INPATIENT REHABILITATION HOSPITALS in the United. Early Effects of the Prospective Payment System on Inpatient Rehabilitation Hospital Performance

198 ORIGINAL ARTICLE Early Effects of the Prospective Payment System on Inpatient Rehabilitation Hospital Performance Michael J. McCue, DBA, Jon M. Thompson, PhD ABSTRACT. McCue MJ, Thompson JM. Early effects of the prospective payment system on inpatient rehabilitation hospital performance. Arch Phys Med Rehabil 2006;87: 198-202. Objective: To assess changes in utilization and financial performance for inpatient rehabilitation facilities (IRFs) that shifted from Medicare s cost-based payment system to the IRF prospective payment system (PPS). Design: A pre-post nonequivalent comparison group design. The intervention group included IRFs that changed to the PPS in fiscal year 2002. The comparison group included IRFs that were paid under the cost-based system. Setting: IRFs in the United States. Participants: Final sample included 120 IRFs, with 26 IRFs in the comparison sample. Interventions: Not applicable. Main Outcome Measures: Outcome measures included utilization (length of stay [LOS], total discharges, Medicare discharges) and financial performance (revenue, expenses, profitability, Medicare payment and cost). Results: PPS IRFs experienced a smaller decline in LOS, whereas Medicare cost per discharge increased at a lower rate. PPS IRFs reduced operating costs per discharge, whereas profit margin had a greater increase. Conclusions: IRFs under PPS implemented cost controls that lead to lower operating costs below the fixed payment to profit under PPS. Discharge growth for PPS IRFs was similar to the comparison group. PPS facilities did not implement a strategy that attempted to admit more patients to increase Medicare payments. Key Words: Financial management; Inpatients; Prospective payment system; Rehabilitation; Rehabilitation centers. 2006 by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation INPATIENT REHABILITATION HOSPITALS in the United States continue to play a significant role in meeting the needs of people who are in need of intervention to optimize cognitive and physical functioning. 1 As life expectancy has progressively increased, more people are in need of rehabilitation services at some point in their life because of a congenital condition, From the Department of Health Administration, Virginia Commonwealth University, Medical College of Virginia Campus, Richmond, VA (McCue); and Health Services Administration Program, James Madison University, Harrisonburg, VA (Thompson). No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit upon the author(s) or upon any organization with which the author(s) is/are associated. Correspondence to Michael J. McCue, DBA, Dept of Health Administration, Virginia Commonwealth University, Medical College of Virginia Campus, PO Box 980203, Richmond, VA 23298-0203, e-mail: mccue@hsc.vcu.edu. Reprints are not available from the author. 0003-9993/06/8702-10118$32.00/0 doi:10.1016/j.apmr.2005.10.029 illness, or traumatic injury. In addition, significant expansion of the older age cohorts because of the aging baby boom generation is expected to result in increased demand for rehabilitation services. 2 Finally, to meet the demand for service, many hospital systems and for-profit providers have established rehabilitation services as part of their business strategy. 2 These factors have helped secure the position of inpatient rehabilitation facilities (IRFs) in the continuum of care in many communities. 2 Rehabilitation facilities have been viewed differently by the Medicare program because of the distinctive nature of services offered and the focus of care. Since 1983, rehabilitation facilities and specialty units within acute care hospitals have been excluded from Medicare s prospective payment system (PPS) and have been reimbursed under Tax Equity and Fiscal Responsibility Act (TEFRA) payment. Under TEFRA, IRFs were paid on the basis of their incurred average costs per discharge, subject to annual adjusted facility-specific limits. 3 The PPS exclusion granted to rehabilitation hospitals and other types of hospitals such as cancer and long-term hospitals was based on the inability of the diagnosis-related, group based PPS for acute care hospitals to account for resource consumption in rehabilitation hospitals. 4,5 IRFs focus on the restoration of cognitive and physical functioning, and, therefore, services are directed to improving functional limitations rather than solely treating an acute medical problem. 