# Medicare Program; Prospective Payment System for Inpatient Psychiatric Facilities > Briefs, arguments, decisions, and more. URL: https://www.frixlaw.com/law-library/documents/fr%3A03-29137 ## Record - **Collection:** Federal Register - **Document type:** Proposed Rule - **Published:** November 28, 2003 - **Citation:** 68 FR 66920 ## Text DEPARTMENT OF HEALTH AND HUMAN SERVICES Centers for Medicare & Medicaid Services 42 CFR Parts 412, 413, and 424 [CMS-1213-P] RIN 0938-AL50 Medicare Program; Prospective Payment System for Inpatient Psychiatric Facilities AGENCY: Centers for Medicare & Medicaid Services (CMS), HHS. ACTION: Proposed rule. SUMMARY: This rule proposes a prospective payment system for Medicare payment of inpatient hospital services furnished in psychiatric hospitals and psychiatric units of acute care hospitals. This rule proposes to implement section 124 of the Medicare, Medicaid, and SCHIP Balanced Budget Refinement Act of 1999 (BBRA), which requires the implementation of a per diem prospective payment system for hospital services of psychiatric hospitals and psychiatric units. The prospective payment system described in this proposed rule would replace the reasonable cost-based payment system currently in effect. DATES: We will consider comments if we receive them at the appropriate address, as provided below, no later than 5 p.m. on January 27, 2004. ADDRESSES: In commenting, please refer to file code CMS-1213-P. Because of staff and resource limitations, we cannot accept comments by facsimile (FAX) transmission. Mail written comments (one original and two copies) to the following address ONLY: Centers for Medicare & Medicaid Services, Department of Health and Human Services, Attention: CMS-1213-P, P.O. Box 8012, Baltimore, MD 21244-8012. Please allow sufficient time for mailed comments to be received timely in the event of delivery delays. If you prefer, you may deliver (by hand or courier) your written comments (one original and two copies) to one of the following addresses: Room 445-G, Hubert H. Humphrey Building, 200 Independence Avenue, SW., Washington, DC 20201, or Room C5-14-03, 7500 Security Boulevard, Baltimore, MD 21244-1850. (Because access to the interior of the HHH Building is not readily available to persons without Federal Government identification, commenters are encouraged to leave their comments in the CMS drop slots located in the main lobby of the building. A stamp-in clock is available for persons wishing to retain a proof of filing by stamping in and retaining an extra copy of the comments being filed.) Comments mailed to the addresses indicated as appropriate for hand or courier delivery may be delayed and could be considered late. For information on viewing public comments, see the beginning of the SUPPLEMENTARY INFORMATION section. FOR FURTHER INFORMATION CONTACT: Janet Samen, (410) 786-4533. Philip Cotterill, (410) 786-6598, for information regarding the regression analysis. SUPPLEMENTARY INFORMATION: Inspection of Public Comments: Comments received timely will be available for public inspection as they are received, generally beginning approximately 4 weeks after publication of a document, at the headquarters of the Centers for Medicare & Medicaid Services, 7500 Security Boulevard, Baltimore, Maryland 21244, Monday through Friday of each week from 8:30 a.m. to 4 p.m. To schedule an appointment to view public comments, phone (410) 786-9994. 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 (202) 512-1800 (or toll-free at 1-888-293-6498) or by faxing to (202) 512-2250. 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 Web site address is: http://www.access.gpo.gov/nara/index.html. To assist readers in referencing sections contained in this document, we are providing the following table of contents. Table of Contents I. Background A. General and Legislative History B. Overview of the Payment System for Psychiatric Hospitals and Psychiatric Units before the BBA 1. Description of the TEFRA Payment Methodology 2. BBA Amendments to TEFRA 3. BBRA Amendments to TEFRA 4. BIPA Amendments to TEFRA II. Overview of the Proposed IPF Prospective Payment System A. Use of Diagnostic Codes for Payment 1. ICD 2. DRGs B. Limitations of the DRG System for Psychiatric Patients C. Proposed DRG Adjustments Under the Proposed IPF Prospective Payment System D. DRGs not Recognized in the Proposed IPF Prospective Payment System E. Applicability of the Proposed IPF Prospective Payment System III. Development of the Proposed IPF Per Diem Payment Amount A. Proposed Market Basket B. Development of the Proposed Case-Mix Adjustment Regression. 1. Proposed Patient-Level Characteristics a. DRGs b. Comorbidities c. Patient Age and Gender d. Length of Stay 2. Proposed Facility-Level Characteristics a. Rural Location b. Teaching Status c. Disproportionate Share Hospital Status d. Psychiatric Units in General Acute Care Hospitals e. Adjustment for Alaska and Hawaii IPFs 3. Proposed Payment Adjustments a. Proposed Outlier Adjustment b. Methodology for Proposed Outlier Payments c. Proposed Implementation of the Outlier Policy 1. Statistical Accuracy of Cost-to-Charge Ratio 2. Adjustment of IPF Outlier Payments d. Computation of Proposed Outlier Payments e. Interrupted Stays C. Development of the Proposed Budget-Neutral Federal Per Diem Base Rate 1. Data Used to Develop the Proposed Federal Per Diem Base Rate 2. Calculation of the Proposed Per Diem Amount 3. Determining the Proposed Update Factors for the Budget-Neutrality Calculation a. Cost Report Data for April 1, 2004 through June 30, 2005 b. Estimate of Total Payments under the TEFRA Payment System c. Payments Under the Proposed Prospective Payment System without a Budget-Neutrality Adjustment d. Calculation of the Proposed Budget-Neutrality Adjustment 4. The Proposed Behavioral Offset 5. Proposed Federal Per Diem Base Rate 6. Proposed Changes to Physician Recertification Requirements E. Proposed Area Wage Adjustment F. Effect of the Proposed Transition on Budget Neutrality G. Calculation of the Proposed Payment IV. Implementation of the Proposed IPF Prospective Payment System A. Proposed Transition B. New Providers C. Claims Processing D. Periodic Interim Payments (PIP) E. Limitation on Beneficiaries Charges V. Future Updates A. Proposed Annual Update Strategy B. Update of the ICD Codes and DRGs C. Future Refinements 1. RTI International” (trade name of Research Triangle Institute) a. Mode of Practice b. Patient Characteristics c. Analysis 2. University of Michigan Research 3. Case-Mix Tool VI. Provisions of the Proposed Rule VII. Collection of Information Requirements VIII. Response to Comments IX. Regulatory Impact Statement A. Overall Impact B. Anticipated Effects 1. Budgetary Impact 2. Impact on Providers 3. Results a. Facility Type b. Location c. Teaching Status 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 Psychiatric Prospective Payment Adjustment Addendum B1: Proposed Pre-Reclassified Wage Index for Urban Areas Addendum B2: Proposed Wage Index for Rural Areas Addendum C: Proposed Case-Mix Assessment Tool Acronyms Because of the many terms to which we refer by acronym in this proposed 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 IRFs Inpatient rehabilitation facilities LTCHs Long-term care hospitals MedPAR Medicare provider analysis and review file PIP Periodic interim payments 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 hospital inpatient 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 hospital inpatient 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 for acute care hospitals in 1983, with the enactment of Pub. L. 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 hospital inpatient services with a prospective payment system. Although most hospital inpatient services became subject to the prospective payment system, certain specialty hospitals were excluded from the prospective payment system 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 (LTCHs), children's hospitals, and rehabilitation hospitals and units. 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 (BIPA) (Pub. L. 106-554) to replace the cost-based methods of reimbursement with a prospective payment system for the following excluded hospitals: • Rehabilitation hospitals (including units in acute care hospitals). • Psychiatric hospitals (including units in acute care hospitals. • LTCHs. 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 I.B.1. of this proposed rule. Section 124 of the BBRA mandated that the Secretary—(1) develop a per diem prospective payment system for inpatient hospital services furnished in psychiatric hospitals and psychiatric units; (2) include in the prospective payment system 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 prospective payment system; and (5) submit a report to the Congress describing the development of the prospective payment system. Section 124 also required that the payment system for inpatient psychiatric services be implemented for cost reporting periods beginning on or after October 1, 2002. The creation of each new payment system requires an extraordinary amount of lead-time to develop and implement the necessary changes to our existing computerize claims processing systems. In order to meet the BBRA requirement to develop an adequate patient classification system, we undertook two research projects. It became apparent that the two research projects could not be completed in time for us to implement an inpatient psychiatric facility prospective payment system by October 1, 2002. It was impossible for us to analyze our existing administrative data in a sufficient amount of time to go through notice and comment rulemaking and implementation of the inpatient psychiatric facility prospective payment system by the statutory deadline. This delay enabled us to analyze our existing administrative data to determine the feasibility and validity of using these data to develop the proposed inpatient psychiatric facility prospective payment system. We are using a combination of available facility and patient specific data for this proposed rule. Our research efforts will continue and will be used to refine the proposed system. In this proposed rule, as required under section 124 of the BBRA, we set forth the proposed Medicare prospective payment system for psychiatric hospitals and psychiatric units of acute care hospitals. We note that many hospitals have “psychiatric units,” however; only those units that are separately certified from the hospital and meet the requirements of § 412.23, § 412.25, and § 412.27 are excluded from the hospital inpatient prospective payment system and would be subject to this proposed prospective payment system. Psychiatric units that are currently paid under the hospital inpatient prospective payment system and do not meet the requirements of § 412.22, § 412.25 and § 412.27 would not be paid under the proposed IPF prospective payment system. The proposed system includes an adequate patient classification system that would result in higher prospective payments to providers treating more costly, resource intensive patients using statistically objective criteria. We are proposing to establish a base payment rate that would be paid to inpatient psychiatric facilities for each day of inpatient psychiatric care (the Federal per diem base rate). The proposed base rate would be adjusted by certain proposed patient-level and facility-level characteristics. B. Overview of the Payment System for Psychiatric Hospitals and Psychiatric Units Before the BBA 1. Description of the TEFRA Payment Methodology Hospitals and units that are excluded from the hospital inpatient prospective payment system under section 1886(d)(1)(B) of the Act are paid for their inpatient operating costs under the provisions of Pub. L. 97-248 (TEFRA). The TEFRA provisions are found in section 1886(b) of the Act and implemented in regulations at 42 CFR Part 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 acute care hospital inpatient prospective payment system. A 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 hospital inpatient prospective payment system 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 a prospective payment system-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 hospital inpatient prospective payment system. A complete explanation of these amendments can be found in the final rule concerning the hospital inpatient prospective