of this part, then CMS may, as appropriate—

++ Withhold (in full or in part) or reduce payment to the IPF until the facility provides adequate assurances of compliance; or

++ Classify the IPF as a hospital subject to the IPPS.

In paragraph (b), we specify that, subject to the special payment provisions of § 412.22(c), an IPF must meet the general criteria set forth in § 412.22 for exclusion from the hospital IPPS as specified in § 412.1(a)(1). Additionally, a psychiatric hospital must meet the criteria set forth in § 412.23(a), § 482.60, § 482.61, and § 482.62 and psychiatric units must meet the criteria set forth in § 412.25 and § 412.27.

In paragraph (c), we specify the prohibited and permitted charges that may be imposed on Medicare beneficiaries.

In paragraph (c)(1), we specify that except as permitted in paragraph (c)(2), an IPF may not charge the beneficiary for any services for which payment is made by Medicare, except as permitted in paragraph (c)(2), even if the IPFs costs are greater than the amount the facility is paid under the IPF PPS.

In paragraph (c)(2), we specify that an IPF receiving payment for a covered stay may charge the Medicare beneficiary or other person for only the applicable deductible and coinsurance amounts under § 409.82, § 409.83, and § 409.87.

In paragraph (d), we specify the following provisions for furnishing IPF services directly or under arrangement:

Applicable payments made under the IPF PPS are considered payment in full for all inpatient hospital services (as defined in § 409.10(a)). In addition, we specify the following—

• Inpatient hospital services do not include physician, physician assistant, nurse practitioner, clinical nurse specialist, certified nurse midwives, qualified psychologist, and certified registered nurse anesthetist services.
• Payment is not made to a provider or supplier other than the IPF, except for services provided by a physician, physician assistant, nurse practitioner, clinical nurse specialist, certified nurse midwives, qualified psychologist, and certified registered nurse anesthetist.
• The IPF must furnish all necessary covered services to the Medicare beneficiary directly or under arrangement (as defined in § 409.3).

In paragraph (e), we specify that IPFs must meet the recordkeeping and cost reporting requirements of § 412.27(c), § 413.20, and § 413.24.

Section 412.422 Basis of Payment

In § 412.422(a), we specify that under the IPF PPS, IPFs will receive a predetermined per diem amount, adjusted for patient characteristics and facility characteristics, for inpatient hospital services furnished to Medicare Part A fee-for-service beneficiaries. In addition, we specify that during the transition period, payment is based on a blend of the Federal per diem payment amount and the facility-specific payment rate as specified in § 412.426.

In § 412.422(b), we specify that payments made under the IPF PPS represent payment in full for inpatient operating and capital-related costs associated with furnishing Medicare covered service in an IPF, but not for the cost of an approved medical education program described in § 413.85 and § 413.86 and for bad debts of Medicare beneficiaries as specified in § 413.80.

Section 412.424 Methodology for Calculating the Federal Per Diem Payment Amount

In § 412.424, we specify the methodology for calculating the Federal per diem base rate for IPFs.

In paragraph (a), we specify the data sources used to calculate the Federal per diem base rate.

In paragraph (b), we specify that we determine the average inpatient operating, ancillary, and capital related per diem cost for which payment is made to IPF as described in paragraph (a)(1).

In paragraph (c), we specify that the methodology used for determining the Federal per diem base rate for cost reporting periods beginning on or after January 5, 2005 through June 30, 2006 includes the following:

• The updated average per diem amount
• The budget-neutrality adjustment factor
• Outlier payments
• Standardization
• Computation of the Federal per diem base rate

In paragraph (d), we specify that the Federal per diem payment amount for IPFs is the product of the Federal per diem base rate, the facility-level adjustments applicable to the IPF and the patient-level adjustments applicable to the case as described below:

• Facility-level adjustments include:

++ Adjustment for wages

++ Rural location

++ Teaching adjustments

++ Cost of living adjustments for IPFs in Alaska and Hawaii

++ IPFs with qualifying emergency departments

• Patient-level adjustments include:

++ Age

++ Diagnosis-related group assignment

++ Principal diagnosis

++ Comorbodities

++ Variable per diem adjustments

• Other payment adjustments include:

++ Outlier payments

++ Stop-loss payments

++ Special payment provision for interrupted stay

++ Patients who receive ECT treatments

++ Adjustment for high-cost outlier cases

In paragraph (d), we specify the special payment provisions for interrupted stays.

Section 412.426 Transition Period

In § 412.426(a), we specify the duration of the transition period to the IPF PPS. In addition, we specify that IPFs receive a payment that is a blend of the Federal per diem payment amount and the facility-specific payment amount the IPF would receive under the TEFRA payment methodology.

In paragraph (b), we specify how the facility-specific payment amount is calculated.

In paragraph (c), we specify that a new IPF, that is, a facility that under present or previous ownership, or both, has its first cost reporting period as an IPF beginning on or after January 1, 2005, is paid based on 100 percent of the full Federal per diem payment.

Section 412.428 Publication of Updated to the IPF PPS

In § 412.428, we specify how we plan to publish information each year in the Federal Register to update the IPF PPS.

Section 412.432 Method of Payment Under the IPF PPS

In § 412.432, we specify the following method of payment used under the IPF PPS:

• General rules for receiving payment
• Periodic interim payments including—

++ Criteria for receiving periodic interim payments

++ Frequency of payments

++ Termination of periodic interim payments

• Interim payment for Medicare bad debts and for costs of an approved education program and other costs paid outside the PPS
• Outlier payments
• Accelerated payments including—

++ General rule for requesting accelerated payments

++ Approval of accelerated payments

++ Amount of the accelerated payment

++ Recovery of the accelerated payment

Section 413.1 Introduction

We are revising the authority citation to include “Section 124 of Public Law 106-113.”

We are amending § 413.1(d)(2)(ii) by removing the words “psychiatric hospitals (as well as separate psychiatric units (distinct parts) of short-term general hospitals).”

We are revising § 413.1 by redesignating paragraphs (d)(2)(iv), (d)(2)(v), (d)(2)(vi), and (d)(2)(vii) as paragraphs (d)(2)(vi), (d)(2)(vii), (d)(2)(viii), and (d)(2)(ix).

We are adding a new paragraph (iv) to specify that for cost reporting periods beginning before January 1, 2005, payment to psychiatric hospitals (as well as separate psychiatric units of short-term general hospitals) that are excluded under subpart B of part 412 of this chapter from the PPS is on a reasonable cost basis, subject to the provisions of § 413.40.

We are adding a new paragraph (v) to specify that for cost reporting periods beginning on or after January 1, 2005, payment to psychiatric hospitals that meet the conditions of § 412.404 of this chapter is made under the PPS as described in subpart N of part 412.

Section 413.40 Ceiling on the Rate of Increase in Hospital Costs

Section 413.40(a)(2)(i) specifies the types of facilities to which the ceiling on the rate of increase in hospital inpatient costs is not applicable.

We are revising § 413.40(a)(2)(i) by redesignating paragraphs (a)(2)(i)(C) and (a)(2)(i)(D) as paragraphs (a)(2)(i)(D) and (a)(2)(i)(E).

We are adding a new paragraph (a)(2)(i)(C) to § 413.40 to clarify that § 413.40 is not applicable to psychiatric hospitals and psychiatric units under subpart N of part 412 of this chapter for cost reporting periods beginning on or after January 1, 2005.

We are republishing paragraph (a)(2)(ii).

We are revising paragraph (a)(2)(ii)(B) to include reference to psychiatric hospitals and psychiatric units as specified in § 412.22, § 412.23, § 412.25, § 412.27, § 412.29, and § 412.30 of this chapter.

