| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||
Correspondence: Address correspondence to Debra Street, Pepper Institute on Aging and Public Policy, Florida State University, Tallahassee, FL 32306-1121. E-mail: dstreet{at}garnet.acns.fsu.edu
 |
Abstract |
|---|
 TOP Abstract Federal and State Health...MethodsConclusionsReferences |
|---|
Key Words: Medicaid long-term care • Medicare long-term care • Skilled nursing facilities
The long-term care system in the United States consists of a complex mix of public and private financing and delivery mechanisms that, until recently, have favored nursing homes as the site of long-term care services, with Medicaid as the primary payer. Federal and state policies enacted over the past two decades have begun to transform the financing and thus delivery of nursing home care across the states in both intended and unintended ways. Florida is an excellent state for evaluating how these policies affect institutional long-term care for older people. In 1996, Florida was the only state with more than 19% of its population 65 and older, compared with a national average of 12.8% (U.S. Bureau of the Census, 1996). The large number of elderly residents has a significant effect both on state expenditures for Medicaid and on federal Medicare expenditures. In 1996, Florida accounted for 6.9% of all Medicare skilled nursing facility (SNF) admissions nationally and was one of three states (along with California and Texas) in which nearly one fourth of all Medicare-covered SNF stays occurred (Health Care Financing Administration [HCFA], 1998). Despite its large elderly population, in 2000, Florida ranked only 44th in the nation in total Medicaid long-term care expenditures (Burwell, 2001).
In this study, we first describe changes in state and federal policies that created a growing disparity in the median SNF revenues from various payers, including Medicaid, Medicare, Insurance/health maintenance organizations (HMOs) and private pay between 1989 and 1997. We then use a series of ordinary least squares (OLS) regression models to examine the effects of the reimbursement gap between Medicaid and other payers on the percentage of Medicaid nursing home admissions. Data analyzed in the regression analysis come from three sources: state Certificate of Need (CON) and Nursing Home Resident Data (NHRD), and data from the federal Online Survey Certification and Reporting (OSCAR) database. Finally, we examine the effect of changes in reimbursement streams on aggregate resident characteristics in terms of age, physical status, length of stay, and place of discharge.
|
Federal and State Health Care Policy |
|---|
|
TOPAbstractFederal and State Health...MethodsConclusionsReferences |
|---|
Medicare Cost Containment
During the 1980s, policy makers sought ways to contain soaring Medicare costs for hospital care. Until 1983, Medicare hospitals were reimbursed on a cost-plus basis, a method that encouraged long hospital stays. The first important policy shift affecting Medicare's long-term care role was the introduction of the Prospective Payment System (PPS) for Medicare hospital benefits, created under the Social Security Amendments of 1983. Under Medicare, PPS hospitals were reimbursed a fixed amount per patient, according to diagnosis-related groups (Moon, 1993). Phased in gradually and fully in place by 1987, PPS gave hospitals an incentive to shorten lengths of stay and to discharge patients to posthospital settings as soon as medically feasible (Petrie, 1992). The effect was to increase rates of nursing home admissions, as hospitals discharged patients to SNFs to recover from surgery, stroke, or other health problems (Gaumer & Stavins, 1992). As a result, the average length of a Medicare hospital stay declined, and SNFs became an increasingly important site of posthospital subacute care (Cohen & Dubay, 1990; Lewin-VHI, Inc., 1995).
The trend toward conventional (i.e., non-HMO) Medicare-financed care being provided in SNFs rather than hospitals accelerated sharply after the Health Care Financing Administration (HCFA) issued a 1988 administrative directive (HCFA transmittal no. 222) that clarified and expanded the definition of what constituted skilled care under the Medicare SNF benefit (Laliberte, Mor, & Berg, 1997). For the first time, the administration of intravenous feedings and intramuscular injections, the insertion of catheters, and the provision of ultrasound, shortwave, and microwave therapies were explicitly identified as skilled services under the Medicare SNF benefit. The new guidelines also expanded services to persons whose health and functional status were not necessarily going to improve (Bishop & Dubay, 1991). Previously, Medicare guidelines encouraged restrictive interpretation of SNF claims by Medicare intermediaries (Liu, Taghavi, & Cornelius, 1992). After the guideline changes, Medicare fiscal intermediaries rejected fewer claims for Medicare SNF reimbursement than they had previously, facilitating the transfer of Medicare patients from hospitals to nursing homes (Bishop & Dubay, 1991). As a result, the percentage of nursing home cost-based expenditures paid by Medicare increased by 7.5% between 1988 and 1989. This increase was caused both by increased Medicare admissions from hospitals and by reclassification of current residents (Aaronson, Zinn, & Rosko, 1994a). Together, Medicare hospital PPS implementation and eased regulations for conventional Medicare SNF benefits expanded payment options for posthospitalization nursing home stays and increased the percentage of SNF residents admitted directly from hospitals.