payment system 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 and rehabilitation hospitals and 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 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, 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, as 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, which amended section 1886(b)(3)(H) of the Act, were explained in detail and implemented in the hospital inpatient prospective payment system interim final rule published in the Federal Register on August 1, 2000 (65 FR 47026) and in the hospital inpatient prospective payment system 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. Overview of the Proposed IPF Prospective Payment System As required by statute, we are proposing a per diem prospective payment system for psychiatric hospitals and psychiatric units (hereinafter referred to as inpatient psychiatric facilities (IPFs)) that would replace the current reasonable cost-based payment system under the TEFRA provisions. In this rule, we are proposing to base the system on data from the 1999 Medicare Provider Analysis and Review (MedPAR) file, which includes patient characteristics (for example, patients' diagnoses and age), and data from the 1999 Hospital Cost Report Information System (HCRIS), which includes facility characteristics (for example, location and teaching status). We are using the 1999 MedPAR and HCRIS data because they are the best available data. Based on our analysis, we are proposing the following methodology as the basis of the proposed IPF prospective payment system: • Compute a Federal per diem base rate to be paid to all psychiatric hospitals and psychiatric units based on the sum of the average routine operating, ancillary, and capital costs for each patient day of psychiatric care in an IPF adjusted for budget neutrality (see section III.C. of this proposed rule). In computing the Federal per diem base rate, our analysis showed that routine operating and capital represent approximately 88 percent of total costs and the remaining 12 percent of total costs are for ancillary services. • Adjust the Federal per diem base rate to reflect certain patient and facility characteristics that were found in the regression analysis to be associated with statistically significant cost differences (see section III.B. of this proposed rule). The variance explained by patient characteristics (19 percent) in the regression analysis is limited by the nature of the administrative data used to develop this system, which assigns average facility routine costs to individual patients. We are conducting research to better understand the relationship between individual patient characteristics and average facility routine costs that could be incorporated into the payment system in future updates. We note that ancillary costs are already identifiable at the individual patient level. • Implement an April 1, 2004 effective date and a 3-year transition period. As explained in section IV of this proposed rule, it ultimately may be necessary to delay implementation beyond April 2004 as well as to increase the length of the transition period. However, the rate development, budget-neutrality adjustment, and impact analysis assume an April 1, 2004 effective date and a 3-year transition period. • Include research information for future refinement of the proposed patient classification system. Part of this research could result in a new patient assessment instrument that could identify additional patient level characteristics. In addition, we are proposing to make the following types of adjustments to appropriately make payments on a per-diem basis: • Patient-level adjustments for age, specified diagnosis-related groups, and selected high cost comorbidity categories. These patient-level characteristics explain approximately 19 percent of the variance in the cost of psychiatric care in the administrative data, which establishes the empirical basis for this methodology. • Facility adjustments that include a wage index adjustment, rural location adjustment, and an indirect teaching adjustment. These facility characteristics explain approximately 13 percent of the variance in the costs of psychiatric care in the administrative data. • Variable per diem adjustments to recognize the higher costs incurred in the early days of a psychiatric stay. • Outlier adjustments to target greater payment to the high cost cases. We are also proposing the following policies: • Interrupted stay policy for the purpose of applying the variable per diem adjustment and the outlier policy. • Coding policy (see section II. A.) that would—(1) require IPFs to report patient diagnoses using the International Classification of Diseases-9th Revision, Clinical Modification (ICD-9-CM) code set to report the psychiatric diagnosis; and (2) select the diagnosis-related groups (DRGs) that would be used for payment adjustments in this proposed rule. A. Use of Diagnostic Codes for Payment The patient's principal diagnosis of his or her physical or mental condition is essential because it typically acts as a guide for treatment and validates payment. It is for these reasons that diagnostic information is routinely reported on hospital claims and is used in other prospective payment systems. In mental health treatment, the principal tool recognized and utilized by the psychiatric community for diagnostic assessment is the Diagnostic and Statistical Manual of Mental Disorders (DSM). The DSM provides a broad and comprehensive description of patients through behavioral domains, or “axes.” This multiaxial system is routinely used by clinical staff to diagnose patients and plan treatment. The DSM is currently in its fourth revision text revision (DSM-IV-TR). Although, the DSM is used for patient assessment by IPFs, the ICD-9-CM coding system is used currently for reporting diagnostic information for payment purposes. 1. ICD The ICD coding system was designed for the classification of morbidity and mortality information for statistical purposes and for the indexing of hospital records by disease. Chapter Five of the ICD-9-CM includes the codes for mental disorders. In addition, the following definitions (as described in the 1984 Revision of the Uniform Hospital Discharge Data Set) are requirements of the ICD-9-CM coding system. • Diagnoses include all diagnoses that affect the current hospital stay. • Principal diagnosis is defined as the condition established, after study, to be chiefly responsible for occasioning the admission of the patient to the hospital for care. • Other diagnoses (also called secondary diagnoses or additional diagnoses) are defined as all conditions that coexist at the time of admission, that develop subsequently, or that affect the treatment received or the length of stay or both. Diagnoses that relate to an earlier episode of care and have no bearing on the current hospital stay are excluded. We are proposing to require IPFs to use the psychiatric diagnosis codes in Chapter Five (“Mental Disorder”) of the ICD-9-CM to report diagnostic information for the proposed IPF prospective payment system. All changes to the ICD coding system that would affect the proposed IPF prospective payment system would be addressed annually in the hospital inpatient prospective payment system rules. The updated codes are effective October 1 of each year and must be used to report diagnostic or procedure information. (Additional information regarding updates to the ICD-9-CM and DRGs is included in section V.B. of this proposed rule). The official version of the ICD-9-CM is available on CD-ROM from the U.S. Government Printing Office. The FY 2004 version can be ordered by contacting the Superintendent of Documents, U.S. Government Printing Office, Department 50, Washington, D.C. 20402-9329, telephone: (202) 512-1800. The stock number is 017-022-01544-7, and the price is $25.00. In addition, private vendors publish the ICD-9-CM. Questions and comments concerning the codes should be addressed to: Patricia E. Brooks, Co-Chairperson, ICD-9-CM Coordination and Maintenance Committee, CMS, Center for Medicare Management, Purchasing Policy Group, Division of Acute Care, Mailstop C4-08-06, 7500 Security Boulevard, Baltimore, Maryland 21244-1850. Comments may be sent via e-mail to: pbrooks@cms.hhs.gov. 2. DRGs DRGs constitute the patient classification system used in the hospital inpatient prospective payment system. DRGs provide a means of relating the types of patients treated by a hospital to the costs incurred by the hospital. While each patient is unique, groups of patients have demographic, diagnostic, and therapeutic attributes in common that determine their level of resource intensity. Currently, IPF claims include ICD-9-CM diagnosis coding information. The TEFRA payment methodology does not use the DRG classification of IPF cases. Nonetheless, when IPF claims are submitted to us, the DRG associated with the patient's principal ICD-9-CM diagnosis code is assigned to the claim by the GROUPER software program. As a result, our administrative data includes the DRG assignments for all IPF cases. We are proposing to require IPFs to use the psychiatric diagnosis codes in Chapter Five (“Mental Disorders”) of the ICD-9-CM. This decision is consistent with the Standards for Electronic Transaction final rule published in the Federal Register on August 17, 2000 (65 FR 50312). The ICD-9-CM coding system is currently designated as the standard medical data code set for capturing cause and manifestation of injury, disease, impairments, or other health problems. These guidelines are available through a number of sources, including the following Web site: http://www.cdc.gov/nch/data/icdguide.pdf. Current regulations at § 412.27 require that a psychiatric unit admit only those patients who have a principal diagnosis that is listed in the DSM or classified in Chapter Five (“Mental Disorders”) of the ICD-9-CM. The hospital must maintain records that substantiate the psychiatric diagnoses of its patients. We specifically request public comments on continuing to reference the DSM in light of the proposed requirement that IPFs use the ICD-9-CM code set in the proposed IPF prospective payment system. B. Limitations of the DRG System for Psychiatric Patients Adopting a patient classification system for IPFs based on diagnosis alone may not explain the wide variation in resource use among patients in IPFs for several reasons. For instance, the diagnosis may not fully capture the reasons for hospitalization. A patient with a chronic disorder, like schizophrenia, may be admitted for a variety of acute problems (suicide attempt, catatonic withdrawal, or psychotic episode) that require very different treatments (Goldman, H.H., Pincus, H.A., Taube, C.A., and Reiger, D.A. (1984). Hospital and Community Psychiatry, 35(5): 460-464). Further, treatment patterns are more variable in psychiatry, with multiple clinically accepted methods of care. As a result, resource use varies substantially between acute care and chronic care patients, and between the facilities that treat predominately one type of patient. For example, public psychiatric hospitals tend to treat the chronically mentally ill, with substantially longer lengths of stay, compared to the patients generally treated in psychiatric units and private psychiatric hospitals. Predicated on the analysis of the administrative data and pending refinements from the research, we believe the DRG is an appropriate method to account for certain, although not all, clinical characteristics and associated resources. Therefore, under this prospective payment system, we are proposing to assign a DRG to each case based on the principal diagnosis (ICD-9-CM code) reported by the IPF as one adjustment to the Federal per diem base rate. In making this decision, we analyzed past research as well as a recent study supported by the American Psychiatric Association (APA). In the study, APA partnered with the Health Economics and Outcomes Research Institute (THEORI), a division of the Greater New York Hospital Association, to assess whether our existing administrative data could be used to develop a prospective payment system for IPFs. This study found that a prospective payment system for IPFs could be developed based on existing CMS administrative data, be clinically relevant, and limit the administrative burden on providers. The system they proposed included an adjustment