6 As a result of their exclusion from acute care hospital PPS and increasing consumer demand for rehabilitation services, significant growth has occurred in the number of IRFs. In 1984, there were 35 freestanding rehabilitation hospitals in the United States. 7 By 2003, there were 168 IRFs a 6-fold increase in 19 years. 8 Despite the exclusion from acute care hospital PPS, the Medicare program had always intended to establish a PPS for rehabilitation but recognized that a different payment method needed to be developed based on the resources required to meet the unique needs of rehabilitation patients. 9 Under TEFRA payment, Medicare experienced 17% average annual increases in expenditures throughout the 1990s because more Medicare beneficiaries were served by IRFs. 5 In fact, Medicare accounts for the largest portion of revenue in IRFs, 3 and approximately 70% of admissions to rehabilitation facilities are Medicare beneficiaries. 5 As a result of the continued increase in Medicare spending for rehabilitation hospital care, Congress passed the Balanced Budget Act in 1997, which provided for caps on TEFRA payments to rehabilitation hospitals and mandated the implementation of a PPS for rehabilitation hospitals based on functional measures associated with a patient s clinical condition. Amendments to the 1997 Balanced Budget Act enacted in 1999 called for the implementation of inpatient rehabilitation PPS, and, in 2001, final regulations were issued by the Centers for Medicaid and Medicare Services (CMS) for the implementation of the rehabilitation PPS beginning in 2002. Under the PPS, IRFs now are paid by Medicare under a prospectively based, per-case fixed payment system that creates an incentive to control costs and develop efficiencies in patient care. Under the current rehabilitation PPS, payment is made to IRFs according to a patient classification system. IRFs

EFFECTS OF PPS ON INPATIENT REHABILITATION PERFORMANCE, McCue 199 are paid predetermined, per-discharge rates for 1 of 385 intensive rehabilitation products, called case-mix groups (CMGs), defined by types of treatment episodes. 3 Patients are assigned to these categories based on the primary reason for their care (eg, stroke, burns), their age and levels of functional and cognitive impairments, and types of comorbidities present during their stay. Base rates for the CMGs reflect weighting to account for variation in resources used by different types of patients and are adjusted for local wage rates, rural hospital status, serving low income patients, and serving high-cost outliers. Despite the conversion to PPS in 2002, there is no knowledge of how facilities are performing. The focus of this study is to assess how the utilization and financial performance changed for IRFs under this new payment system. We examine IRFs performance both before and after PPS implementation and specifically address the following 2 questions: (1) How has PPS affected the utilization of IRFs in terms of discharges and length of stay (LOS)? and (2) How has PPS affected the financial performance of IRFs in terms of revenue, operating expenses, and profitability? As in the case of acute care hospitals, the prospective fixed payment creates an incentive to shorten LOS, reduce costs and increase discharges in order to profit from turning over their beds. 10 Generating a positive contribution margin to cover the high fixed overhead costs also entices hospitals to grow their admissions. 10 Therefore, we hypothesized the following: IRFs that convert to a PPS will have higher discharges, shorter LOS, greater Medicare discharges, and higher profit margin than IRFs under the TEFRA payment system. Such findings are helpful to both policymakers and rehabilitation hospital administrators. IRF administrators would be interested in findings that show any changes in operational practices such as managing LOS and operating expenses that are related to the new payment system and the associated incentives for greater efficiency. Policymakers would be interested in findings that show initial effects of the conversion to PPS on rehabilitation hospital utilization and financial performance. Findings that show the impact on performance would be helpful to show policymakers that IRFs responded to the PPS incentives for efficiency and effectiveness. METHODS Research Design To evaluate the change to the new payment system on IRF operational and financial performance measures, the study followed a pre-post nonequivalent comparison group design analysis. 