The Expansion of Managed Care Organizations
A second factor that altered conditions for SNFs was the growth of managed care. By the mid-1990s, managed care organizations were dominant providers in the private health insurance market, representing approximately 70% of private sector insurance (General Accounting Office, 1996). Although managed care in the public sector was not as extensive, in the 1980s and 1990s both Medicare and state Medicaid programs increased contracting with managed care organizations.
Florida was among the few states with significant concentrations of HMO Medicare beneficiaries during the 1990s. Medicare managed care for Floridians was dominated by HMOs, which negotiated risk-contract arrangements with the HCFA (now Center for Medicare and Medicaid Services). Under risk contracts, Medicare HMOs received a fixed payment for each beneficiary enrolled. Although Medicare HMOs assumed a level of risk for the cost of providing care at least as comprehensive as conventional Medicare, they also had greater flexibility in determining the conditions for making SNF payments. By the mid-1990s, Florida ranked fourth among the states in the proportion of its Medicare population enrolled in risk-contract HMOs. In 1995, 15.6% of all Florida Medicare beneficiaries were enrolled in Medicare prepaid health plans, compared with just 7% nationwide (General Accounting Office, 1996). As is true of managed care generally, capitation (i.e., set payment per beneficiary) gave Medicare HMOs a strong incentive to provide health care in the least expensive medically appropriate setting. This made the provision of subacute care for Medicare HMO beneficiaries in SNFs rather than hospitals an attractive option for Medicare HMOs.
Regulations Regarding Payment Determinations
The third factor that altered the environment in which Florida SNFs were operating was the growing payment disparity between Medicaid and Medicare reimbursement levels. Conventional Medicare payments to SNFs nationwide were determined by a cost-based reimbursement method until 1997. As a result, between 1989 and 1997, Medicare's average payment per nursing home day increased 12% annually on average, even though the SNF market basket index, which measures yearly changes in the prices of goods and services purchased by nursing homes, rose only 3% on average (General Accounting Office, 2000).
During the same period, Florida sought to constrain its Medicaid nursing home costs through a prospective payment system, which is similar to the federal government's PPS strategy for Medicare hospital cost constraint. Under prospective payment, the rate for Medicaid nursing home reimbursement is based on property, operating (based on a Florida market basket index) and resident care costs, and a return on equity or use allowance (Agency for Health Care Administration, 2000a). Rates are set semiannually (prior to January 1 and June 1), and are adjusted for facility size and geographic location in the state. A target rate system was added in 1988, which limited operating and patient care per diem reimbursement growth to a function of an inflation index (Clarke, 1992). During the early 1990s, the Florida Medicaid program also absorbed more than $1 billion in budget cuts, with a substantial share occurring in institutional provider payments (Wiener & Stevenson, 1998). The combined effect of these measures was to slow the growth of Medicaid reimbursements to Florida SNFs.
Because Medicare's cost-based formula provided more generous reimbursement for SNF care than Medicaid's prospective formula, the disparity in revenue per bed day between the two payment sources grew in both absolute and relative terms between 1989 and 1997. Figure 1 shows the emerging gap among payer sources for Florida SNFs, derived from CON financial data (described in more detail later). Between 1989 and 1997, median Medicaid payments per bed day increased approximately 55%, from $60 to $92, whereas median Medicare payments per bed day grew more than 250% from $104 to $374. During the same period, median Insurance/HMO total payments per bed day more than doubled, from $104 to $242. (The Insurance/HMO category includes SNF revenues from private insurers and HMOs [both for Medicare and non-Medicare HMOs].) Although Insurance/HMO payments were lower than Medicare payments per bed day, they were still substantially higher than Medicaid.