for DRG assignment. In summary, we acknowledge that the psychiatric community uses the DSM as a tool to diagnose a patient's mental illness and to aid in treatment planning. However, we are proposing to require IPFs to report diagnoses in Chapter Five of the ICD-9-CM as required by the Administrative Simplification Provisions found in 45 CFR subchapter C. In addition, we are proposing to identify specific DRGs for payment adjustment under the proposed IPF prospective payment system. The rationale for the selection of the proposed DRGs for use in the proposed IPF prospective payment system is described below. C. Proposed DRG Adjustments Under the Proposed IPF Prospective Payment System As noted above, the principal diagnosis is defined as the condition, after study (clinical evaluation), to be chiefly responsible for admitting the patient to the hospital for care. Despite this longstanding definition, our review of hospital claims data that were used to develop the proposed IPF prospective payment system indicates that a substantial number of claims have non-psychiatric diagnoses identified as the principal diagnosis. Medicare regulations as specified in § 412.27(a) require psychiatric units of acute care hospitals to admit only those patients with a principal diagnosis in the DSM or Chapter Five (“Mental Disorders”) in the ICD-9-CM. Therefore, if a patient is admitted to a general hospital for a medical condition such as pneumonia, and also presents psychiatric symptoms, which necessitates an admission to the psychiatric unit, the principal diagnosis for the admission to the psychiatric unit should be the psychiatric symptoms exhibited by the patient in accordance with § 412.27(a). We note that current regulations applicable to psychiatric hospitals (§ 412.23(a)) do not include these requirements, however, historically, psychiatric hospitals have limited admissions to psychiatric patients. Section 412.27(a) also requires that patients be admitted to the psychiatric units for active treatment that is of an intensity that can be furnished appropriately only in an inpatient hospital setting. For this reason, in order to be paid under the proposed IPF prospective payment system, patients must be capable of participating in an active treatment program. In selecting the proposed DRGs for payment adjustment, we analyzed the DRG assignments for ICD-9-CM diagnosis codes in Chapter Five. In addition, as noted previously, IPFs use the DSM-IV-TR to establish diagnoses and current regulations at § 412.27(a) refer to DSM diagnoses. However, most, but not all, DSM codes crosswalk to the codes in Chapter Five of the ICD-9-CM. Although, all the DSM codes are psychiatric, some of the corresponding ICD-9-CM codes are located in other chapters of the ICD-9-CM coding system and are linked to the body system affected. For example, the DSM diagnosis, Male Erectile Disorder, crosswalks to ICD-9-CM code 607.84, Impotence of Organic Nature which is found in Chapter 10, Diseases of the Genitourinary Systems. Accordingly, we also analyzed the DRG assignments for certain ICD-9-CM codes that are based on DSM diagnoses but are not in Chapter Five of the ICD-9-CM. These codes are discussed in the next section of this proposed rule. As a result of this analysis, we identified 25 DRGs with one or more psychiatric diagnoses that are included in Chapter Five of the ICD-9-CM as well as those diagnoses that are in other chapters of the ICD-9-CM. We are proposing payment adjustments for 15 out of the 25 DRGs we analyzed. The remaining 10 DRGs include codes for a specific range of diseases other than psychiatric, but have a few codes for DSM diagnoses that are included in Chapter Five or other body system chapters of the ICD-9-CM. The rationale for our decisions regarding these 10 codes is provided in section II.D. below. Table 1 below lists the DRGs that we are proposing to recognize under the proposed IPF prospective payment system and the proposed adjustment factors. This information also is presented in Addendum A. Table 1.—Proposed IPF Prospective Payment System DRGs DRG Description Adjustment Factor 12 Degenerative Nervous System Disorders 1.07 23 Nontraumatic Stupor and Coma 1.10 424* O.R. Procedure with Principal Diagnosis of Mental Illness 1.22 425 Acute Adjustment Reaction and Psychosocial Dysfunction 1.08 426 Depressive Neurosis 1.00 427 Neurosis Except Depressive 1.01 428 Disorders of Personality and Impulse Control 1.03 429 Organic Disturbances and Mental Retardation 1.02 430 Psychosis 1.00 431 Childhood Mental Disorders 1.02 432 Other Mental Disorder Diagnoses 0.96 433** Alcohol/Drug Abuse or Dependence, Left Against Medical Advice 0.88 521 Alcohol/Drug Abuse or Dependence with Complication or Comorbidity 1.02 522 Alcohol/Drug Abuse or Dependence with Rehabilitation Therapy without Complication or Comorbidity 0.97 523 Alcohol/Drug Abuse or Dependence without Rehabilitation Therapy without Complication or Comorbidity 0.88 * DRG 424—is an O.R. procedure code that must be billed with a principal diagnosis of mental disorder. ** DRG 433—is used when providers indicate a patient left against medical advice (discharge status code 07). D. DRGs Not Recognized in the Proposed IPF Prospective Payment System We are proposing not to recognize the following 10 DRGs in the proposed IPF prospective payment system. They were determined not to be clinically significant because the principal diagnoses did not result in enough admissions to IPFs in order to establish an adjustment to the payment rate: • DRGs 34 and 35 include a range of cases for disorders of the nervous system. The diagnoses in these DRGs also include five ICD-9-CM codes for DSM diagnoses: Codes 333.1 (Tremor not elsewhere classified), code 333.82 (Orofacial Dyskinesia), code 333.92 (Neuroleptic Malignant Syndrome), code 347 (Cataplexy and Narcolepsy), and code 307.23 (Gilles de La Tourette's Disorder). In the 1999 MedPAR records for admissions to IPFs, only one patient was grouped in these DRGs. In addition, patients with these diagnoses generally do not require management in an IPF unless there is a concomitant psychiatric disorder. • DRGs 182, 183, and 184 include a range of gastrointestinal conditions, including esophagitis, gastroenteritis, and other digestive system diseases. The diagnoses in these DRGs include one that is listed in Chapter Five of the ICD-9-CM, code 306.4 (Psychogenic GI Disease). In the 1999 MedPAR records for admissions to IPFs, we found that only a few patients with this ICD-9-CM diagnosis were grouped in these DRGs. • DRG 352 includes a range of diagnoses affecting the testes, prostate, and male external genitalia. This DRG includes DSM diagnoses that are not in Chapter Five of the ICD-9-CM: code 607.84 (Impotence of an Organic Origin), and code 608.89 (Male Genital Diseases, not elsewhere classified). In the 1999 MedPAR records for admissions to IPFs, we were able to identify only one patient grouped in DRG 352. • DRGs 358, 359, and 369 include a range of cases in which procedures have been performed on the uterus and fallopian tubes (Adnexa). These DRGs include two diagnoses that are in Chapter Five of the ICD-9-CM: code 306.51 (Psychogenic Vaginismus), and code 306.52 (Psychogenic Dysmenorrhea). In the 1999 MedPAR records for admissions to IPFs, we were able to identify only 11 patients grouped into DRGs 358, 359, and 369, and there were no patients diagnosed with codes 306.51 or 306.52. • DRG 467 includes a range of cases in which other factors influence health status. This DRG contains only one diagnosis code listed in Chapter Five of the ICD-9-CM, code 305.1 (tobacco use disorder). Patients with this diagnosis do not require inpatient treatment in an IPF unless there is a concomitant psychiatric disorder. We are proposing not to recognize these 10 DRGs for payment adjustments (34, 35, 182, 183, 184, 352, 358, 359, 369, and 467) because they generally do not include a psychiatric diagnosis. We believe that failure to recognize these DRGs will not affect the care of Medicare beneficiaries because our analysis shows few, if any, of the patients with these diagnoses are admitted or treated in an IPF. In addition, we believe that these cases would be classified into one of the selected DRGs and grouped with other beneficiaries with similar symptoms and requiring similar care. This approach would avoid creating case-mix groups based on small numbers of cases. We believe there is value in selecting only those DRGs that contain a large enough number of psychiatric cases to ensure that individual variability can be averaged. We specifically invite public comments on this issue. E. Applicability of the Proposed IPF Prospective Payment System The following psychiatric hospitals and psychiatric units, currently paid under section 1886(b) of the Act, would be paid under the proposed IPF prospective payment system for cost reporting periods beginning on or after April 1, 2004. We are proposing that the IPF prospective payment system would apply to inpatient hospital services furnished by Medicare participating entities that are classified as psychiatric hospitals or psychiatric units as specified in § 412.22, § 412.23, § 412.25, and § 412.27. We note that psychiatric units that are currently paid under the hospital inpatient prospective payment system and do not meet the requirements of § 412.25 and § 412.27 would not be paid under the proposed IPF prospective payment system. As specified in § 400.200, the United States means the fifty States, the District of Columbia, the Commonwealth of Puerto Rico, the Virgin Islands, Guam, American Samoa, and the Northern Mariana Islands. Therefore, IPFs located within the United States would be subject to the proposed IPF prospective payment system. However, the following hospitals are paid under special payment provisions specified in § 412.22(c) and, therefore, would not be paid under the proposed IPF prospective payment system: • 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 Pub. L. 90-248 (42 U.S.C. 1395b-1) or section 222(a) of Pub. L. 92-603 (42 U.S.C. 1395b-1(note)). • Non-participating hospitals furnishing emergency services to Medicare beneficiaries. This proposed rule would not change the basic criteria for a hospital or hospital unit to be classified as a psychiatric hospital or psychiatric unit that is excluded from the hospital prospective payment systems under sections 1886(d) and 1886(g) of the Act, nor would it revise the survey and certification procedures applicable to entities seeking this classification. We note that we are proposing a technical change to § 412.27(a). We are proposing to replace the Third Edition with the Fourth Edition, Text Revision, of the DSM so that our rules reflect the most current edition of the DSM. As noted previously, we are requesting public comments on continuing to require a DSM diagnosis for patients admitted to a psychiatric unit in light of the proposed requirement that IPFs use the ICD-9-CM code set in the proposed IPF prospective payment system. III. Development of the Proposed IPF Per Diem Payment Amount The primary goal in developing the proposed IPF prospective payment system is to pay each IPF an appropriate amount for the efficient delivery of care to Medicare beneficiaries. The system must be able to account adequately for each IPF's case-mix in order to ensure both fair distribution of Medicare payments and access to adequate care for those beneficiaries who require more costly care. The proposed IPF prospective payment system would establish a standard per diem payment amount for inpatient psychiatric services provided to Medicare beneficiaries. The proposed per diem amount would reflect the average daily cost of inpatient psychiatric care in an IPF, including capital-related costs. This proposed per diem payment amount, after adjustment for budget neutrality, is then modified by factors for patient and facility characteristics that account for variation in patient resource use. The proposed IPF prospective payment system would also include an outlier policy and account for interrupted stays. This section includes a discussion of how the proposed Federal per diem base rate was created, the factors that we considered to adjust the proposed Federal per diem base rate, and how the proposed per diem payment amount is calculated. A. Proposed Market Basket We are proposing to use a 1997-based excluded hospital with capital market basket. We periodically revise and rebase the market basket to reflect more current cost data. Rebasing means moving the base year for the structure of costs (in this case from 1992 to 1997), while revising means changing data sources, cost categories, or price proxies used. The proposed updated market basket would replace the 1992-based excluded hospital with capital market basket. This rebased (1997-base year) and revised market basket would be used to update FY 1999 IPF costs to the proposed 15-month period beginning April 1, 2004, the first year under the IPF prospective payment system. 