11 The intervention group included IRFs that changed their payment system from the TEFRA-based payment system to PPS in fiscal year (FY) 2002. The comparison group included IRFs that were still reimbursed under the old payment system of TEFRA for FY 2002. To determine if the payment system had an effect on utilization and performance, we computed the difference between the pre-pps and post-pps measures for both groups. In addition, to determine if the 2 groups were equivalent in their measures at the FY before the change in reimbursement, we computed the mean differences between the 2 groups during the FY before the change. We used a t test to compare the differences in the percentage changes in the pre- and postperiods between PPS and TEFRA comparison groups. A t test analysis was also computed to compare the differences between the 2 groups for the FY before the change in payment system. Given the small sample size for TEFRA comparison group and high standard deviation (SD) values for some of the measures, we also used a nonparametric median test to test significance of these findings. The median test findings also confirmed our t test findings. Data Sources IRFs with cost-reporting periods beginning on or after January 1, 2002, and before October 1, 2002, were identified as intervention facilities that changed their payment system from TEFRA to PPS in FY 2002. The comparison group included IRFs that had cost-reporting periods after October 1, 2002, and did not shift to PPS until FY 2003 and were paid under a TEFRA basis. Therefore, the comparison group included IRFs that were still reimbursed under the old payment system of TEFRA for FY 2002. Financial and operational data for IRFs were collected from Health Care Cost Report Information System data for the fiscal years prior to and after the change. The total number of freestanding IRFs within the CMS Health Care Cost Report Information System database was 168. Missing data and cost reporting fiscal data less than 12 months resulted in a final sample of 120 IRFs that changed from TEFRA to PPS. The comparison group IRFs final sample totaled 26. For each IRF, we collected annual cost report data for 2 fiscal time periods before the change to PPS and 1 annual fiscal period after the change to PPS. Because we were unable to collect cost report data from inpatient rehabilitation units within acute care hospitals, our analysis only focused on the financial performance of IRFs. For IRFs to be paid by Medicare under this PPS, they needed to first meet the criteria of an IRF, which includes serving at least 75% of an inpatient population requiring extensive rehabilitation services for 1 of 10 medical conditions. The transition to IRF PPS was phased in over time. PPS IRFs with costreporting periods beginning on or after January 1, 2002, and before October 1, 2002 (FY 2002), could have elected for either a blended payment rate combining 66.66% of the federal IRF PPS rate and the remaining 33.33% payment rate representing the payment rate if the IRF was not under a PPS or 100% of the federal IRF PPS rate. The study was unable to distinguish PPS IRFs from either the blended or full PPS rate. For IRFs with cost-reporting periods beginning on or after October 1, 2002, the only option was to be paid 100% of the federal IRF PPS rate in FY 2003. In terms of IRF characteristics that initially selected PPS compared with those IRFs that remained in TEFRA, 52% of the TEFRA IRFs were for-profit compared with 81% for PPS IRFs. In terms of bed size, PPS IRFs had an average bed size of 70 compared with 82 for the TEFRA IRFs. Therefore, it appears initially that smaller for-profit IRFs had a greater propensity to select PPS rate rather than TEFRA rate. Variables Table 1 lists and defines the variables and presents the means and SDs for each measure by PPS versus TEFRA groups before the change in payment. Outcome or dependent variables included utilization and financial performance measures. Utilization measures included LOS, Medicare discharges, and total discharges. Financial measures included net patient revenue, operating expenses, Medicare payment, and Medicare costs. To measure profitability, we computed the profitability ratios on a total and operational basis. Because these were ratio values, we analyzed the absolute change in total profit margin ratio, which is measured as total income earned by the IRF, and absolute change in operating margin ratio measured as income earned from patient operations.