We are revising paragraph (a)(2)(iii) by redesignating paragraphs (a)(2)(iii) and (a)(2)(iv) as paragraphs (a)(2)(iv) and (a)(2)(v).

We are revising paragraph (a)(2)(ii)(C) by removing reference to “paragraph (a)(2)(iv)” and adding the reference to “paragraph (a)(2)(v)” in its place.

We are adding a new paragraph (a)(2)(iii) to specify psychiatric facilities are excluded from the prospective payment system as specified in § 412.1(a)(1) and paid under § 412.1(a)(2) for cost reporting periods beginning on or after January 1, 2005.

Section 413.64 Payment to Providers: Special Rules

We are amending § 413.64(h)(2)(i) to add a reference to hospitals paid under the IPF PPS.

Section 413.70 Payment for Services of a CAH

We are revising paragraph (e) to specify that for cost reporting periods beginning before January 1, 2005, payment is made on a reasonable cost basis, subject to the provisions of § 413.40. For cost reporting periods beginning on or after January 1, 2005, payment is based on prospectively determined rates under subpart N § 412.400 through § 412.432) of part 412 of this subchapter.

XI. Collection of Information Requirements

These regulations do not impose any new information collection requirements. The burden of the requirements in § 412.404(e), reporting and recordkeeping requirements, are captured in the burden for the cross-referenced § 412.27(c), § 413.20, and § 413.24 under OMB approval numbers 0938-0301, 0938-0050, 0938-0358, and 0938-0600.

XII. Regulatory Impact Analysis

A. Overall Impact

We have examined the impact of this final rule as required by Executive Order 12866 (September 1993, Regulatory Planning and Review), the Regulatory Flexibility Act (RFA) (September 19, 1980, Pub. L. 96-354), section 1102(b) of the Act, the Unfunded Mandates Reform Act of 1995 (UMRA) (Pub. L. 104-4), and Executive Order 13132).

Executive Order 12866 (as amended by Executive Order 13258, which merely reassigns responsibility of duties) directs 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). A regulatory impact analysis (RIA) must be prepared for major rules with economically significant effects ($100 million or more in any 1 year).

Based on analysis of the aggregate dollar impacts for each of the different facility types, we have determined that the re-distributive impact of the IPF PPS among facility types is $96 million in the first year the system is fully implemented. In addition, our analysis showed that an estimated payment “reduction” of almost $48 million would occur for psychiatric units and an estimated payment “increase” of $18 million would occur for for-profit hospitals, $27 million for government-operated hospitals, and slightly more than $3 million for non-profit hospitals. Although this final rule does not meet the $100 million threshold established by Executive Order 12866 in its first year of implementation, we have determined that this final rule is a major rule within the meaning of Executive Order 12866 in its first year of implementation, because the re-distributive effects are estimated to be close to constituting a shift of $100 million in the first year of implementation. In addition, although we have not estimated the distributional impact of this rule in subsequent years, because of the trends in medical expenditure discussed below, we believe it is likely that the rule would have distributional impacts greater than $100 million in subsequent years, relative to TEFRA payments. In addition, because the IPF PPS must be budget neutral in accordance with section 124(a)(1) of Public Law 106-113, we estimate that there will be no budgetary impact for the Medicare program as discussed later in this analysis.

The RFA requires agencies to analyze options for regulatory relief of small businesses. For purposes of the RFA, small entities include small businesses, nonprofit organizations, and small government jurisdictions. Most hospitals and most other providers and suppliers are small entities, either by nonprofit status or by having revenues of $29 million or less in any 1 year. Medicare fiscal intermediaries are not considered to be small entities. Individuals and States are not included in the definition of a small entity.

HHS considers that a substantial number of entities are affected if the rule impacts more than 5 percent of the total number of small entities as it does in this rule. We included all freestanding psychiatric hospitals (79 are non-profit hospitals) in the analysis since their total revenues do not exceed the $29 million threshold. We also included psychiatric units of small hospitals, that is, fewer than 100 beds. We did not include psychiatric units within larger hospitals in the analysis because we believe this final rule would not significantly impact total revenues of the entire hospital that supports the unit. We have provided the following RFA analysis in section B, to emphasize that although the final rule would impact a substantial number of IPFs that were identified as small entities, we do not believe it would have a significant economic impact. Based on the analysis of the 1063 psychiatric facilities that were classified as small entities by the definitions described above, we estimate the combined impact of the IPF PPS will be a 5-percent increase in payments relative to their payments under TEFRA. We have prepared the following analysis to describe the impact of the final rule in order to provide a factual basis for our conclusions regarding small business impact.

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. We have determined that this final rule would have a substantial impact on hospitals classified as located in rural areas. As discussed earlier in this preamble, we are providing a payment adjustment of 17 percent for IPFs located in rural areas. In addition, we are establishing a 3-year transition to the new system to allow IPFs an opportunity to adjust to the new system. Therefore, the impacts shown in Table 10 below reflect the adjustments that are designed to minimize or eliminate any potentially significant negative impact that the IPF PPS may otherwise have on small rural IPFs.

Section 202 of the UMRA also requires that agencies assess anticipated costs and benefits before issuing any final rule that may result in expenditures in any 1 year by State, local, or tribal governments, in the aggregate, or by the private sector, of $110 million or more. This final rule does not mandate any requirements for State, local, or tribal governments nor would it result in expenditures by the private sector of $110 million or more in any 1 year.

Executive Order 13132 establishes certain requirements that an agency must meet when it promulgates a final rule that imposes substantial direct requirement costs on State and local governments, preempts State law, or otherwise has Federalism implications.

We have examined this final rule under the criteria set forth in Executive Order 13132 and have determined that the final rule will not have any negative impact on the rights, roles, and responsibilities of State, local, or tribal governments or preempt State law.

B. Anticipated Effects

Below, we discuss the impact of this final rule on the Federal Medicare budget and on IPFs.

1. Budgetary Impact

Section 124(a)(1) of Public Law 106-113 requires us to set the payment rates contained in this final rule to ensure that total payments under the IPF PPS are projected to equal the amount that would have been paid if the IPF PPS had not been implemented. As a result of this analysis, which is discussed in section V.B.2.b. of this final rule, we are establishing a budget-neutrality adjustment to the Federal per diem base rate. Thus, there will be no budgetary impact to the Medicare program by implementation of the IPF PPS.

2. Impacts on Providers

To understand the impact of the IPF PPS on providers, it is necessary to compare estimated payments that would be made under the current TEFRA payment methodology (current payments) to estimated payments under the IPF PPS. The IPFs were grouped into the categories listed below based on characteristics provided in the Online Survey and Certification and Reporting (OSCAR) file and the 2002 cost report data from HCRIS:

• Facility Type
• Location
• Teaching Status Adjustment
• Census Region
• Size

To estimate the impacts among the various categories of IPFs, we had to compare estimated future payments that would have been made under the TEFRA payment methodology to estimated payments under the IPF PPS. We estimated the impacts using the same set of providers (1,806 IPFs) that was used for the regression analysis to calculate the budget-neutral Federal per diem base rate, and to determine the appropriateness of various adjustments to the Federal per diem base rate. A detailed explanation of the methods we used to simulate TEFRA payments and estimate payments under the IPF PPS is provided in section V. of this final rule.

The impacts reflect the estimated “losses” or “gains” among the various classifications of IPF providers for the first year of the IPF PPS. Prospective payments were based on the budget-neutral Federal per diem base rate of $572 adjusted by the IPFs' estimated patient-level, facility-level adjustments, and simulated outlier amounts. This simulated PPS payment was compared to the IPF's payments based on its cost from the cost report inflated to the midpoint of the implementation period (January 1, 2005 through June 30, 2006) and subject to the updated per discharge target amount. Table 10 below illustrates the aggregate impact of the IPF PPS on various classifications of IPFs. The first column identifies the type of IPF, the second column indicates the number of IPFs for each type of IPF, and the third column indicates the ratio of IPF PPS payments to the current TEFRA payments in the first period of the transition.