|
Much of the increased Medicare SNF spending during the period of the study was driven by the cost-based reimbursement for ancillary services for traditional Medicare beneficiaries (non-HMO beneficiaries; HCFA, 1998). As shown in Figure 2, from 1989 to 1997, median ancillary services revenue per bed day trended nearly flat for Medicaid and private pay, whereas Medicare and Insurance/HMO ancillary services revenue per bed day rose substantially over the period, creating a widening gap between payers. Differences in median routine services revenues by payment source were far less dramatic. As Figure 3 illustrates, in 1989, median routine service revenues clustered relatively closely, with revenues from routine services per bed day ranging from $59 for Medicaid, to $66 for Medicare, to $73 for Insurance/HMO. By 1997, the gap between median routine services revenue per Medicaid bed day and all other payer sources had increased, but not nearly as much as ancillary services.
|
|
Hypotheses
We hypothesize that the growing gap between Medicaid and other payers would lead Florida SNFs to admit fewer residents whose care is reimbursed through Medicaid. We expect this relationship to exist even after controlling for a variety of facility characteristics known to affect the payer mix of SNFs, such as profit status, chain ownership, facility size, and urban/rural status.
|
Methods |
|---|
|
TOPAbstractFederal and State Health...MethodsConclusionsReferences |
|---|
|
NHRD data were linked to CON facility-level bed and financial data, gathered annually by the Florida Agency for Health Care Administration. The most recent year for which financial data are available is 1997. The Medicare revenue gaps are the difference between Medicare and Medicaid revenue per bed day; the Insurance/HMO reimbursement gaps are the difference between Insurance/HMO and Medicaid revenue per bed day. Because cost-based reimbursement for ancillary services has been identified as driving much of the increased Medicare SNF spending during the period of the study (HCFA, 1998), we decompose the revenue variables into measures of ancillary services revenue (Ancillary Medicare Gap or Ancillary Insurance/HMO gap) and routine services revenue (Routine Medicare Gap or Routine Insurance/HMO gap). We lag these variables because we expect that the effects of these gaps will be realized in the year after they occur. In other words, we use measures of the reimbursement gaps in the preceding year to predict the current level of Medicaid admissions.
The facility characteristics included in the model are profit status (coded 0 = for-profit, 1 = nonprofit and government), chain-operated (coded 0 = independent, 1 = chain), size, and rural or urban location. After convention, we created four dummy variables for facility size (Harrington et al., 2001): fewer than 60 beds, 60–119 beds, and more than 160 beds, with facilities from 120 to 160 beds as the excluded category. A rural/urban variable (coded 0 = rural, 1 = urban) is also included in the model.
To analyze resident physical status (an indirect measure of facility case-mix), we also used data from the OSCAR database. OSCAR gathers data on residents in each licensed SNF certified by Medicare and/or Medicaid throughout the United States at least once every 9–15 months. Yearly observation files were created for Florida facilities spanning from 1989 to 1997 and merged with the data provided by the state. Because OSCAR identifiers for some facilities changed over time (e.g., when facilities change ownership), cases were matched based on facility name, address, and nursing home identification number available in both state and OSCAR data. The physical status of residents was captured by activities of daily living (ADLs) and mobility measures in the OSCAR data. Three ADL measures—requiring assistance to eat, transfer, and use the toilet—are well-established as characteristics related to resource utilization in nursing homes (Harrington & Carillo, 1999; Mukamel, 1997). Before 1995, OSCAR ADL measures were counts of residents who could perform tasks (1) independently, (2) with some supervision, (3) with limited assistance, (4) with extensive assistance, or (5) were totally dependent. From 1995, the OSCAR measures counted residents who could perform the same tasks (1) independently, (2) with the assistance of one or two staff, or (3) were dependent. Because of the measurement change, the only consistent measure of ADLs over time was "residents' independence." Consequently, ADL variables were coded to measure the percentage of residents in each facility that could eat, use the toilet, and transfer independently. OSCAR measures of the percentage of chair-bound, bedfast, and independently ambulatory facility residents were also used to assess the relationship between admission source and resident physical status.
In the following sections, we first report bivariate findings over time to show trends affecting SNF revenue sources. We then use regression models to test the effects of growing disparities between Medicaid and other payers on the percentage of residents admitted to SNFs with Medicaid as primary payer at admission. Finally, we run bivariate correlations to assess the effect of the revenue trends on SNF resident characteristics.