200 EFFECTS OF PPS ON INPATIENT REHABILITATION PERFORMANCE, McCue Table 1: Variables and Operational Definitions Measured in the Time Periods Before PPS Change Variables Operational Definitions PPS Group TEFRA Comparison Group Total discharges Total discharges 1106.00 1044.00 1206.00 649.00 Medicare discharges Total Medicare discharges 770.00 455.00 803.00 604.00 LOS (d) Inpatient days/discharges 16.73 2.97 16.24 3.07 Medicare payment per discharge ($) Medicare payment/total discharges 10,316 2178 9971 2219 Medicare cost per discharge ($) Medicare cost/total discharge 9069 2705 8941 2071 Net patient revenue per discharge ($) Net patient revenue/total discharge 15,399 6333 16,374 7211 Operating expense per discharge ($) Operating expense/total discharge 14,941 9644 16,928 8716 Operating profit margin (%) Operating income/net patient revenue 4.19 16.4 1.15 12.1* Total profit margin (%) Net income/(net patient revenue other operating revenue) 5.98 15.1 2.77 9.8 NOTE. Values are mean standard deviation (SD). *Significant at the.01 level. RESULTS The results were first analyzed by comparing the mean values of the utilization and financial measures of the 2 groups (PPS vs TEFRA) in the 2 years before they changed to PPS (see table 1). The second analysis (table 2) presents the changes in the utilization and financial measures between those IRFs that switched from TEFRA reimbursement to PPS compared with IRFs that remained under the TEFRA payment system. Before PPS Results Table 1 shows the mean differences between the 2 groups for the period before the change in the payment system. There were no statistically mean differences in total discharges and Medicare discharges before the change in the payment system. This finding suggests for the fiscal period before the change to the PPS, IRFs that changed to the PPS had similar LOS and discharges to IRFs that did not change their payment system. From a level analysis, the 2 groups were similar before the change. For financial performance measures (see table 1), the operating margin ratio was significantly higher for those IRFs that Table 2: Impact of PPS on Selected Variables Variables PPS Group TEFRA Comparison Group Percentage change in total discharges (%) 19.39 18.3 18.62 20.3 Percentage change in Medicare discharges (%) 22.80 23.9 26.00 32.4 Absolute change in LOS (d) 1.35 1.66 2.33 1.75* Percentage change in Medicare payment per discharge (%) 30.80 20.8 23.50 41.1 Medicare cost per discharge (%) 2.6 15.44 12.83 15.10* Net patient revenue per discharge (%) 8.85 17.2 7.79 11.02 Operating expense per discharge (%) 5.80 10.6 0.4 10.23* Absolute change in Operating margin percentage points 12.34 10.45 6.47 10.21* Total profit margin percentage points 12.64 10.44 4.68 11.21* NOTE. Values are mean SD. *Significant at the.01 level. changed to PPS compared with those IRFs that did not change. Total profit margin, net patient revenue per discharge, and operating expense per discharge did not differ significantly before the change to PPS. Change to PPS Results Table 2 lists the results from the differences in pre- and postchange values between the 2 groups. Significant results occurred for the changes in LOS, Medicare cost per discharge, operating costs per discharge, operating profit margin, and total profit. In terms of the change in LOS, IRFs that changed to PPS reduced their LOS by 1.35 days, whereas IRFs that did not change their payment system reduced the LOS by 2.33 days. For the financial measures, percentage change in Medicare cost grew at a lower rate, 2.6%, for the PPS group compared with 12.83% for the TEFRA comparison group, whereas there was no significant difference in the percentage change in Medicare payment. Similar findings occurred for change in total net patient revenue and operating expenses. The percentage increase in cost differed significantly between the 2 groups. For the PPS group, operating expenses per discharge declined by 6% compared with.4% for the TEFRA comparison group. For the percentage increase in net patient revenue, there was no significant difference between the 2 groups. In terms of profitability, absolute change in operating margin ratio was statistically significant and resulted in a 12.34 percentage point increase for the PPS group compared with only a 6.47 percentage point increase for the comparison group. A similar outcome occurred for the change in total profit margin ratio. DISCUSSION PPS for IRFs was implemented in 2002, and Medicare now pays for care prospectively on a per-case basis. The intent of the PPS program is to create greater efficiency in care by paying for care commensurate with the resource needs of specific groups of patients. The purpose of this study was to assess the change in the utilization and financial performance measures of IRFs under PPS compared with the TEFRA payment system. In the period before the onset of the new payment system, we found no statistically significant differences for all the level measures except for operating profit margin ratio between IRFs that were paid under the PPS compared with the TEFRA. Thus, differences before the new payment system should not have influenced the changes in utilization and financial performance measures, except in the case of the operating margin ratio.