3. Results

We measured the impact of the IPF PPS by comparing estimated payments under the IPF PPS relative to current TEFRA payments. This was computed as a ratio of IPF PPS payment to current TEFRA payment for each classification of IPF. We have prepared the following summary of the impact of the IPF PPS set forth in this final rule.

a. Facility type

We grouped the IPFs into the following four categories: (1) Psychiatric units; (2) government-operated hospitals; (3) for-profit hospitals; and (4) non-profit hospitals. Roughly 77 percent of all IPFs are psychiatric units. The impact analysis in Table 10 indicates that under the IPF PPS, freestanding psychiatric hospitals receive an estimated “increase” relative to the current payment. Psychiatric units have an estimated IPF PPS payment to current TEFRA payment ratio of 0.98, the government-operated hospitals have an estimated IPF PPS payment to current TEFRA payment ratio of 1.13, and the non-profit and for-profit hospitals have an estimated IPF PPS payment to current TEFRA payment ratio of 1.02 and 1.05, respectively.

b. Location

Approximately 24 percent of all IPFs are located in rural areas. The impact analysis in Table 10 indicates that under the IPF PPS, the estimated IPF PPS payment to current TEFRA payment ratio is approximately 1.00 for rural and urban IPFs. When we group all of the IPFs by facility type within urban and rural locations, the impact analysis indicates that the estimated IPF PPS payment to current TEFRA payment ratios would be between approximately 0.98 and 1.05 for all IPFs except government-operated hospitals. Under the IPF PPS, the payment ratios for rural and urban government-operated hospitals are estimated to be 1.14 and 1.12, respectively.

c. Teaching Status Adjustment

Using the ratio of interns and residents to the average daily census for each facility as a measure of the magnitude of the teaching status, we grouped facilities into the following four major categories: (1) Non teaching; (2) less than 0.10 (it is not a percent) ratio of interns and residents to average daily census; (3) 0.10 to 0.30 ratio of interns and residents to average daily census; and (4) more than 0.30 ratio of interns and residents to average daily census. Facilities with a teaching ratio greater than 0.10, have payment ratios less than 1.00.

d. Census Region

Under the IPF PPS, IPFs in the Mid-Atlantic region receive a payment ratio of approximately 1.03 when compared to IPFs in other regions that receive payment ratios between approximately 0.98 and 1.01. Specifically, the New England States, the West North Central States, and the Mountain States receive payment ratios of 1.00. The South Atlantic States, East North Central States, and the Pacific States, receive payments ratios of approximately 0.99. The East South Central States have a payment ratio of 1.01, and the West South Central States have a ratio of 0.98.

e. Size

We grouped the IPFs into 5 categories for each group of psychiatric facilities based on bed size: (1) Under 12 beds; (2) 12 to 25 beds; (3) 25 to 50 beds; (4) 50 to 75 beds; and (5) over 75 beds. Under the IPF PPS, the majority of IPFs' bed sizes were categories in which the payment ratio would be greater than 0.98. Under the IPF PPS, large IPFs with over 75 beds receive the highest payment ratio (1.10 for psychiatric hospitals and 1.01 for psychiatric units), while psychiatric units with less than 10 beds receive the lowest payment ratio of 0.96.

4. Effect on the Medicare Program

Based on actuarial projections resulting from our experience with other prospective payment systems, we estimate that Medicare spending (total Medicare program payments) for IPF services over the next 5 years would be as follows:

These estimates are based on the current estimate of increases in the number of proposed excluded hospitals with capital market basket as follows:

• 3.4 percent for FY 2005;
• 3.0 percent for FY 2006;
• 2.8 percent for FY 2007;
• 2.7 percent for FY 2008;
• 3.0 percent for FY 2009; and
• 3.0 percent for FY 2010.

We estimate that there would be a change in fee-for-service Medicare beneficiary enrollment as follows:

• 0.5 percent in FY 2005;
• -7.3 percent in FY 2006;
• -4.7 percent in FY 2007;
• -0.2 percent in FY 2008;
• -0.1 percent in FY 2009; and
• 1.4 percent in FY 2010.

Consistent with the statutory requirement for budget neutrality in the initial implementation period, we intend for estimated aggregate payments under the IPF PPS to equal the estimated aggregate payments that would be made if the IPF PPS were not implemented. Our methodology for estimating payments for purposes of the budget-neutrality calculations uses the best available data.

After the IPF PPS is implemented, we will evaluate the accuracy of the assumptions used to compute the budget-neutrality calculation. We intend to analyze claims and cost report data from the first year of the IPF PPS to determine whether the factors used to develop the Federal per diem base rate are not significantly different from the actual results experienced in that year. We are planning to compare payments under the final IPF PPS (which relies on an estimate of cost-based TEFRA payments using historical data from a base year and assumptions that trend the data to the initial implementation period) to estimated cost-based TEFRA payments based on actual data from the first year of the IPF PPS. The percent difference (either positive or negative) would be applied prospectively to the established prospective payment rates to ensure the rates accurately reflect the payment levels intended by the statute. We intend to perform this analysis within the first 5 years of the implementation of the IPF PPS.

Section 124 of Public Law 106-113 provides the Secretary broad authority in developing the IPF PPS, including the authority for appropriate adjustments. In accordance with this authority, as stated above, we may make a one-time prospective adjustment to the Federal per diem base rate in an effort to ensure that the best historical data available forms the foundation of the prospective payment rates in future years.

5. Effect on Beneficiaries

Under the IPF PPS, IPFs will receive payment based on the average resources consumed by patients for each day. We do not expect changes in the quality of care or access to services for Medicare beneficiaries under the IPF PPS. In fact, we believe that access to IPF services would be enhanced due to the patient and facility level adjustment factors, all of which are intended to adequately reimburse IPFs for expensive cases. Finally, the stop-loss policy is intended to assist IPFs during the transition. In addition, we expect that paying prospectively for IPF services will enhance the efficiency of the Medicare program.

6. Computer Hardware and Software

We do not anticipate that IPFs will incur additional systems operating costs in order to effectively participate in the IPF PPS. We believe that IPFs possess the computer hardware capability to handle the billing requirements under the IPF PPS. Our belief is based on indications that approximately 99 percent of hospital inpatient claims are submitted electronically. In addition, we are not adopting significant changes in claims processing (see section IV. C. of this final rule).

C. Alternatives Considered

We considered the following alternatives in developing the IPF PPS: One option we considered incorporated not only the patient-level and facility-level variables described previously, but also a site-of-service distinction. Under this approach, psychiatric units would have received a higher per diem payment, all other factors being equal, based on the assumption that psychiatric units on average treat a more complex and costly case-mix. A psychiatric unit adjustment to the otherwise applicable per diem payment rate would reflect the absence of a more sophisticated patient classification system specifically linked to resource use. Our analysis of the FY 2002 cost report and billing data used to develop the final IPF PPS reveals that an adjustment would have increased the otherwise applicable per diem payment to psychiatric units by approximately 33 percent. The average 2002 IPF per diem costs was $615 for psychiatric units, $534 for non-profit hospitals, $448 for proprietary providers, and $378 for governmental-operated facilities. While some of the higher than average per diem cost in psychiatric units may be due to a greater medical and surgical acuity among patients treated in psychiatric units, part of the difference is likely attributable to economy of scale inefficiencies associated with operating small units, including higher overhead expenses, and generally lower occupancy rates. A psychiatric unit site-of-service distinction in payment rates would represent a proxy adjuster in lieu of a more sophisticated patient classification system.