Change in Reimbursement Streams
During the 1990s, patterns of primary payer at admission changed significantly for Florida SNFs. As Figure 4 shows, residents admitted with Medicare as primary payer increased from 39% in 1989 to nearly 60% of annual SNF admissions by 1997, whereas the percentage of residents with Insurance/HMO as primary payer increased from 2% to 14%. Because state data collapse Medicare HMO, non-Medicare HMO and private insurance into a single category, we cannot specify what portion of the Insurance/HMO admissions were Medicare beneficiaries. However, the Insurance/HMO trend identified is consistent with the growth and magnitude of Florida Medicare HMOs and with the tendency of SNFs to serve predominantly aged residents. By contrast, annual SNF admissions with Medicaid as primary payer declined from 26% in 1989 to just 16% in the later years. Residents who paid privately at admission declined from nearly 30% of all SNF admissions in 1989 to only 10% by 1998.
|
|
|
|
With the inclusion of percentage admitted from hospitals and facility controls, the weak effect of the ancillary gap on Medicaid admissions reverses direction. To reconcile this unexpected finding, we explored the possibility of interaction effects with facility profit status. Prior research suggests why facilities might respond differently to the Medicare–Medicaid reimbursement gap, depending on their profit status. Profit status may reflect differing management objectives and styles of care and has been identified as a predictor of the quality of SNF care (Aaronson, Zinn, & Rosko, 1994b), of reported deficiencies and staffing levels (Harrington, Carrillo, Thollaug, Summers, & Wellin, 2000) and of the probability of admitting Medicare SNF residents (Bishop & Dubay, 1991). The significant interaction coefficient in Model 5 between the Medicare–Medicaid gap in ancillary revenue and profit status suggests that for-profit and nonprofit facilities vary in their response to the gap. As the ancillary revenue gap increases, nonprofit and government facilities admit a larger share of Medicaid residents, whereas for-profit facilities admit relatively fewer Medicaid-paying residents. By contrast, the interaction coefficient in Model 6 shows that there is no statistically significant difference by profit status in the response to the magnitude of the Medicare–Medicaid routine revenue gap.
The interaction effects in Models 5 and 6 help explain why the Medicare–Medicaid gap in routine service revenue displays a consistent negative relationship in the previous models and why the Medicare–Medicaid gap in ancillary service revenue displays inconsistent relationships under differing model specifications. For-profit nursing facilities reduce their percentage of admissions with Medicaid as primary payer in response to the increasing magnitude of the Medicare–Medicaid gap in ancillary service revenue; nonprofit nursing homes take up the slack by admitting a higher percentage of residents with Medicaid as primary payer at admission.
In the regressions of the Insurance/HMO gaps on the percentage of Medicaid admissions, shown in Table 3, only SNFs reporting any Insurance/HMO revenue have been included (thus the low Ns for the models). Among these facilities, the results suggest a relatively weak relationship between the Insurance/HMO gaps and percentage of admissions with Medicaid as primary payer. At the zero-order level, the Insurance–Medicaid gap in ancillary services is negative and statistically significant, suggesting that an increase in the gap between ancillary service reimbursement for Insurance/HMO and Medicaid is associated with a decrease in the percentage of Medicaid admissions in the following year. However, the relationship is not statistically significant once additional variables are included in the model. Likewise, there is no statistically significant relationship between the Insurance/HMO–Medicaid gap in routine revenue and Medicaid admissions in subsequent years. The percentage of admissions from hospitals is negatively related to the percentage of Medicaid admissions, such that an increase in admissions from hospitals is associated with a decrease in the percentage of Medicaid admissions. The facility characteristics presented in Table 3 display effects similar to those presented in Table 2, with a few exceptions. Overall, nonprofit and government facilities admit fewer Medicaid-paying residents as a proportion of total admissions than for-profit facilities. Also, urban location is no longer associated with the percentage of admissions from hospitals. This is likely because we are selecting only SNFs that report nonzero Insurance/HMO revenue, which are predominantly located in urban areas of the state.