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 hospital inpatient prospective payment system 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 Medicare cost reports, the 1997 Business Expenditure Survey, and the 1997 Annual Input-Output data from the Bureau of the Census. As explained in that August 1, 2002 final rule, we revised the market basket by making two methodological revisions to 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. 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, it holds that 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. Two sets of weights for the capital portion of the revised and rebased market basket needed to be determined. 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 hospital inpatient prospective payment final rule published in the Federal Register on August 1, 2002 (67 FR 50045 through 50047), for a discussion of how vintage weights are determined). The cost categories, price proxies, and base-year FY 1992 and proposed FY 1997 weights for the excluded hospital with capital market basket are presented in Table 2 below. The vintage weights for the proposed 1997-based excluded hospital with capital market basket is presented in Table 2(A) below. TABLE 2.—Proposed Excluded Hospital With Capital Input Price IndeX (FY 1992 and Proposed FY 1997) Structure and Weights Cost category Price wage variable Weights (%) base-year 1992 Proposed weights (%) base-year 1997 TOTAL 100.000 100.000 Compensation 57.935 57.579 Wages and Salaries ECI—Wages and Salaries, Civilian Hospital Workers 47.417 47.355 Employee Benefits ECI—Benefits, Civilian Hospital Workers to capture total costs (operating and capital), In order to capture total costs (operating and capital), HCFA Occupational Benefit Proxy 10.519 10.244 Professional fees: Non-Medical ECI—Compensation: Prof. & Technical 1.908 4.423 Utilities 1.524 1.180 Electricity WPI—Commercial Electric Power 0.916 0.726 Fuel Oil, Coal, etc. WPI—Commercial Natural Gas 0.365 0.248 Water and Sewerage CPI-U—Water & Sewage 0.243 0.206 Professional Liability Insurance HCFA—Professional Liability Premiums 0.983 0.733 All Other Products and Services 28.571 27.117 All Other Products 22.027 17.914 Pharmaceuticals WPI—Prescription Drugs 2.791 6.318 Food: Direct Purchase WPI—Processed Foods 2.155 1.122 Food: Contract Service CPI-U—Food Away from Home 0.998 1.043 Chemicals WPI—Industrial Chemicals 3.413 2.133 Blood and Blood Products WPI—Blood and Derivatives 0.748 Medical Instruments WPI—Med. Inst. & Equipment 2.868 1.795 Photographic Supplies WPI—Photo Supplies 0.364 0.167 Rubber and Plastics WPI—Rubber & Plastic Products 4.423 1.366 Paper Products WPI—Convert. Paper and Paperboard 1.984 1.110 Apparel WPI—Apparel 0.809 0.478 Machinery and Equipment WPI—Machinery & Equipment 0.193 0.852 Miscellaneous Products WPI—Finished Goods excluding Food and Energy 2.029 0.783 All Other Services 6.544 9.203 Telephone CPI-U—Telephone Services 0.574 0.348 Postage CPI-U—Postage 0.268 0.702 All Other: Labor ECI—Compensation: Service Workers 4.945 4.453 All Other: Non-Labor Intensive CPI-U—All Items (Urban) 0.757 3.700 Capital-Related Costs 9.080 8.968 Depreciation 5.611 5.586 Fixed Assets Boeckh-Institutional Construction: 23 Year Useful Life Life Y_y_YYF e. 3.570 3.503 Movable Equipment WPI—Machinery & Equipment: 11 Year Useful life 2.041 2.083 Interest Costs 3.212 2.682 Non-profit Avg. Yield Municipal Bonds: 23 Year Useful Life 2.730 2.280 For-profit Avg. Yield AAA Bonds: 23 Year Useful Life 0.482 0.402 Other Capital Related Costs CPI-U—Residential Rent 0.257 0.699 Note: The operating cost category weights in the proposed excluded hospital market basket add to 100.0. When we add an additional set of cost category weights (total capital weight = 8.968 percent) to this original group, the sum of the weights in the new index must still add to 100.0. Because capital costs account for 8.968 percent of the market basket, then operating costs account for 91.032 percent. Each weight in the proposed 1997-based excluded hospital market basket was multiplied by 0.91032 to determine its weight in the proposed 1997-based excluded hospital with capital market basket. Note: Weights may not sum to 100.0 due to rounding. Table 2(A).—Proposed Excluded Hospital With Capital Input Price Index (FY 1997) Vintage Weights Year from farthest to most recent Fixed assets (23-year weights) Movable assets (11-year weights) Interest: capital-related (23-year weights) 1 0.018 0.063 0.007 2 0.021 0.068 0.009 3 0.023 0.074 0.011 4 0.025 0.080 0.012 5 0.026 0.085 0.014 6 0.028 0.091 0.016 7 0.030 0.096 0.019 8 0.032 0.101 0.022 9 0.035 0.108 0.026 10 0.039 0.114 0.030 11 0.042 0.119 0.035 12 0.044 0.039 13 0.047 0.045 14 0.049 0.049 15 0.051 0.053 16 0.053 0.059 17 0.057 0.065 18 0.060 0.072 19 0.062 0.077 20 0.063 0.081 21 0.065 0.085 22 0.064 0.087 23 0.065 0.090 Total 1.0000 1.0000 1.0000 Note: Weights may not sum to 1.000 due to rounding. Table 2(B) below compares the 1992-based excluded hospital with capital market basket to the proposed 1997-based excluded hospital with capital market basket. As shown below, the rebased and revised market basket grows slightly faster over the 1999 through 2001 period than the 1992-based market basket. The main reason for this growth is the switching of the wage and benefit proxy to the ECI for hospital workers from the previous occupational blend. This revision had a similar impact on the hospital inpatient prospective payment system and excluded hospital market baskets, as described in the final rule published in the Federal Register on August 1, 2002 (67 FR 50032 through 50041). Table 2(B).—Percent Changes in the 1992-Based and Proposed 1997-Based Excluded Hospital With Capital Market Baskets, FYs 1999 Through 2004 Fiscal year Percent change, 1992-based market basket Percent change, proposed 1997-based market basket 1999 2.3 2.7 2000 3.4 3.1 2001 3.9 4.0 Average historical: 3.2 3.3 2002 2.7 3.6 2003 3.0 3.5 2004 3.0 3.3 Average forecast: 2.9 3.5 Source: Global Insights, Inc, 4th Qtr 2002,@ USMARCO.MODTREND@CISSIM/TL1102.SIM. Historical data through 3rd Qtr 2002. Based upon the analysis mentioned below, we believe the excluded hospital with capital market basket provides a reasonable measure of the price changes facing IPFs. However, we have also been researching the feasibility of developing a market basket specific to IPF services. This research includes analyzing data sources for cost category weights, specifically the Medicare cost reports, and investigating other data sources on cost, expenditure, and price information specific to IPFs. Our analysis of the Medicare cost reports indicates that the distribution of costs among major cost report categories (wages, pharmaceuticals, and capital) for IPFs is not substantially different from the 1997-based excluded hospital with capital market basket we propose to use. In addition, the only data available to us for these cost categories (wages, pharmaceuticals, and capital) presented a potential problem since no other major cost category weights would be based on IPF data. Based on the research discussed below, at this time, we are not proposing to develop a market basket specific to IPF services. 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 excluded hospital with capital market basket. Other cost categories were recalibrated using ratios available from the hospital inpatient prospective payment system hospital market basket. 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 hospital inpatient prospective payment system update framework. Based upon this analysis, we believe that the excluded hospital with capital market basket is doing an adequate job of reflecting the price changes facing IPFs. We will continue to solicit comments about issues particular to IPFs that should be considered in our development of the proposed 1997-based excluded hospital with capital market basket, as well as encourage suggestions for additional data sources that may be available. Our hope is that the additional cost data being collected under the proposed IPF prospective payment system will eventually allow for the development of a market basket based primarily on IPF data. We welcome comments on issues particular to IPFs that should be considered in our use of the proposed 1997-based excluded hospital with capital market basket, as well as on suggestions for additional data sources that may be readily available on the cost structure of IPFs. As discussed more fully in section IV of this proposed rule, we are proposing to implement the proposed IPF prospective payment system for IPF cost reporting periods that begin on or after April 1, 2004. The first update, however, would not be until July 1, 2005. This extends the first year for 3 additional months in order to adjust the update cycle for this proposed payment system. As a result, the effective period for this proposed rule is April 1, 2004 through June 30, 2005. To update payments between FY 2003 and the effective period, the update must reflect the market basket increase over this period, which is currently estimated at 5.3 percent. This would represent the proposed increase in the excluded hospital with capital market basket for FY 2004 and the first 9 months of FY 2005. B. Development of the Proposed Case-Mix Adjustment Regression In order to ensure that the proposed IPF prospective payment system 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. Based on the regression analysis, we are proposing to adjust the per diem payment for differences in the patient's DRG, age, comorbidities, and the day of the stay. Also, we are proposing adjustments for area wage levels, rural IPFs, and teaching IPFs. We computed a per diem cost for each Medicare inpatient psychiatric stay, including routine operating, ancillary, and capital components using information from the 1999 MedPAR file and data from the 1999 Medicare cost reports. The method described below that was used to construct the proposed per diem cost for IPFs is a standard method that has been used to construct a Medicare cost per discharge for inpatient acute care (Newhouse, J.P., S. Cretin, and C. Witsberger. Predicting Hospital Accounting Costs, Health Care Financing Review, V.11, No. 1. Fall 1989). We believe that this method provides a full account of IPF's per diem costs. 