EFFECTS OF PPS ON INPATIENT REHABILITATION PERFORMANCE, McCue 201 In evaluating the IRFs that changed from cost based to PPS, our findings show that IRFs did not make any significant changes in utilization, specifically total discharges and Medicare discharges. Significant changes did occur with respect to LOS. IRFs under the new PPS reduced their LOS by 1 day compared with 2 days for IRFs under the TEFRA system. This outcome may be a reflection of the change in patient mix. Under TEFRA, as Carter and Paddock 12 noted, IRFs had an incentive to discharge early costly patients because the TEFRA payment applied to all patients regardless of their medical condition. Conversely, under PPS, IRFs are now compensated for treating high-cost, complex patients. 12 Therefore, this finding may imply that IRFs under the TEFRA payment were reducing their LOS because they were not being compensated for treating complex cases requiring a longer LOS. This finding is consistent with the study findings of Ottenbacher et al 13 in their analysis of rehabilitation facilities in the period 1994 to 2001 and shows a continued decline in LOS for rehabilitation hospital patients. However, we found that IRFs under the new payment system had lower growth in Medicare costs per discharge compared with IRFs under TEFRA, whereas total operating expense per discharge declined for the PPS IRFs. The cost-based system for the TEFRA IRFs did not create the pressure to reduce costs. Given that there was no difference in revenue per discharge or Medicare payment per discharge, PPS IRFs earned higher increases in operating and total profit margins by lowering their operating expenses. This finding is consistent with results of prior studies that have examined acute care hospital response to PPS 14 as well as skilled nursing facility (SNF) response to PPS. 15,16 Furthermore, this finding is also consistent with expectations as to how IRF administrators will adapt to the new IRF PPS system 17 and suggests that IRFs have emulated acute care hospitals and SNFs by implementing similar strategies to reduce costs in order to maximize profit. From a policy standpoint, the findings of this study suggest that IRFs under PPS implemented cost controls that allowed them to lower their operating costs below the fixed payment in order to profit under the new payment system. In the case of acute care hospitals that are paid under a fixed PPS, higher profits are generated when facilities increase census, reduce LOS, and lower costs. For rehabilitation facilities, we found that discharge growth for the IRFs under PPS was similar to the comparison group. Therefore, these facilities did not implement an operational strategy that attempted to admit more patients to increase their Medicare payments. This finding is consistent with the study results of Thompson and McCue 18 who found that higher profit facilities under TEFRA had lower dependency on Medicare patients and revenues. In this current study, capacity constraint may be a reason why IRFs were unable to increase patient volume. During 2002, IRFs were operating close to capacity with an average occupancy rate of 75%. Instead, it appears that IRFs focused on reducing operating costs either through reducing staffing mix and wages, limiting the type of care for complex cases, and/or using fewer supplies and services. In addition, high-cost IRF patients may have been transferred to other providers, such as long-term care hospitals, SNFs, or home health agencies, whereby level and units of payment may result in higher compensation. As a result, reimbursement rather than the medical condition would be the force behind the type of provider used to treat the patient. Finally, the findings of this study show that for profits (81%) were more likely to select initially the PPS than the TEFRA system (54%). Under the fixed payment system of PPS, an IRF can now be accurately compensated for treating high-cost, complex cases than under the TEFRA payment system. This outcome suggests that the for-profit IRFs viewed this new payment system as a greater opportunity to maximize their profits for a given diagnosis by controlling their costs below the payment. However, excessive profits by IRFs, in the future, may cause Congress and the CMS to lower Medicare payments to IRFs. One approach to reduce payments is by lowering the relative weights for each case-mix category. Another approach is not to provide an increase in the market basket index factor that adjusts for higher prices in the mix of goods and services to the base payment rate. Study Limitations There are several limitations to our study that should be noted. First, this is a descriptive study that compares mean IRF utilization and performance between 2 groups: IRFs that converted to PPS and those that did not. The study does not address those factors that might be associated with observed changes in outcome variables. This could be addressed through a multivariate study that identifies the relative contribution of predictor variables on variation in the dependent variable of choice, such as LOS. Rather, the current study addresses changes in the performance variables in 2 groups between 2 time periods. A multivariate study is the subject for additional, future research. Second, we were unable to control for case mix, which may affect the financial and operating performance of IRFs under the TEFRA system. Some facilities may serve, overall, less complex patients and patients who have less comorbidity. A casemix measure was not in the database, and therefore we could not include it. This is a limitation of rehabilitation databases that has been identified in prior research. 