We considered alternative policies in order to reduce financial risk to facilities in the event that they experience substantial reductions in Medicare payments during the period of transition to the IPF PPS. As discussed previously in this final rule, we have adopted a provision that would guarantee each facility an average payment per case under the IPF PPS that is estimated to be no less than a minimum proportion of its average payment per case under TEFRA. We analyzed the impact on losses if we were to make a payment adjustment to ensure that the minimum IPF PPS per case payment to an IPF is at least 70 percent of its TEFRA payment.

The stop-loss adjustment will be applied to the IPF PPS portion of Medicare payments during the transition. For example, during year 1 of the 3-year transition period, three-quarters of the payment is based on TEFRA, and one-quarter of the payment is based on the Federal rate. We would apply the stop-loss adjustment to the portion of the IPF's payments during the transition based on the Federal rate. We estimate that the combined effects of the transition and the stop-loss policies will ensure that per case payments relative to pre-IPF PPS TEFRA per case payments are no less than 92.5 percent in year 1, 85 percent in year 2, and 77.5 percent in year 3. We estimate that about 10 percent of IPFs will receive additional payments under the stop-loss policy.

The 70 percent of TEFRA stop-loss policy would require a reduction in the per diem rate to make the stop-loss policy budget neutral. As a result, we made a reduction to the Federal per diem base rate of 0.4 percent in order to maintain budget neutrality.

In accordance with the provisions of Executive Order 12866, this rule was reviewed by OMB.

List of Subjects

42 CFR Part 412

• Administrative practice and procedure
• Health facilities
• Medicare
• Puerto Rico
• Reporting and recordkeeping requirements

42 CFR Part 413

• Health facilities
• Kidney diseases
• Medicare
• Puerto Rico
• Reporting and recordkeeping requirements

For the reasons set forth in the preamble, the Centers for Medicare & Medicaid Services amends 42 CFR chapter IV as follows:

PART 412—PROSPECTIVE PAYMENT SYSTEMS FOR INPATIENT PSYCHIATRIC SERVICES

1. The authority citation for part 412 is revised to read as follows:

Authority: Secs. 1102 and 1871 of the Social Security Act (42 U.S.C. 1302 and 1395hh), Sec. 124 of Pub. L. 106-113, 113 Stat. 1515, and Sec. 405 of Pub. L. of 108-173, 117 Stat. 2266, 42 U.S.C. 1305, 1395.

Subpart A—General Provisions

2. Section 412.1 is amended as follows:

a. Redesignating paragraphs (a)(2) and (a)(3) as paragraphs (a)(3) and (a)(4).

b. Adding a new paragraph (a)(2).

c. Redesignating paragraphs (b)(12) and (b)(13) as paragraphs (b)(13) and (b)(14).

d. Adding a new paragraph (b)(12).

e. Amending newly redesignated paragraph (b)(13) by removing the reference “paragraph (a)(3)” and adding the reference “paragraph (a)(4)” in its place.

f. Amending newly redesignted paragraph (b)(14) by removing the reference “paragraph (a)(2)” and adding the reference “paragraph (a)(3)” in its place.

The additions read as follows:

§ 412.1

Scope of part.

(a) * * *

(2) This part implements section 124 of Public Law 106-113 by establishing a per diem prospective payment system for the inpatient operating and capital costs of hospital inpatient services furnished to Medicare beneficiaries by a psychiatric facility that meets the conditions of subpart N of this part.

* * * * *

(b) * * *

(12) Subpart N describes the prospective payment system specified in paragraph (a)(2) of this section for inpatient psychiatric facilities and sets forth the general methodology for paying the operating and capital-related costs of inpatient hospital services furnished by inpatient psychiatric facilities effective with cost reporting periods beginning on or after January 1, 2005.

* * * * *

Subpart B—Hospital Services Subject to and Excluded From the Prospective Payment Systems for Inpatient Operating Costs and Inpatient Capital Related Costs

3. Section 412.20 is amended as follows:

a. Revising paragraph (a).

b. Redesignating paragraphs (b), (c), and (d) as paragraphs (c), (d), and (e).

c. Adding a new paragraph (b).

The revision and addition read as follows:

§ 412.20

Hospital services subject to the prospective payment systems.

(a) Except for services described in paragraphs (b), (c), (d), and (e) of this section, all covered hospital inpatient services furnished to beneficiaries during the subject cost reporting periods are paid under the prospective payment system as specified in § 412.1(a)(1).

(b) Effective for cost reporting periods beginning on or after January 1, 2005, covered inpatient hospital services furnished to Medicare beneficiaries by a inpatient psychiatric facility that meets the conditions of § 412.404 are paid under the prospective payment system described in subpart N of this part.

* * * * *

4. Section 412.22 is amended by revising paragraph (b).

§ 412.22

Excluded hospitals and hospital units: General rules.

* * * * *

(b) Cost reimbursement. Except for those hospitals specified in paragraph (c) of this section, and § 412.20(b), (c), and (d), all excluded hospitals (and excluded hospital units, as described in § 412.23 through § 412.29) are reimbursed under the cost reimbursement rules set forth in part 413 of this chapter, and are subject to the ceiling on the rate of hospital cost increases as specified in § 413.40 of this chapter.

* * * * *

5. Section 412.23 is amended as follows:

a. Republishing paragraph (a) introductory text.

b. Redesignating paragraphs (a)(1) and (a)(2) as paragraphs (a)(2) and (a)(3).

c. Adding a new paragraph (a)(1).

d. Amending the introductory text to paragraph (b) by removing the reference “§ 412.1(a)(2)” and adding the reference to “§ 412.1(a)(3)” in its place.

e. Amending paragraph (b)(9) by removing the reference to “§ 412.2(a)(2)” and adding the reference to “§ 412.1(a)(3)” in its place.

f. Revising the introductory text to paragraph (e).

The republication and addition read a follows:

§ 412.23

Excluded hospitals: Classifications.

* * * * *

(a) Psychiatric hospitals. A psychiatric hospital must—

(1) Meet the following requirements to be excluded from the prospective payment system as specified in § 412.1(a)(1) and to be paid under the prospective payment system as specified in § 412.1(a)(2) and in subpart N of this part;

* * * * *

(e) Long-term care hospitals. A long-term care hospital must meet the requirements of paragraph (e)(1) and (e)(2) of this section and, when applicable, the additional requirement of § 412.22(e), to be excluded from the prospective payment system specified in § 412.1(a)(1) and to be paid under the prospective payment system specified in § 412.1(a)(4) and in Subpart O of this part.

* * * * *

6. Section 412.25 is amended by revising the paragraph (a) introductory text to read as follows:

§ 412.25

Excluded hospital units: Common requirements.

(a) Basis for exclusion. In order to be excluded from the prospective payment systems as specified in § 412.1(a)(1) and to be paid under the prospective payment system as specified in 412.1(a)(2), a psychiatric unit must meet the following requirements.

* * * * *

7. Section 412.27 is amended as follows:

a. Revising the introductory text.

b. Amending paragraph (a) by removing the words “Third Edition”, and adding in its place, “Fourth Edition, Text Revision”.

The revision reads as follows:

§ 412.27

Excluded psychiatric units: Additional requirements.

In order to be excluded from the prospective payment system as specified in § 412.1(a)(1), and paid under the prospective payment system as specified in § 412.1(a)(2), a psychiatric unit must meet the following requirements:

* * * * *

§ 412.29

[Amended]

8. In § 412.29, the introductory text is amended by removing the reference “§ 412.1(a)(2)” and adding the reference “§ 412.1(a)(3)” in its place.

9. Section 412.116 is amended as follows:

a. Redesignating paragraphs (a)(3) and (a)(4) as paragraphs (a)(4) and (a)(5).

b. Adding a new paragraph (a)(3).