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 1999 Medicare cost report by the number of Medicare covered days on the 1999 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. We used the best available data and methods for this proposed IPF prospective payment system. However, the data are potentially limited for the purpose of determining the extent to which differences in patient characteristics influence the per diem cost of inpatient psychiatric care. This potential limitation results from Medicare cost accounting practices in which routine per diem costs are calculated as an average and, therefore, do not vary among patients within a facility (that is, a patient requiring intensive staff attention is assigned the same routine cost as a patient requiring little staff attention). This potential limitation assumes heightened importance for IPFs because routine costs represent about 88 percent of total costs. As a result, our cost measure may not capture the degree of variation in routine cost attributable to differences in patient characteristics. Patient differences are reflected in our measure of routine cost only to the extent that facilities tend on average to treat different proportions of patients with differing routine resource needs. For example, one IPF may have higher routine per diem costs because it treats a higher proportion of older patients (or patients who require continuous monitoring) than another IPF. However, our cost variable will not measure the extent to which older patients within the same IPF are more costly than younger patients. We are currently conducting a research study with the RTI International ® (trade name of Research Triangle Institute) that will provide information as to the effects of this data limitation. As a result, we expect to have more information about the extent to which routine costs vary by certain patient characteristics. We solicit suggestions on other data sets or studies that could provide additional information on the relationship between individual patients and average facility routine costs. This routine cost limitation does not apply to ancillary costs because they can be measured at the patient level using Medicare claims as reported in the MedPAR file. However, there are differences in charging practices between psychiatric hospitals and psychiatric units that affect our measurement of ancillary costs. For example, there are approximately 100 hospitals in our MedPAR data file that do not bill ancillary charges; the majority of these providers are State psychiatric hospitals who bill a single average per diem rate that includes routine, ancillary, and other costs. The proposed payment adjustors were derived from regression analysis of 100 percent of the 1999 MedPAR data file. The MedPAR data file used for the final regression contains 467,372 cases although the complete file contains 476,541 cases. We deleted 5,822 cases (1.24 percent) from this file because routine cost data for certain IPFs was not available. In order to include as many IPFs as possible in the regression, we substituted the 1998 Medicare cost report data for routine cost and ancillary cost to charge ratios (using the 1998 Medicare cost report data). For the remaining 470,719 cases, we used the following method to trim extraordinarily high or low cost values that most likely contained data errors, in order to improve the accuracy of our results. The means and standard deviations of the logged per diem total cost were computed separately for cases from psychiatric hospitals and psychiatric units. Separate statistics were computed for the groups of IPFs, because we did not want to systematically exclude a larger proportion of cases from the higher cost psychiatric units. Before calculating the means of the logged per diem total cost, we trimmed cases from the file when covered days were zero, or routine costs were less than $100 or greater than $3,000, (because we believe this range captured the grossly aberrant cases), so that the means would not be distorted. We trimmed cases when the logged per diem cost was outside the standard and generally used statistical trim points of plus or minus 3 standard deviations from the respective means for hospitals and psychiatric units. These criteria eliminated another 3,347 cases, leaving 467,372 cases that were used in the final regression. 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. To control for psychiatric hospitals that do not bill ancillary costs, we included a categorical variable that identified them. The proposed per diem cost was adjusted for differences in labor cost across geographic areas using the FY 1999 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 proposed wage adjustment factor for each case by multiplying the Medicare hospital wage index for each facility by the proposed labor-related share (.72828) and adding the proposed non-labor share (.27172). We used the proposed excluded hospital with capital market basket to determine the labor-related share (see section III.A. of this proposed rule). 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 proposed payment adjustment for teaching facilities, 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 proposed payment adjustment factors represent the proportional effect of each variable relative to a reference variable. 1. Proposed Patient-Level Characteristics Subject to the limitations of the proposed cost variable described above and the availability of patient characteristic information contained in the administrative data, we attempted to use patient characteristics to explain the cost variation amongst IPFs. By adjusting for DRGs, comorbidities, age, and day of the stay, we were able to explain approximately 19 percent of the variation in the per diem cost. This result is comparable to that obtained by THEORI in the analysis they conducted for the APA. The study is described in section II.B. of this proposed rule. a. DRGs The principal diagnosis ICD code listed on the claim is used to assign each case to one of the 15 DRGs that we are proposing to recognize in this IPF prospective payment system (see section II.C of this proposed rule). The coefficients of these DRGs from the cost regression analysis were used to determine the magnitude of the payment adjustment for each of the proposed 15 DRGs. The payment adjustments are expressed relative to the most frequently assigned DRG (DRG 430, Psychoses). That is, the proposed adjustment factor for DRG 430 would be 1.00, and the proposed adjustment factors for the other 14 DRGs would vary above and below 1.00. For 8 DRGs, the proposed adjustments would be relatively small (between .96 and 1.04, that is, between 4 percent lower to 4 percent higher). The following 4 DRGs would receive relatively large payment adjustments: • DRG 424 (Surgical procedure with Principal Diagnosis of Mental Illness) would have the largest payment adjustment of approximately 1.22. • DRG 023 (Non-traumatic stupor and coma) would receive an adjustment of approximately 1.10. • DRG 425 (Acute Adjustment Reaction and Psychosocial Dysfunction) would receive an adjustment of approximately 1.08. • DRG 12 (Degenerative Nervous System Disorders) would receive an adjustment of approximately 1.07. Both of the following two DRGs would be paid substantially less than DRG 430 with payment adjustments of approximately 0.88: • DRG 433 (Alcohol/Drug Abuse or Dependence, left against medical advice). • DRG 523 (Alcohol/Drug Abuse or Dependence, without Complications and/or Comorbidity and without Rehabilitation Therapy). Cases in our MedPAR data file whose principal diagnosis classified them in DRGs other than one of the 15 DRGs that we are proposing to recognize in this proposed IPF prospective payment system were grouped into a single “other” category. b. Comorbidities Our analysis of the data indicates that patients who have certain comorbid conditions in addition to their psychiatric condition generally require more expensive care while they are hospitalized. After a thorough review of the ICD-9-CM codes, some comorbid conditions were identified as being more costly on a per diem basis. Groups of similar diagnosis codes were created to describe these conditions, which tend to be chronic illnesses that require additional medications, supplies, laboratory, or diagnostic testing in addition to the care provided for their psychiatric condition. Conditions in which the patient is acutely ill requiring care in a general hospital, for example, myocardial infarction, were not included in our analysis. Based upon this analysis, we are proposing payment adjustments for 17 comorbidity categories that we would recognize for payment adjustments under the proposed IPF prospective payment system. Table 3 below provides a listing of the proposed comorbidity categories, the ICD-9-CM diagnostic codes comprising each category, and the payment adjustment factors. The adjustment factors are also in Addendum A. As in the case of the DRGs, the cost regression analysis was used to determine the magnitude of the proposed payment adjustments for the comorbidity groups. Of the 17 comorbidity categories, the following 4 groups would have proposed payment adjustment factors ranging from 1.11 to 1.17 more than a case that did not have any of the 17 comorbid conditions: (1) Coagulation factor deficits; (2) renal failure, chronic; (3) chronic cardiac conditions; and (4) atherosclerosis of extremity with gangrene. Seven categories would be paid payment adjustments from 1.08 to 1.14: (1) Tracheotomy; (2) renal failure, acute; (3) malignant neoplasms; (4) severe protein calorie malnutrition; (5) chronic obstructive pulmonary disease; (6) poisoning; and (7) severe musculoskeletal and connective tissue diseases. The remaining 6 comorbidity categories would receive payment adjustments ranging from 1.03 to 1.10: (1) HIV; (2) infectious diseases; (3) uncontrolled type I diabetes mellitus; (4) artificial openings digestive and urinary; (5) drug and/or alcohol induced mental disorders; and (6) eating and conduct disorders. Other potential conditions were considered as potentially more expensive, but the small number of cases in the MedPAR data file made it impossible to propose an appropriate adjustment for those conditions. We solicit comments suggesting other conditions that may be expected to increase the per diem cost of care in IPFs. In addition, we expect that as facilities become aware of the importance of providing accurate information on the diagnoses of patients, we will have more data to use as a basis for refinements to the list of proposed comorbid conditions affecting the per diem cost of care. Table 3.—Diagnosis Codes for Proposed Comorbidity Categories Description of proposed comorbidity ICD-9-CM code Proposed adjustment factor HIV 042 1.06 Coagulation Factor Deficits 2860 through 2864 1.11 Tracheotomy 51900 and V440 1.14 Renal Failure, Acute 5846 through 5849; 7885; 9585; V451; V560, V561; and V562 1.08 Renal Failure, Chronic 40301; 40311; 40391; 40402; 40412; 40492, 585; and 586 1.14 Malignant Neoplasms 1400 through 1720; 1740 through 1840; and 1850 through 2080 1.10 Uncontrolled Type I Diabetes-Mellitus, with or without complications 25003; 25083; 25013; 25023; 25033; 25093; 25043; 25053; 25063; and 25073 1.10 Severe Protein Calorie Malnutrition 260 through 262 1.12 Eating and Conduct Disorders 3071; 30750; 31203; 31233; and 31234 1.03 Infectious Diseases 01000 through 04110; 04500 through 05319, 05440 through 05449; 0550 through 0770; 0782 through 0789; and 07950 through 07595 1.08 Drug and/or Alcohol Induced Mental Disorders 2920; 2922; 2910; 29212; 30300; and 30400 1.03 Cardiac Conditions 3910; 3911; 3912; 40201; 41403; 4160; and 4210 1.13 Atherosclerosis of Extremity with Gangrene 44024 1.17 Chronic Obstructive Pulmonary Disease 5100; 51883; 51884; 4920; 494; 49120 through 49122, and V461 1.12 Artificial Openings-Digestive and Urinary 56960; V441 through V443; and V4450 1.09 Severe Musculoskeletal and Connective Tissue Diseases 6960; 7100; 73000 through73009; 73010 through 73019; 73020 through 73029; and 7854 1.12 Poisoning 96500 through 96509; and 9654; 9670 through 9700; 9800 through 9809; 9830 through 9839; 986; 9890 through 9897 1.14 c. Patient Age and Gender The cost regressions explored several alternative configurations of age and gender variables. The results indicate that the per diem cost rises as a patient's age increases, and the per diem cost are higher for female patients. We examined the variation in the per diem cost for 5-year age intervals ranging from age 40 to 80 with open-ended categories ranging above age 80 and below 40 and determined that the effect of age was statistically significant. We initially ran the regression for three age groups consistent with the natural breaks in the distribution of age (under 55, 55 to 64, and 65 and over). The distribution showed that most Medicare psychiatric patients are under age 55 and over age 65. In addition, the distribution showed that the age group between 55 and 65 years of age increased the predictive power of the model only by a factor of .002 percent because there were few patients in that age category. For this reason, we are not proposing adjustments reflecting the three age groups. Rather, we are proposing to make a single adjustment of 13 percent for patients 65 years and over. We are proposing two age groups (under 65 and over 65) to correspond with the major populations within Medicare: the disabled and the elderly, which we believe are largely responsible for the age-related cost differences that we observed. In addition, preliminary results from the RTI International ® research that used estimates of patient-specific routine cost per day (from a sample of 40 IPFs) found that splitting age into two groups (under 65 and over 65) has greater explanatory power than alternative age group configurations. The research study is described in more detail in