18,19 Third, we also acknowledge that we were unable to distinguish between those facilities that selected either the blended or 100% PPS rate. This may have affected how these IRFs reacted to this change and influenced the outcome of their utilization and financial measures. In addition, we did not address the operational and financial performance of inpatient rehabilitation units within acute care hospitals. Although these units are an important source of rehabilitation care, our study focus was on rehabilitation hospitals because these facilities face distinct organization-wide pressures on performance. CONCLUSIONS The results reflect IRF performance in the first year of the PPS for IRFs. Observed changes in performance may or may not be consistent over time. Accordingly, measurement of IRF performance over a much longer time frame is needed through a longitudinal study. Such a study will assist in determining if observed changes in rehabilitation hospital performance reported here are sustained over time. References 1. Esselman P. Inpatient rehabilitation outcome trends: implications for the future. JAMA 2004;292:1746-8. 2. Wheatley B, DeJong G, Sutton J. Consolidation of the inpatient medical rehabilitation industry. Health Aff (Millwood) 1998;17: 209-15. 3. Medicare Payment Advisory Commission. Rehabilitation facilities (inpatient) payment system. Briefing. July 13, 2004. Available at: http://www.medpac.gov/publications/other_reports/071304_ PPS13_Rehab.pdf. Accessed November 3, 2005. 4. Grimaldi PL. Inpatient rehabilitation facilities are now paid under prospective rates. J Health Care Finance 2002;28:32-8.

202 EFFECTS OF PPS ON INPATIENT REHABILITATION PERFORMANCE, McCue 5. Medicare Payment Advisory Commission. Report to Congress: Medicare payment policy. Mar 2000. Available at: http://www. medpac.gov/publications/congressional_reports/mar00%20table %20of%20Contents.pdf. Accessed November 7, 2005. 6. Carter GM, Relles DA, Ridgeway GK, Rimes CM. Measuring function for Medicare inpatient rehabilitation payment. Health Care Financ Rev 2003;24:25-44. 7. Langenbrunner J, Willis P, Jencks S, Dobson A, Iezzoni L. Developing payment refinements and reforms under Medicare for excluded hospitals. Health Care Financ Rev 1989;10:91-107. 8. Centers for Medicaid and Medicare Services. 2004 CMS statistics. Washington (DC): U.S. Department of Health and Human Services; Oct 2004. CMS Publication No. 03455. 9. Health Care Financing Administration. Report to Congress: developing a prospective payment system for excluded hospitals. Washington (DC): Dept of Health and Human Services; 1987. 10. Hadley J, Zuckerman S, Feder J. Profits and fiscal pressure in the prospective payment system: their impacts on hospitals. Inquiry 1989;26:354-65. 11. Cook DC, Campbell DT. Quasi-experimentation design & analysis issues for field setting. Boston: Houghton Mifflin; 1979. 12. Carter GM, Paddock SM. Preliminary analyses of changes in coding and case mix under inpatient rehabilitation facility prospective payment system. Santa Monica: Rand; 2002. 13. Ottenbacher KJ, Smith PM, Illig SB, Linn RT, Sotir GV, Granger CV. Trends in length of stay, living setting, functional outcome, and mortality following medical rehabilitation. JAMA 2004;292: 1687-95. 14. Feder J, Hadley J, Zuckerman S. How did Medicare s prospective payment system affect hospitals? N Engl J Med 1987;317: 867-73. 15. White C. Rehabilitation therapy in skilled nursing facilities: effects of Medicare s new prospective payment system. Health Aff (Millwood) 2003;22:214-23. 16. Zinn JS, Mor V, Intrator O, Feng Z, Angellini J, Davis JA. The impact of the prospective payment system for skilled nursing facilities on therapy service provision: a transaction cost approach. Health Serv Res 2003;38:1467-85. 17. Ward GA. Preserving high-quality care. Greg A. Ward, Rehabilitation Institute of Chicago. Healthc Financ Manage 2002;56: 36-9. 18. Thompson JM, McCue MJ. Organizational and market factors associated with Medicare dependence in inpatient rehabilitation hospitals. Health Serv Manage Res 2004;17:13-23. 19. Chan L, Ciol M. Medicare s payment system: its effect on discharges to skilled nursing facilities from rehabilitation hospitals. Arch Phys Med Rehabil 2000;81:715-9.