The addition reads as follows:

§ 412.116

Method of payment.

(a) * * *

(3) For cost reporting periods beginning on or after January 1, 2005, payments for inpatient hospital services furnished by an inpatient psychiatric facility that meets the conditions of § 412.404 are made as described in § 412.432.

* * * * *

§ 412.130

[Amended]

10. In § 412.130, paragraphs (a)(1) and (a)(2) are amended by removing the reference “§ 412.1(a)(2)” and adding the reference “§ 412.1(a)(3)” in its place.

11. A new subpart N is added to read as follows:

Subpart N—Prospective Payment System for Hospital Inpatient Services of Inpatient Psychiatric Facilities

412.400

Basis and scope of subpart.

412.402

Definitions.

412.404

Conditions for payment under the prospective payment system for inpatient hospital services of psychiatric facilities.

412.422

Basis of payment.

412.424

Methodology for calculating the Federal per diem payment amount.

412.426

Transition period.

412.428

Publication of Updates to the inpatient psychiatric facility prospective payment system.

412.432

Method of payment under the inpatient psychiatric facility prospective payment system.

Subpart N—Prospective Payment System for Inpatient Hospital Services of Inpatient Psychiatric Facilities

§ 412.400

Basis and scope of subpart.

(a) Basis. This subpart implements section 124 of Public Law 106-113, which provides for the implementation of a per diem-based prospective payment system for inpatient hospital services of inpatient psychiatric facilities.

(b) Scope. This subpart sets forth the framework for the prospective payment system for the inpatient hospital services of inpatient psychiatric facilities, including the methodology used for the development of the Federal per diem rate, payment adjustments, implementation issues, and related rules. Under this system, for cost reporting periods beginning on or after January 1, 2005, payment for the operating and capital-related costs of inpatient hospital services furnished by inpatient psychiatric facilities to Medicare Part A fee-for-service beneficiaries is made on the basis of prospectively determined payment amount applied on a per diem basis.

§ 412.402

Definitions.

As used in this subpart—

Comorbidity means all specific patient conditions that are secondary to the patient's primary diagnosis and that coexist at the time of admission, develop subsequently, or that affect the treatment received or the length of stay or both. Diagnoses that relate to an earlier episode of care that have no bearing on the current hospital stay are excluded.

Federal per diem base rate means the payment based on the average routine operating, ancillary, and capital-related cost of 1 day of hospital inpatient services in an inpatient psychiatric facility.

Federal per diem payment amount means the Federal per diem base rate with all applicable adjustments.

Fixed dollar loss threshold means a dollar amount by which the costs of a case exceed payment in order to qualify for an outlier payment.

Inpatient psychiatric facilities means hospitals that meet the requirements as specified in § 412.22, § 412.23(a), § 482.60, § 482.61, and § 482.62, and units that meet the requirements as specified in § 412.22, § 412.25, and § 412.27.

Interrupted stay means a Medicare inpatient is discharged from an inpatient psychiatric facility and is admitted to any inpatient psychiatric facility within 3 consecutive calendar days following discharge. The 3 consecutive calendar days begins with the day of discharge from the inpatient psychiatric facility and ends on midnight of the third day.

Outlier payment means an additional payment beyond the Federal per diem payment amount for cases with unusually high costs.

Principal diagnosis means the condition established after study to be chiefly responsible for occasioning the admission of the patient to the inpatient psychiatric facility also referred to as primary diagnosis. Principal diagnosis is also referred to as primary diagnosis.

Qualifying emergency department means an emergency department that is staffed and equipped to furnish a comprehensive array of emergency services and meting the definitions of a dedicated emergency department as specified in § 489.24(b).

Rural area means any area outside an urban area.

Urban area means an area as defined in § 412.62(f)(1)(ii).

§ 412.404

Conditions for payment under the prospective payment system for inpatient hospital services of psychiatric facilities.

(a) General requirements. (1) Effective for cost reporting periods beginning on or after January 1, 2005, an inpatient psychiatric facility must meet the conditions of this section to receive payment under the prospective payment system described in this subpart for inpatient hospital services furnished in to Medicare Part A fee-for-service beneficiaries.

(2) If an inpatient psychiatric facility fails to comply fully with these conditions, CMS may, as appropriate—

(i) Withhold (in full or in part) or reduce Medicare payment to the inpatient psychiatric facility until the facility provides adequate assurances of compliance; or

(ii) Classify the inpatient psychiatric facility as an inpatient hospital that is subject to the conditions of subpart C of this part and is paid under the prospective payment system as specified in § 412.1(a)(1).

(b) Inpatient psychiatric facilities subject to the prospective payment system. Subject to the special payment provisions of § 412.22(c), an inpatient psychiatric facility must meet the general criteria set forth in § 412.22. In order to be excluded from the hospital inpatient prospective payment system as specified in § 412.1(a)(1), a psychiatric hospital must meet the criteria set forth in § 412.23(a), § 482.60, § 482.61, and § 482.62 and psychiatric units must meet the criteria set forth in § 412.25 and § 412.27.

(c) Limitations on charges to beneficiaries—(1) Prohibited charges. Except as permitted in paragraph (c)(2) of this section, an inpatient psychiatric facility may not charge a beneficiary for any services for which payment is made by Medicare, even if the facility's cost of furnishing services to that beneficiary are greater than the amount the facility is paid under the prospective payment system.

(2) Permitted charges. An inpatient psychiatric facility receiving payment under this subpart for a covered hospital stay (that is, a stay that included at least one covered day) may charge the Medicare beneficiary or other person only the applicable deductible and coinsurance amounts under § 409.82, § 409.83, and § 409.87 of this chapter and for items or services as specified under § 489.20(a) of this chapter.

(d) Furnishing of inpatient hospital services directly or under arrangement. (1) Subject to the provisions of § 412.422, the applicable payments made under this subpart are payment in full for all inpatient hospital services, as specified in § 409.10 of this chapter. Hospital inpatient services do not include the following:

(i) Physicians' services that meet the requirements of § 415.102(a) of this chapter for payment on a fee schedule basis.

(ii) Physician assistant services, as specified in section 1861(s)(2)(K)(i) of the Act.

(iii) Nurse practitioners and clinical nurse specialist services, as specified in section 1861(s)(2)(K)(ii) of the Act.

(iv) Certified nurse midwife services, as specified in section 1861(gg) of the Act.

(v) Qualified psychologist services, as specified in section 1861(ii) of the Act.

(vi) Services of a certified registered nurse anesthetist, as specified in section 1861(bb) of the Act and defined in § 410.69 of this subchapter.

(2) CMS does not pay providers or suppliers other than inpatient psychiatric facilities for services furnished to a Medicare beneficiary who is an inpatient of the inpatient psychiatric facility, except for services described in paragraphs (d)(1)(i) through (d)(1)(vi) of this section

(3) The inpatient psychiatric facility must furnish all necessary covered services to a Medicare beneficiary who is an inpatient of the inpatient psychiatric facility, either directly or under arrangements (as specified in § 409.3 of this chapter).

(e) Reporting and recordkeeping requirements. All inpatient psychiatric facilities participating in the prospective payment system under this subpart must meet the recordkeeping and cost reporting requirements as specified in § 412.27(c), § 413.20, § 413.24, and § 482.61 of this chapter.

§ 412.422

Basis of payment.

(a) Method of Payment. (1) Under the inpatient psychiatric facility prospective payment system, inpatient psychiatric facilities receive a predetermined Federal per diem base rate for inpatient hospital services furnished to Medicare Part A fee-for-service beneficiaries.

(2) The Federal per diem payment amount is based on the Federal per diem base rate plus applicable adjustments as specified in § 412.424.