section V.C.1. of this proposed rule. The cost regression implies that female patients are approximately 3 percent more costly than male patients. However, the explanatory power of the equation increases by less than .002 percentage points. There is also a small reduction in the age effect for the 65 and over age group (less than one percentage point). We also examined the alternative of including gender along with the three age groups (under 55, 55 to 64, and 65 and over) and compared the results to the regression without gender and with two age groups (under 65 and 65 and over). The fuller specification of age and gender only increased the explanatory power by .003 points and had little effect on the size of the age effects. We know that the elderly and women are more frequently treated in psychiatric units than in freestanding psychiatric hospitals. When an indicator variable for psychiatric units is included in the cost regression, the age and gender effects decrease (the 65 and over age effect declines from approximately 13 percent to approximately 9 percent, and the gender effect decreases from approximately 3 percent to 2 percent). We are unable to determine the extent to which this 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, RTI International's ® preliminary results, which used a better patient-specific cost variable for a sample of 40 hospitals found a much stronger effect for age than for gender. This is because the evidence currently available to us is limited and we believe we cannot identify a direct link between the costs of psychiatric care in psychiatric units and treatment of female IPF patients. We are not proposing to adjust the per diem payment rate to account for gender. We invite comments on the appropriateness of including a gender variable as a payment adjustment as well as comments on the age categories used to identify variations in costs. We will continue to assess the effects of gender and age as we analyze more current data in the development of the final rule. d. Length of Stay Cost regressions indicate that the per diem cost declines as the length of stay increases. We are proposing adjustments to account for ancillary and certain administrative costs that occur disproportionately in the first days after being admitted to an IPF (the variable per diem adjustments). We examined the per diem cost over a range of 1 to 14 days. According to the 1999 MedPAR data file, the per diem costs were highest on day 1 and declined for days 2 through 8 as indicated below. Per diem costs for days 9 and thereafter remained relatively consistent with the median length of stay in an IPF for Medicare beneficiaries. The cost regression analysis was used to determine the following proposed payment adjustments. Relative to a stay of 9 or more days, the resulting adjustments for the first 8 days of a stay that we are proposing to use in this IPF prospective payment system are as follows: • The variable per diem adjustment for day 1 would be an increase of approximately 26 percent. • The variable per diem adjustment for days 2 to 4 would be an increase of approximately 12 percent. • The variable per diem adjustment for days 5 to 8 would be an increase of approximately 5 percent. • No variable per diem adjustment would be paid after the 8th day. The higher payments for earlier days are offset through the budget neutrality adjustment, which has the effect of lowering the average payment to account for the increased payments. 2. Proposed Facility-Level Characteristics As noted earlier, we were able to explain 19 percent of the variation in wage-adjusted per diem cost using patient characteristics. We explored a variety of ways to incorporate facility characteristics into the cost regressions in order to raise the explanatory power and refine the proposed payment system to better align payments with cost differences across facility types. Per diem costs are strongly related to facility occupancy, because occupancy (as measured by the ratio of actual days to available days) measures the extent to which the facility is efficiently utilizing its capacity. When occupancy is low, fixed costs must be spread across relatively few days of care and the per diem costs are high. Because we do not want to pay for inefficiency, we are not proposing that occupancy be used as a payment adjuster. However, this variable is included in the cost regression to improve the estimates of the effects of other factors that may more appropriately be used to adjust payments. An analysis of the facility-level characteristics we considered follows. To summarize the analysis, we are proposing that payments be adjusted based on the IPF's wage index, rural location, and teaching status. We considered, and explain below, the reasons why we are proposing not to provide adjustments for psychiatric units, disproportionate share intensity, or IPFs in Alaska or Hawaii. a. Rural Location We found that, controlling for the patient characteristics and other facility variables included in our cost regression, facilities located in non-metropolitan area counties had per diem costs about 16 percent higher than facilities located in metropolitan area counties. Most of the higher cost of rural IPFs is related to the fact that the vast majority are psychiatric units within small general acute care hospitals. Small-scale facilities are more costly on a per diem basis because there are minimum levels of fixed costs that cannot be avoided. Based on this analysis, we are proposing to make an adjustment of 16 percent for IPFs located in rural areas. b. Teaching Status One option for paying psychiatric teaching facilities for their higher costs relies on past experience with the teaching adjustment for other Medicare prospective payment systems. As in other inpatient prospective payment systems, we measured teaching status as one plus the ratio of the number of interns and residents assigned to the facility divided by the IPF's average daily census (ADC). Similarly for psychiatric units, we used the number of interns and residents assigned to the psychiatric unit. The advantages of using the ADC rather than the number of beds for the denominator of the ratio noted above was discussed in the final rule we published in the Federal Register on August 30, 1991 (56 FR 43380) for putting inpatient hospital capital payments under a prospective payment. As described in that rule, the two key advantages of the ADC are that it is—(1) easier to define more precisely than number of beds; and (2) less subject to understatement in an effort to increase the size of the teaching variable. We believe that these advantages apply equally to IPFs. The teaching variable in our cost regressions, that is, the logarithm of one plus the ratio of interns and residents to ADC, has a coefficient value of .5215. This cost effect is converted to a payment adjustment by treating the regression coefficient as an exponent and raising the teaching variable to the .5215 power. Applying this method for a facility with a teaching variable of 1.10 would yield a 5.1 percent increase in the per diem payment; for a facility with a teaching variable of 1.25, there would be a 12.3 percent higher payment. Our impact tables are based on the assumption that we would pay a proposed IPF teaching adjustment in this manner and our proposed regulatory text is also based on this approach. However, we are considering alternatives because we are concerned that this method creates incentives for teaching hospitals to add residents and to increase their payments under an open-ended formula that pays higher teaching payments as teaching intensity, as measured by resident to ADC ratios, increases. The BBA, sections 4621 and 4623, limited the incentives to add residents in hospitals paid under the hospital inpatient prospective payment system by adopting caps for both direct and indirect teaching payments. The number of residents was capped for the purpose of computing both the direct and indirect teaching adjustments and the resident to ADC was capped for purposes of computing the indirect teaching adjustment. Because IPFs would now be paid on a prospective basis similar to acute care hospitals, we are considering extending the indirect teaching caps to IPF teaching hospitals. Regulations, as specified at § 413.86, already apply the BBA caps to direct medical education payments for all teaching hospitals. We are also exploring whether there are other alternatives for paying IPF teaching hospitals their higher teaching costs. We are interested in developing methodologies for estimating these higher costs and then, based on the newly available estimates and current data, distributing those costs fairly to individual teaching hospitals. We invite comments on obtaining the estimates and current data and on other approaches to paying psychiatric teaching hospitals for their higher medical-education costs based on that data. c. Disproportionate Share Hospital Status We measured the extent to which a facility provides care to low income patients using the disproportionate share hospital (DSH) variable used in other Medicare prospective payment systems (that is, the sum of the proportion of Medicare days of care provided to recipients of Supplemental Security Income and the proportion of the total days of care provided to Medicaid beneficiaries). For psychiatric units, both proportions are specific to the unit and not the entire hospital. A limitation of the Medicaid proportion as applied to psychiatric hospitals is that Medicaid does not pay for services provided to individuals under the age of 65 in an institution for mental diseases (IMD), as specified in section 1905(h) of the Act. As a result, low-income beneficiaries in IMDs cannot be identified as Medicaid beneficiaries, and the Medicaid proportion will be biased downwards. The DSH variable was highly significant in our cost regressions; however, we found that facilities with higher DSH had lower per diem costs. We note that the previously cited study for the APA also found the same results. The relationship of high DSH with lower costs cannot be attributed to downward bias in the Medicaid proportion due to the IMD exclusion. This is because public psychiatric hospitals already have lower costs on average than other types of IPFs. Therefore, if we propose a DSH adjustment based on the regression analysis, IPFs with high DSH shares would be paid lower per diem rates. We tried a variety of supplemental analyses in an attempt to better understand the observed relationship, but did not find a positive relationship between the per diem cost and the DSH ratio. Therefore, we are not proposing a payment adjustment for DSH intensity but will monitor the effect of DSH for possible future adjustments. d. Psychiatric Units in General Acute Care Hospitals On average, psychiatric units have higher per diem costs than psychiatric hospitals. According to the 1999 MedPAR file, the average per diem cost for psychiatric units was $615, compared to $444 for psychiatric hospitals. Some of the patient characteristics and facility variables that we included in our cost regressions explain part, but not all, of the cost difference between hospitals and psychiatric units. Controlling for facility size, occupancy, and selected comorbidities reduces the magnitude of the estimated cost difference from approximately 37 percent to 19 percent. Several factors may account for the remaining 19 percent difference: (1) A large proportion of psychiatric admissions to these units enter the hospital through the emergency room (ER), and ER charges are included on the inpatient claims used in our analysis (this issue will not be relevant to IPF payment in the future because ER services have been paid under the outpatient hospital prospective payment system since August 2000); (2) some of these admissions have medical conditions in addition to psychiatric symptoms and require more treatments resulting in higher costs due to more services and equipment; (3) psychiatric hospitals and psychiatric units may utilize different patterns of care and staffing; and (4) accounting differences may account for some of the cost difference. We have decided not to propose a specific adjustment for psychiatric units. We are concerned about applying such an adjustment to all psychiatric units regardless of an individual unit's costs, efficiency, or case mix. We hope that with further research, we will be able to gain a better understanding of the cost differences that would enable us to propose even more refined payment adjustments to directly measure the differences in patient care needs in psychiatric units. e. Adjustment for Alaska and Hawaii IPFs Some of the prospective payment systems that have been developed include a cost-of-living adjustment for the unique circumstances of Medicare providers located in Alaska and Hawaii. Therefore, we analyzed our data to determine the existence of IPFs located in Alaska and Hawaii. Currently, in Alaska, there are only two psychiatric hospitals and no psychiatric units. In Hawaii, there is one psychiatric hospital and one psychiatric unit. In the absence of a cost-of-living adjustment, our analyses indicates that some facilities in Alaska and Hawaii would “profit” and other facilities would experience a “loss.” Due to the limited number of cases, the results of our analysis are inconclusive regarding whether a cost-of-living adjustment would improve payment equity for these facilities. Therefore, we are not proposing an adjustment for IPFs located in Alaska and Hawaii. We will continue to assess the impact of the proposed IPF prospective payment system on IPFs located in Alaska and Hawaii as we obtain more current data. 3. Proposed Payment Adjustments a. Proposed Outlier Adjustment While we are not statutorily required to provide outlier payments, we believe that it is appropriate to propose an outlier payment policy in connection with this prospective payment system in order to both ensure that IPFs treating unusually costly cases do not incur substantial “losses” and promote access to IPFs for patients requiring expensive care. Providing additional payments for costs that are beyond the IPF's control can strongly improve the accuracy of the proposed IPF prospective payment system in determining resource costs at the patient and facility level. Notwithstanding the factors that we are proposing to recognize in the IPF prospective payment system as proposed adjustments to the per diem payment rate, the cost of care for some psychiatric patients may still substantially exceed the otherwise applicable payments during the course of a stay. This may occur because of multiple comorbid conditions and complications that require a high utilization of ancillary services. Since this is a per diem payment system, the extent to which length of stay is a factor would be mitigated because payment is made for each day of the stay. We have determined that it is important to provide some protection from financial risk caused by treating patients who require more costly care and to reduce the incentives to under serve these patients. Therefore, in order to protect IPFs from significant “losses” on very costly cases, we are proposing to provide outlier payments and set outlier numerical criteria prospectively so that outlier payments are projected to equal 2 percent of total payments under the proposed IPF prospective payment system. Based on the regression analysis and payment simulations, we believe that using a 2 percent threshold optimizes our ability to protect vulnerable IPFs while providing adequate payment for all other cases that are not outlier cases. We are proposing, in § 412.424(c), to make an outlier payment for any case in which the estimated total cost exceeds an outlier threshold amount equal to the total IPF prospective payment system payment amount plus a fixed dollar loss amount. The fixed dollar loss amount is the amount used to limit the loss that an IPF would incur under the proposed outlier policy (see section III.C.3. of this proposed rule for an explanation of how the fixed dollar loss amount is calculated). Once the cost of a case exceeds the outlier threshold amount, an outlier payment would be made. A basic principle of an outlier policy is that outlier payments should cover less than the full amount of the additional costs above the outlier threshold in order to preserve the incentive to contain costs once a case qualifies for outlier payments (see Emmett B. Keeler, Grace M. Carter, and Sally Trude, “Insurance Aspects of DRG Outlier Payments,” The Rand Corporation, N-2762-HHS, October 1988). This results in Medicare and the IPF sharing financial risk in the treatment of extraordinarily costly cases. b. Methodology for Proposed Outlier Payments We are proposing to make outlier payments on a per case basis rather than on a per diem basis. Outlier payments would be made for IPF cases when the estimated cost of the entire stay exceeds the outlier threshold amount. We believe it is appropriate to determine outlier status on a per case basis in order to accurately assess the “losses” associated with the care of a patient for the entire stay. If we propose to establish a per diem fixed dollar loss threshold, outlier payments could occur for part of an inpatient stay when no “losses” actually occur. If we review the stay in terms of the resources expended each day, the facility may incur a “loss” on some days of the stay and may experience “gains” on other days of the stay. Thus, assessing the resources expanded over the course of the entire stay provides a fuller picture of the actual resources needed to provide care for the complete episode of care. After assessing the entire stay, one can determine if a “loss” was actually incurred by the IPF. Therefore, we are proposing to define the outlier threshold amount as the total IPF prospective payment for an IPF stay, plus a fixed dollar loss amount. As explained below, the fixed dollar amount is determined to be the dollar amount per stay that achieves a total outlier percentage of 2 percent of the proposed prospective payments. The proposed outlier payment would be defined as a proportion of the estimated cost beyond the outlier threshold. The proportion of additional costs paid as outlier payments is referred to as the loss-sharing ratio. We chose to propose the fixed dollar loss amount and the loss-sharing ratios to allow the estimated total outlier payments to be 2 percent of the total estimated proposed IPF prospective payments. In order to determine the most appropriate outlier policy, our goal was to analyze the extent to which the various outlier percentages reduce financial risk, reduce incentives to under serve costly beneficiaries, and improve the overall fairness of the payment system. Our analysis showed that the higher the outlier percentage, the more cases qualified for outlier payments, and the less payment was made per case. Conversely, a low outlier percentage resulted in a higher fixed dollar loss threshold and although fewer cases exceeded the threshold, the amount paid was more substantial. We began our analysis by determining that if approximately 10 percent of IPF cases received an outlier payment, we would be maintaining the basic premise behind establishing an outlier policy, that is, to compensate IPFs for their truly high cost cases. Also, this percentage of cases, that is 10 percent, is not inconsistent with the percentage of total outlier cases paid in other prospective payment systems. Initially, we believed that a 5 percent outlier policy would result in outlier payments for approximately 10 percent of total IPF cases. However, our analysis showed that a 5 percent outlier policy resulted in outlier payments for approximately 20 percent of IPF cases, paying an average of $1,975 per case. Since 20 percent of IPF cases would receive an outlier payment, we do not believe that a 5 percent outlier policy limits outlier payments to only the truly high cost cases. We then reduced the outlier policy to 3 percent and found that 12 percent of IPF cases received outlier payments, with an average payment of $2,125 per case. Although a 3 percent outlier policy reduced the number of cases that would qualify for outlier payments, 12 percent of cases still exceeded our target of 10 percent of total IPF cases. However, we have determined that an outlier policy of 2 percent of the total proposed IPF payments would allow us to achieve a balance of the above stated goals. A 2 percent outlier policy would appropriately compensate for the truly high cost cases with a much more appropriate level of payment and reduced financial risk without causing a significant reduction in the per diem base rate. Under a 2 percent outlier policy, approximately 7 percent of IPF cases qualify for outlier payments with an average payment of $2,350 per case. Providing outlier payments to 7 percent of cases meets the 10 percent target and would provide outlier payment for only the high cost IPF cases. Accordingly, we are proposing the outlier policy to be 2 percent of the total proposed IPF payments. The amount of outlier payments would be funded by prospectively reducing the non-outlier payment rates in a budget-neutral manner. Under our proposed outlier policy, we would make outlier payments for discharges in which estimated costs exceed an adjusted threshold amount ($4,200 multiplied by the IPF's facility adjustments, that is wages, rural location, and teaching status) plus the total IPF prospective payment system adjusted payment amount for the discharge. The estimated cost for a case would be calculated by multiplying the overall facility-specific cost-to-charge ratio by the total charges for the inpatient stay. In establishing the loss-sharing ratio, we considered establishing a single ratio consistent with the hospital inpatient prospective payment system, which is set at a marginal cost of 80 percent of the difference between the cost for the discharge and the adjusted threshold amount. However, the proposed IPF prospective payment system unlike the hospital inpatient prospective payment system is a per diem payment system, we are concerned that a single loss-sharing ratio at 80 percent might provide an incentive to increase length of stay in order to receive additional outlier payments. Therefore, we are proposing to reduce the loss-sharing ratio when the length of the stay increases beyond the median length of stay. We believe that a reduction to the outlier loss-sharing ratio should occur in a similar manner to the declining per diem payment. The per diem payment amount under the proposed IPF prospective payment system is highest on days 1 through 4, declines on days 5 through 8, and declines further for all days beyond 8. Similarly, we are proposing to establish an 80-percent loss-sharing ratio for days 1 through 8 in order to reflect higher costs early in an IPF stay and reduce the ratio by 20 percent for days 9 and thereafter. This is consistent with the median length of stay for IPFs. Reducing the amount Medicare would share in the loss of high cost cases would provide an incentive for an IPF to contain costs once a case qualifies for outlier payments. We solicit comments on this approach. c. Proposed Implementation of the Outlier Policy The intent of proposing an outlier policy is to adequately pay for truly high-cost cases. However, we have become aware that under the hospital inpatient prospective payment system, some hospitals have taken advantage of two system features in the outlier policy to maximize their outlier payments. The first is the time lag between the current charges on a submitted claim and the cost-to-charge ratio taken from the most recent settled cost report. Second, statewide average cost-to-charge ratios are used in those instances in which an acute care hospital's operating or capital cost-to-charge ratios fall outside reasonable parameters. We set forth these parameters and the statewide cost-to-charge ratios for acute care hospitals in the annual publication of prospective payment rates that are published by August 1 of each year in accordance with § 412.8(b)(2). Currently, these parameters represent 3.0 standard deviations (plus or minus) from the geometric mean of cost-to-charge ratios for all hospitals. Hospitals could arbitrarily increase their charges so far above costs that their cost-to-charge ratios would fall below 3 standard deviations from the geometric mean of the cost-to-charge ratio. Thus, a higher statewide average cost-to-charge ratio would be applied to determine if the hospital should receive an outlier payment. This disparity results in their cost-to-charge ratios being set too high, which in turn results in an overestimation of their current costs per case. The intention of the outlier policy under both the hospital inpatient prospective payment system and the proposed IPF prospective payment system is to make payments only when the cost of care is extraordinarily high in relation to the average cost of treating comparable conditions or illnesses. We believe that if hospitals' charges are not sufficiently comparable in magnitude to their costs, the legislative purpose underlying payment for outliers is thwarted. Thus, on June 9, 2003, we published a final rule in the Federal Register (68 FR 34494) to ensure that outlier payments are paid for truly high-cost cases under the hospital inpatient prospective payment system. We believe the use of parameters is appropriate for determining cost-to-charge ratios to ensure these values are reasonable and that outlier payments can be made in the most equitable manner possible. Further, we believe the proposed methodology of computing IPF outlier payments is susceptible to the same payment enhancement practices identified under the hospital inpatient prospective payment system because it depends on the cost-to-charge ratio to determine the IPF's cost. Accordingly, as discussed below, we are proposing provisions for implementing the outlier policy to ensure the statistical accuracy of cost-to-charge ratios and appropriate adjustment of IPF outlier payments. 