(3) During the transition period, payment is based on a blend of the Federal per diem payment amount as specified in § 412.424, and the facility-specific payment rate as specified in § 412.426.

(b) Payment in full. (1) The payment made under this subpart represents payment in full (subject to applicable deductibles and coinsurance as specified in subpart G of part 409 of this chapter) for inpatient operating and capital-related costs associated with furnishing Medicare covered services in an inpatient psychiatric facility, but not the cost of an approved medical education program as specified in § 413.79 through § 413.75 of this chapter.

(2) In addition to the Federal per diem payment amounts, inpatient psychiatric facilities receive payment for bad debts of Medicare beneficiaries, as specified in § 413.80 of this chapter.

§ 412.424

Methodology for calculating the Federal per diem payment amount.

(a) Data sources. (1) To calculate the Federal per diem base rate (as specified in paragraph (b) of this section for inpatient psychiatric facilities, as specified in paragraph (b) of this section, CMS uses the following data sources:

(2) The best Medicare data available to estimate the average inpatient operating and capital-related costs per day made as specified in part 413 of this chapter.

(i) Patient and facility cost report data capturing routine and ancillary costs.

(ii) An appropriate wage index to adjust for wage differences.

(iii) An increase factor to adjust for the most recent estimate of increases in the prices of an appropriate market basket of goods and services provided by inpatient psychiatric facilities.

(b) Determining the average per diem cost of inpatient psychiatric facilities for FY 2002. CMS determines the average inpatient operating, ancillary, and capital-related per diem cost for which payment is made to each inpatient psychiatric facility, using the available data described in paragraph (a) of this section.

(c) Determining the Federal per diem base rate for cost reporting periods beginning on or after January 1, 2005 through June 30, 2006. (1) General. Payment under the inpatient psychiatric facility prospective payment system is based on a standardized per diem payment referred to as the Federal per diem base rate. The Federal per diem base rate is the unadjusted cost for 1 day of inpatient hospital services in an inpatient psychiatric facility in a base year as described in paragraph (b) of this section. The unadjusted cost per day is adjusted in accordance with paragraphs (c)(2) through (c)(5) of this section.

(2) Update of the average per diem cost. CMS applies the increase factor described in paragraph (a)(2)(iii) of this section to the updated average per diem cost to the midpoint of the January 1, 2005 through June 30, 2006, under the update methodology described in section 1886(b)(3)(B)(ii) of the Act.

(3) Budget neutrality. (i) CMS adjusts the updated average per diem cost so that the aggregate payments in the first 18 months (for January 1, 2005 through June 30, 2006) under the inpatient psychiatric facility prospective payment system are estimated to equal the amount that would have been made to the inpatient psychiatric facilities under part 413 of this chapter if the inpatient psychiatric facility prospective payment system described in this subpart were not implemented.

(ii) CMS evaluates the accuracy of the budget-neutrality adjustment within the first 5 years after implementation of the inpatient psychiatric facility prospective payment system. CMS may make a one-time prospective adjustment to the Federal per diem base rate to account for significant differences between the historical data on cost-based TEFRA payments (the basis of the budget-neutrality adjustment at the time of implementation) and estimates of TEFRA payments based on actual data from the first year of the prospective payment system.

(4) Outlier payments. CMS determines a reduction factor equal to the estimated proportion of outlier payments described in paragraph (d)(3)(i) of this section.

(5) Standardization. CMS determines a reduction factor to reflect estimated increases in the Federal per diem base rate as defined in § 412.402 resulting from the facility-level and patient-level adjustments described in paragraph (d) of this section.

(6) Computation of the Federal per diem base rate. The Federal per diem base rate is computed as follows:

(i) For cost reporting periods beginning on or after January 1, 2005 and on or before June 30, 2006, the Federal per diem base rate is computed in accordance with paragraph (c) of this section.

(ii) For inpatient psychiatric facilities beginning on or after July 1, 2006, the Federal per diem base rate will be the Federal per diem base rate for the previous year, updated by an increase factor described in paragraph (a)(2)(iii) of this section.

(d) Determining the Federal per diem payment amount. The Federal per diem payment amount is the product of the Federal per diem base rate established under paragraph (c) of this section, the facility-level adjustments applicable to the inpatient psychiatric facility, and the patient-level adjustments applicable to the case.

(1) Facility-level adjustments. (i) Adjustment for wages. CMS adjusts the labor portion of the Federal per diem base rate to account for geographic differences in the area wage levels using an appropriate wage index. The application of the wage index is made on the basis of the location of the inpatient psychiatric facility in an urban or rural area as defined in § 412.402.

(ii) Rural location. CMS adjusts the Federal per diem base rate for inpatient psychiatric facilities located in a rural area as defined in § 412.402.

(iii) Teaching adjustment. CMS adjusts the Federal per diem base rate by a factor to account for indirect medical education costs.

(A) An inpatient psychiatric facility's teaching adjustment is based on the ratio of the number of residents training in the inpatient psychiatric facility divided by the facility's average daily census.

(B) The number of full-time equivalent residents used in calculating the teaching adjustment cannot exceed the number of full-time equivalent residents in a base year.

(1) The base year is the inpatient psychiatric facility's most recently filed cost report filed with its fiscal intermediary before November 15, 2004. Residents with less than full-time status and residents rotating through the inpatient psychiatric facility for less than a full year will be counted in proportion to the time they spend in the inpatient psychiatric facility.

(2) The teaching status adjustment for new inpatient psychiatric facilities as defined in § 412.426 is made in accordance with § 413.79(e)(1)(i) and (ii).

(C) If an inpatient psychiatric facility has fewer full-time equivalent residents than in its base year payment of the teaching adjustment will be based on the actual number of full-time equivalent residents. The inpatient psychiatric facility may add residents in subsequent years up to its resident cap established under section (1)(iii)(B) of this paragraph.

(iv) Inpatient psychiatric facilities located in Alaska and Hawaii. CMS adjusts the non-labor portion of the Federal per diem base rate to reflect the higher cost of living of inpatient psychiatric facilities located in Alaska and Hawaii.

(v) Adjustment for IPF with qualifying emergency departments. (A) CMS adjusts the Federal per diem base rate to account for the costs associated with maintaining a qualifying emergency department. A qualifying emergency department is staffed and equipped to furnish a comprehensive array of emergency services and meets the requirements of § 489.24(b) and § 413.65.

(B) Where the inpatient psychiatric facility is part of an acute care hospital that has a qualifying emergency department as described in paragraph (d)(1)(v)(A) of this section and an individual patient is discharged to the inpatient psychiatric facility from that acute care hospital, CMS would not apply the emergency adjustment.

(2) Patient-level adjustments. (i) Age. CMS adjusts the Federal per diem base rate to account for patient age based on age groupings specified by CMS.

(ii) Diagnosis-related group assignment. The inpatient psychiatric facility must identify a principal diagnosis as specified in § 412.27(a) for each patient. CMS adjusts the Federal per diem base rate by a factor to account for the CMS inpatient psychiatric facility prospective payment system recognized diagnosis-related group assignment associated with each patient's principal diagnosis.

(iii) Principal diagnosis. The inpatient psychiatric facility must identify a principal psychiatric diagnosis as specified in § 412.27(a) for each patient. CMS adjusts the Federal per diem base rate by a factor to account for the diagnosis-related group assignment associated with the principal diagnosis, as specified by CMS.

(iv) Comorbidities. CMS adjusts the Federal per diem base rate by a factor to account for certain comorbidities as specified by CMS.

(v) Variable per diem adjustments. CMS adjusts the Federal per diem base rate by factors as specified by CMS to account for the cost of each day of inpatient psychiatric care relative to the cost of the median length of stay.