1. 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 propose to compute this 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 propose to 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 are proposing to 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 are proposing to 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 propose to 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 proposing 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 would 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 are proposing to 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, we are proposing that IPFs would 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. 2. Adjustment of IPF Outlier Payments As discussed in the hospital inpatient prospective payment system final rule for outliers, we have implemented changes to the 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 proposing the following: • Include in proposed § 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, we are proposing that fiscal intermediaries would use more recent data when determining an IPF's cost-to-charge ratio. Specifically, as provided in § 412.84(i), we are proposing that fiscal intermediaries would 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, we are proposing that 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. We are proposing that 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 are still assessing the procedural changes that would be necessary to implement this change. d. Computation of Proposed Outlier Payments In order to illustrate the proposed outlier payment mechanism, we present the following example of how we would calculate the outlier payment. Example: John Smith was hospitalized at a non-teaching IPF facility in Richmond, Virginia for 14 days. His total allowable billed charges for the 14 days was $20,000. The prospective payment amount (per diem payments plus adjustments) was $8,000. To determine whether this case qualifies for outlier payments, it would be necessary to compute the cost of the case by multiplying the facility's overall cost-to-charge ratio of .72 by the allowable charge of $20,000. In this case, the total allowable costs for Mr. Smith's case is $14,400 ($20,000 × .72). Because the IPF is a non-teaching urban facility, the fixed dollar threshold is adjusted by the wage index 0.9477. Table 4.—Computation Example of the Proposed Outlier Payment Steps to Calculate the Proposed Outlier Payment Calculate the Fixed Dollar Loss Threshold: Fixed Dollar Threshold $4,200 Wage adjusted labor share (.72828×4,200)*0.9477 $2,899 Non Labor Share (0.27172 × $4.200) 1,141 Adjusted Fixed Dollar Threshold ($2,899+$1,141) 4.040 Calculate Eligible Outlier Costs: Hospital Costs 14,400 Adjusted Fixed Dollar Threshold 4,040 Prospective Payment System Adjusted Payment 8,000 Eligible for Outlier Costs ($14,400−$4,040−$8,000) 2,360 Calculate the Loss Sharing Ratio Amount: Per Diem Outlier Costs ($2,360/14 days) 169 Loss-sharing Ratio Days 1 through 8 ($169×.80×8 days) 1,079 Loss-sharing Ratio Days 9 through 14 $169×.60 ×6 days) 607 The Total Outlier Payment Amount $1,079+$607) 1,686 e. Interrupted Stays Since per diem payments under the proposed IPF prospective payment system would be higher for the first 8 days of a stay (the variable per diem adjustment discussed earlier in this section), we are proposing to adopt an interrupted stay policy. The policy is intended to reduce incentives to move patients among Medicare-covered sites of care in order to maximize Medicare payment. We are concerned that IPFs could maximize payment by prematurely discharging patients after the 8 days during which they receive higher payments (the variable per diem adjustments), and then readmitting the same patient. In some cases a discharge and subsequent readmission within a short period of time may be appropriate. For example, we are concerned, in particular, that when there is a psychiatric unit within an acute care hospital, a patient could be transferred from the unit after only a few days of care to another part of the hospital and then be readmitted to the psychiatric unit. In this scenario, the hospital could receive the per diem adjustments for both stays in the psychiatric unit as well as receive the DRG payment associated with the acute hospital stay. In proposed § 412.402, we define an interrupted stay as one in which the patient is discharged from an IPF and returns to the same IPF within 5 consecutive calendar days. Specifically, we are proposing in § 412.424(d) that if a patient is discharged from an IPF and returns to the same IPF within 5 consecutive calendar days, we would treat both stays as a single stay. Therefore, we would not apply the variable per diem adjustment for the second admission and would combine the costs of both stays for the purpose of determining whether the case qualifies for outlier payments. We considered defining an interrupted stay as a readmission within 8 days of discharge since the variable per diem adjustments are not applied after the 8th day of the stay. We are not proposing this definition for an interrupted stay because we believe that after an 8-day absence from the IPF, many of the services that account for increased costs early in an inpatient psychiatric stay would need to be repeated, for example, assessments and laboratory testing. After a shorter absence from the IPF of 1 through 4 days, however, many of those admission-related services such as psychiatric evaluations and the patient's medical history would not need to be repeated. Therefore, we believe the lower end of the last range of payment adjustment, that is, 5 days, would provide for appropriate per diem payment adjustment as well as provide a disincentive to inappropriately shift patients between Medicare-covered sites of care. In addition, we intend to monitor the extent and timing of readmissions to IPFs and plan to account for changes in practice patterns as we refine the proposed IPF prospective payment system. Public comments are welcome on the proposed definition of an interrupted stay. For the purposes of counting the 5-calendar day time period to determine the length of the interrupted stay, the day of discharge would be counted as “day 1”, with midnight of that day serving as the end of that calendar day. The 4 calendar days that immediately follow day 1 would be days 2 through 5. C. Development of the Proposed Budget-Neutral Federal Per Diem Base Rate 1. Data Used To Develop the Proposed Federal Per Diem Base Rate Based on the regression analysis, we are proposing a prospective payment system for IPFs based on a per diem payment amount calculated from average costs adjusted for budget neutrality. The per diem amount would be adjusted by a budget-neutrality factor to arrive at the Federal per diem base rate used as the standard payment per day for the proposed IPF prospective payment system. The proposed Federal per diem base payment would be adjusted by the proposed wage index and the proposed patient-level and facility-level characteristics identified in the regression analysis. To calculate the proposed per diem amount, we would estimate the average cost per day for—(1) routine services from the most recent available cost report data (cost reports beginning in FY 1999 supplemented with 1998 cost reports if the 1999 cost report is missing); and (2) ancillary costs per day using data from the 1999 Medicare bills and corresponding data from facility cost reports. 2. Calculation of the Proposed Per Diem Amount For routine services, the proposed per diem operating and capital costs would be 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 developing the proposed per diem amount, we summed 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 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 for the groups of IPFs, 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 standard and generally used statistical trim points of plus or minus 3 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 would delete the data for that provider from the per diem rate development file. This method of trimming is consistent with the method used for the regression analysis. After trimming the data, the average routine cost per day would be $495. For the ancillary services, we would calculate the costs by converting charges from the 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 statistically valued 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). Because the distribution of ratios of cost-to-charges is not normally distributed and because there is no limit to the upper ceiling of the ratio, the mean value tends to 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 by applying the cost-to-charge ratios to the total ancillary charges for each patient stay, we would determine the average ancillary amount per day by dividing the total ancillary costs for all stays by the total covered Medicare days. Using this methodology, the average ancillary cost per day would be $67. Adding the average ancillary costs per day ($67) and the facility's average routine costs per day including capital costs ($495) provides the base payment amount ($562) for the estimated average per diem amount for each patient day of inpatient psychiatric care. 3. Determining the Update Factors for the Budget-Neutrality Calculation Section 124(a)(1) of Pub. L. 106-113 requires that the proposed IPF prospective payment system be budget neutral. In other words, the amount of total payments under the proposed IPF prospective payment system, including any payment adjustments, must be projected to be equal to the amount of total payments that would have been made if the proposed prospective payment system were not implemented. Therefore, we are proposing to calculate the budget-neutrality factor for the implementation period by setting the total estimated prospective payment system payments equal to the total estimated payments that would have been made under the TEFRA methodology had the proposed prospective payment system not been implemented. As discussed in section IV of this proposed rule, the implementation date of the proposed IPF prospective payment system is cost reporting periods beginning on or after April 1, 2004. In order to create a more even and efficient process of updates for the various Medicare payment systems, we are recommending that the first Federal base rate update occur on July 1, 2005. Therefore, we calculated the proposed Federal base rate to be budget neutral for the 15-month period April 1, 2004 through June 30, 2005. The data sources we used to calculate the budget-neutrality factor were the most complete data available for IPFs and included cost report data from FY 1999 and the 1999 Medicare claims data from the June 2001 update of the MedPAR files. We updated the cost report data for each IPF to the midpoint of that 15-month period (April 1, 2004 through June 30, 2005) and used t [Text truncated at 120,000 characters. The full text is on the page linked above.] --- Source: Frix Law Library, https://www.frixlaw.com/law-library/documents/fr%3A03-29137. Public record. Not legal advice.