(3) Other adjustments. (i) Outlier payments. CMS provides an additional payment if an inpatient psychiatric facility's estimated total cost for a case exceeds a fixed dollar loss threshold as defined in § 412.402 plus the Federal per diem payment amount for the case.

(A) The fixed dollar loss threshold is adjusted for the inpatient psychiatric facility's adjustments for wage area, teaching, rural location, and cost of living adjustment for facilities located in Alaska and Hawaii.

(B) The outlier payment equals 80 percent of the difference between the IPF's estimated cost for the case and the adjusted threshold amount for days 1 through 9, and 60 percent for day 10 and thereafter.

(C) For discharges occurring in cost reporting periods beginning on or after January 1, 2005, outlier payments are subject to the adjustments specified at § 412.84(i) and § 412.84(m) of this part, except that national urban and rural median cost-to-charge ratios would be used instead of statewide average cost-to-charge ratios.

(ii) Stop-loss payments. CMS will provide additional payments during the transition period, specified in § 412.426(a)(1) through (3), to an inpatient psychiatric facility to ensure that aggregate payments under the prospective payment system are at least 70 percent of the amount the inpatient psychiatric facility would have received under reasonable cost reimbursement had the prospective payment system not been implemented.

(iii) Special payment provision for interrupted stays. If a patient is discharged from an inpatient psychiatric facility and is admitted to the same or another inpatient psychiatric facility within 3 consecutive calendar days following the discharge, the case is considered to be continuous for the purposes listed below. The 3 consecutive calendar days begins with the day of discharge from the inpatient psychiatric facility and ends on midnight of day 3.

(A) Determining the appropriate variable per diem adjustment, as specified in paragraph (d)(2)(v) of this section, applicable to the case.

(B) Determining whether the total cost for a case meets the criteria for outlier payments, as specified in paragraph (d)(3)(i)(C) of this section.

(iv) Payment for electroconvulsive therapy treatments. CMS provides an additional payment to reflect the cost of electroconvulsive therapy treatments received by a patient during an inpatient psychiatric facility stay in a manner specified by CMS.

(v) Adjustment for high-cost cases. CMS provides for an additional payment if the estimated total cost for a case exceeds a fixed dollar loss threshold plus the total per diem payment amount for the case.

(A) The fixed dollar loss threshold is adjusted for area wage levels, teaching status, and rural location.

(B) The additional payment equals 80 percent of the difference between the estimated cost of the case and the Federal per diem payment amount for days 1 through 9, and 60 percent for days 10 and beyond.

(C) Effective for discharges occurring in cost reporting periods beginning on or after January 1, 2005, additional payments made under this section would be subject to the adjustments at § 412.84(i) and § 412.84(m) of this part, except that the national urban and rural median cost-to-charge ratios would be used instead of statewide averages, and at § 412.84(m) of this part.

§ 412.426

Transition period.

(a) Duration of transition period and composition of the blended transition payment. Except as provided in paragraph (c) of this section, for cost reporting periods beginning on or after January 1, 2005 through June 30, 2008, an inpatient psychiatric facility receives a payment comprised of a blend of the estimated Federal per diem payment amount, as specified in § 412.424(c) and a facility-specific payment as specified under paragraph (b).

(1) For cost reporting periods beginning on or after January 1, 2005 and on or before June 30, 2006, payment is based on 75 percent of the facility-specific payment and 25 percent is based on the Federal per diem payment amount.

(2) For cost reporting periods beginning on or after July 1, 2006 and on or before June 30, 2007, payment is based on 50 percent of the facility-specific payment and 50 percent is based on the Federal per diem payment amount.

(3) For cost reporting periods beginning on or after July 1, 2007 and on or before June 30, 2008, payment is based on 25 percent of the facility-specific payment and 75 percent is based on the Federal per diem payment amount.

(4) For cost reporting periods beginning on or after July 1, 2008, payment is based entirely on the Federal per diem payment amount.

(b) Calculation of the facility-specific payment. The facility-specific payment is equal to the estimated payment for each cost reporting period in the transition period that would have been made without regard to this subpart. The facility's Medicare fiscal intermediary calculates the facility-specific payment for inpatient operating costs and capital costs in accordance with part 413 of this chapter.

(c) Treatment of new inpatient psychiatric facilities. New inpatient psychiatric facilities, are facilities that under present or previous ownership or both have their first cost reporting period as an IPF beginning on or after January 1, 2005. New IPFs are paid based on 100 percent of the Federal per diem payment amount.

§ 412.428

Publication of Updates to the inpatient psychiatric facility prospective payment system.

CMS will publish annually in the Federal Register information pertaining to updates to the inpatient psychiatric facility prospective payment system. This information includes:

(a) A description of the methodology and data used to calculate the updated Federal per diem base payment amount.

(b) The rate of increase factor as described in 412.424(a)(2)(iii), which is based on the excluded hospital with capital market basket under the update methodology of 1886(b)(3)(B)(ii) of the Act for each year.

(c) The best available hospital wage index and information regarding whether an adjustment to the Federal per diem base rate is needed to maintain budget neutrality.

(d) Updates to the fixed dollar loss threshold in order to maintain the appropriate outlier percentage.

(e) Describe the ICD-9-CM coding changes and DRG classification changes discussed in the annual update to the hospital inpatient prospective payment system regulations.

(f) Update the electroconvulsive therapy adjustment by a factor specified by CMS.

§ 412.432

Method of payment under the inpatient psychiatric facility prospective payment system.

(a) General rule. Subject to the exceptions in paragraphs (b) and (c) of this section, an inpatient psychiatric facility receives payment under this subpart for inpatient operating cost and capital-related costs for each inpatient stay following submission of a bill.

(b) Periodic interim payments (PIP). (1) Criteria for receiving PIP.

(i) An inpatient psychiatric facility receiving payment under this subpart may receive PIP for Part A services under the PIP method subject to the provisions of § 413.64(h) of this chapter.

(ii) To be approved for PIP, the inpatient psychiatric facility must meet the qualifying requirements in § 413.64(h)(3) of this chapter.

(iii) A hospital that is receiving periodic interim payments also receives payment under this subpart for applicable services furnished by its excluded psychiatric unit.

(iv) As provided in § 413.64(h)(5) of this chapter, intermediary approval is conditioned upon the intermediary's best judgment as to whether payment can be made under the PIP method without undue risk of resulting in an overpayment to the provider.

(2) Frequency of payment. For facilities approved for PIP, the intermediary estimates the annual inpatient psychiatric facility's Federal per diem prospective payments, net of estimated beneficiary deductibles and coinsurance, and makes biweekly payments equal to 1/26 of the total estimated amount of payment for the year. If the inpatient psychiatric facility has payment experience under the prospective payment system, the intermediary estimates PIP based on that payment experience, adjusted for projected changes supported by substantiated information for the current year. Each payment is made 2 weeks after the end of a biweekly period of service as specified in § 413.64(h)(6) of this chapter. The interim payments are reviewed at least twice during the reporting period and adjusted if necessary. Fewer reviews may be necessary if an inpatient psychiatric facility receives interim payments for less than a full reporting period. These payments are subject to final settlement.

(3) Termination of PIP. (i) Request by the inpatient psychiatric facility. Subject to the provisions of paragraph (b)(1)(iii) of this section, an inpatient psychiatric facility receiving PIP may convert to receiving prospective payments on a non-PIP basis at any time.

(ii) Removal by the intermediary. An intermediary terminates PIP if the inpatient psychiatric facility no longer meets the requirements of § 413.64(h) of this chapter.

(c) Interim payments for Medicare bad debts and for costs of an approved education program and other costs paid outside the prospective payment system. For Medicare bad debts and for costs of an approved education program and other costs paid outside the prospective payment system, the intermediary determines the interim payments by estimating the reimbursable amount for the year based on the previous year's experience, adjusted for projected changes supported by substantiated information for the current year, and makes biweekly payments equal to 1/26 of the total estimated amount. Each payment is made 2 weeks after the end of the biweekly period of service as specified in § 413.64(h)(6) of this chapter. The interim payments are reviewed at least twice during the reporting period and adjusted if necessary. Fewer reviews may be necessary if an inpatient psychiatric facility receives interim payments for less than a full reporting period. These payments are subject to final cost settlement.

(d) Outlier payments. Additional payments for outliers are not made on an interim basis. Outlier payments are made based on the submission of a discharge bill and represents final payment subject to the cost report settlement specified in § 412.84(i) and § 412.84(m).

(e) Accelerated payments. (1) General rule. Upon request, an accelerated payment may be made to an inpatient psychiatric facility that is receiving payment under this subpart and is not receiving PIP under paragraph (b) of this section if the inpatient psychiatric facility is experiencing financial difficulties because of the following:

(i) There is a delay by the intermediary in making payment to the inpatient psychiatric facility.

(ii) Due to an exceptional situation, there is a temporary delay in the inpatient psychiatric facility's preparation and submittal of bills to the intermediary beyond the normal billing cycle.

(2) Approval of accelerated payment. An inpatient psychiatric facility's request for an accelerated payment must be approved by the intermediary and CMS.

(3) Amount of accelerated payment. The amount of the accelerated payment is computed as a percent of the net payment for unbilled or unpaid covered services.

(4) Recovery of accelerated payment. Recovery of the accelerated payment is made by recoupment as inpatient psychiatric facility bills are processed or by direct payment by the inpatient psychiatric facility.

PART 413—PRINCIPLES OF REASONABLE COST REIMBURSEMENT; PAYMENT FOR END-STAGE RENAL DISEASE SERVICES; PROSPECTIVELY DETERMINED PAYMENT FOR SKILLED NURSING FACILITIES

1. The authority citation for part 413 is revised to read as follows:

Authority: Secs. 1102, 1812(d), 1814(b), 1815, 1833(a), (i), and (n), 1861 (v), 1871, 1881, 1883, and 1886 of the Social Security Act (42 U.S.C. 1302, 1395d(d), 1395f(b), 1395g, 1395l(a), (i), and (n), 1395x(v), 1395hh, 1395rr, 1395tt, and 1395ww) Sec 124 of Pub. L. 106-113, 113 Stat. 1515.

2. Section 413.1 is amended as follows:

a. Revising paragraph (d)(2)(ii).

b. Redesignating paragraphs (d)(2)(iv), (d)(2)(v), (d)(2)(vi), and (d)(2)(vii) as paragraphs (d)(2)(vi), (d)(2)(vii), (d)(2)(viii), and (d)(2)(ix).

(c) Adding new paragraphs (d)(2)(iv) and (d)(2)(v).

The revision and additions read as follows:

§ 413.1

Introduction.

* * * * *

(d) * * *

(2) * * *

(ii) Payment to children's hospitals that are excluded from the prospective payment systems under subpart B of part 412 of this chapter, and hospitals outside the 50 States and the District of Columbia is on a reasonable cost basis, subject to the provisions of § 413.40.

* * * * *

(iv) For cost reporting periods beginning before January 1, 2005, payment to psychiatric hospitals (as well as separate psychiatric units (distinct parts) of short-term general hospitals) that are excluded under subpart B of part 412 of this chapter from the prospective payment system is on a reasonable cost basis, subject to the provisions of § 413.40.

(v) For cost reporting periods beginning on or after January 1, 2005, payment to inpatient psychiatric facilities that meet the conditions of § 412.404 of this chapter, is made under the prospective payment system described in subpart N of part 412 of this chapter.

* * * * *

3. Section 413.40 is amended as follows:

a. Redesignating paragraphs (a)(2)(i)(C) and (a)(2)(i)(D) as paragraphs (a)(2)(i)(D) and (a)(2)(i)(E).

b. Adding a new paragraph (a)(2)(i)(C).

c. Republishing paragraph (a)(2)(ii) introductory text.

d. Revising paragraph (a)(2)(ii)(B).

e. Amending paragraph (a)(2)(ii)(C) by removing reference to “paragraph (a)(2)(iv)” and adding the reference “paragraph (a)(2)(v)” in its place.

f. Redesignating paragraphs (a)(2)(iii) and (a)(2)(iv) as paragraphs (a)(2)(iv) and (a)(2)(v).

g. Adding a new paragraph (a)(2)(iii).

The revision and additions read as follows:

§ 413.40

Ceiling on the rate of increase in hospital inpatient costs.

(a) * * *

(2) * * *

(i) * * *

(C) Psychiatric hospitals and psychiatric units that are paid under the prospective payment system for inpatient psychiatric facilities described in subpart N of part 412 of this chapter for cost reporting periods beginning on or after January 1, 2005.

* * * * *

(ii) For cost reporting periods beginning on or after October 1, 1983, this section applies to—

* * * * *

(B) Psychiatric and rehabilitation units excluded from the prospective payment systems, as specified in § 412.1(a)(1) of this chapter and in accordance with § 412.25 through § 412.30 of this chapter, except as limited by paragraphs (a)(2)(iii) and (a)(2)(iv) of this section with respect to psychiatric and rehabilitation hospitals and psychiatric and rehabilitation units as specified in § 412.22, § 412.23, § 412.25, § 412.27, § 412.29 and § 412.30 of this chapter.

* * * * *

(iii) For cost reporting periods beginning on or after October 1, 1983 and before January 1, 2005 this section applies to psychiatric hospitals and psychiatric units that are excluded from the prospective payment systems as specified in § 412.1(a)(1) of this chapter and paid under the prospective payment system as specified in § 412.1(a)(2) of this chapter.

* * * * *

4. Section 413.64 is amended by revising paragraph (h)(2)(i) to read as follows:

§ 413.64

Payment to providers: Specific rules.

* * * * *

(h) * * *

(2) * * *

(i) Part A inpatient services furnished in hospitals that are excluded from the prospective payment systems, as specified in § 412.1(a)(1) of this chapter under subpart B of part 412 of this subchapter, or are paid under the prospective payment systems described in subpart N, O, and P of part 412 of this chapter.

* * * * *

5. Section 413.70 is amended by revising paragraph (e) to read as follows:

§ 413.70

Payment for services of a CAH.

* * * * *

(e) Payment for service of distinct part psychiatric and rehabilitation units of CAHS. Payment for inpatient services of distinct part psychiatric units of CAHs—

(1) For cost reporting periods beginning before January 1, 2005, payment is made on a reasonable cost basis, subject to the provisions of § 413.40.

(2) For cost reporting periods beginning on or after January 1, 2005, payment is made in accordance with regulations governing inpatient psychiatric facilities at subpart N (§ 412.400 through § 412.432) of Part 412 of this subchapter.

(3) Payment for inpatient services of distinct part rehabilitation units of CAHs is made in accordance with regulations governing the inpatient rehabilitation facilities prospective payment system at Subpart P (§ 412.600 through § 412.632) of Part 412 of this subchapter.

(Catalog of Federal Domestic Assistance Program No. 93.778, Medical Assistance Program)

(Catalog of Federal Domestic Assistance Program No. 93.773, Medicare—Hospital Insurance; and Program No. 93.774, Medicare—Supplementary Medical Insurance Program)

Dated: October 26, 2004.

Mark B. McClellan,

Administrator, Centers for Medicare & Medicaid Services.

Approved: November 2, 2004.

Tommy G. Thompson,

Secretary.

Note:

The following Addenda will not appear in the Code of Federal Regulations

Addendum A—Psychiatric Prospective Payment Adjustment Rate and Adjustment Factors