When examining the interaction effects in Models 5 and 6, for-profit and nonprofit facilities differ in their response to routine service revenue, but not in response to ancillary service revenue. As the Insurance/HMO–Medicaid gap in routine revenue increases, nonprofit and government facilities admit a higher percentage of Medicaid-paying residents. This helps explain the inconsistency in the routine revenue coefficient in the previous models.
As noted previously, because conventional Medicare SNF admissions must be preceded by a three-day hospital stay, we expect there to be a positive relationship between admission source and Medicare as the primary payer at admission. In Table 4, we calculate Pearson's r values (correlation coefficients) to measure the strength of the linear relationships between admission source and primary payer at admission at the facility level. As expected, the percentage of SNF admissions from hospitals is positively correlated with Medicare as the primary payer at admission. It is also positively correlated with Insurance/HMO as the primary payer at admission, although this correlation is weaker, in part because Medicare HMOs can permit SNF admissions without a hospital stay. Also, as expected, the percentage of SNF admissions from home, assisted living facilities, and different nursing homes is negatively correlated with Medicare or Insurance/HMO as the primary payer at admission.
|
Resident Characteristics
Age
We expected higher percentages of SNF admissions from hospitals to be positively correlated with increased admission of younger residents, because many posthospital SNF admissions are younger Medicare beneficiaries receiving subacute care rather than the traditionally older, very frail, custodial admissions. As Table 4 shows, there is a significant positive correlation between "young" nursing home admissions (i.e., residents aged 84 or younger) and admissions from hospitals, whereas admissions from assisted living facilities, home, and other nursing homes are positively correlated with "older" (85 and older) admissions.
Physical Status
Residents admitted from hospitals for subacute SNF care would likely have more complicated health conditions and require more assistance than traditional nursing home residents. Consequently, admissions from hospitals should be positively correlated with higher resident acuity levels. Acuity is assessed indirectly using resident physical status variables, including measures of ADLs (independence in eating, toilet use, and transfer reflect lower acuity levels) and other mobility characteristics (bedfast and chair-bound indicate higher acuity levels, independently ambulatory indicates lower acuity levels) expected to discriminate between subacute and traditional residents. (Analysis of ADL measures was confined to 1991 through 1997 because of questionable comparability of the coding for 1989 and 1990.) As Table 4 shows, the negative association between the percentage of residents admitted from hospitals and the percentage of residents who can feed themselves, transfer, use the toilet, or ambulate independently is statistically significant. As expected, the percentage of residents admitted from hospitals is positively correlated with residents who are bedfast and chair-bound. The opposite patterns of association are true for physical status variables and nonhospital sources of admission. There are positive, statistically significant relationships between the percentage of residents admitted from home, an assisted living facility, or a nursing home who are able to eat, transfer, use the toilet, and ambulate. Being chair-bound or bedfast is negatively correlated with residents admitted from nonhospital sources. Residents admitted from hospitals should have a greater need for help with activities of daily living because they are recovering from an acute illness episode in contrast to more traditional nursing home residents who are likely to have chronic health problems.
Length of Stay
Between 1989 and 1997, long-term stays (101–365 days, and stays longer than 1 year) declined as a percentage of all discharges, whereas short-term stays (1–20 days, 21–100 days) increased. As a percentage of all discharges, stays from 1 to 20 days increased from 24% to 35% of discharges between 1989 and 1997. Over the same period, stays from 101 to 365 days decreased from 18% to 12% of all discharges, and stays more than 1 year decreased from 25% to 17% of all discharges. The steep increase in stays from 1 to 20 days and the modest increase in stays from 21 to 100 days (from 34% to 36% of discharges) are consistent Medicare reimbursement practices. The large increase in 1- to 20-day stays reflects Medicare reimbursement for all costs associated with the SNF subacute admission. The requirement that beneficiaries pay a substantial copayment after the first 20 days of a Medicare-covered SNF stay suppresses the "Medicare effect" on stays from 21 to 100 days. Because the copayment can be higher than either the Medicaid payment or private charges, some Medicare patients convert to private pay or Medicaid on the 21st day, even though they are still eligible for Medicare SNF care (Liu & Kenney, 1993). Still others may be discharged to other sites to receive subacute care.
The correlation between length of stay categories and admission sources shown in Table 4 shows that higher percentages of residents admitted from hospitals are positively correlated with shorter stays. Facilities admitting higher proportions of residents from home, an assisted living facility, or a different nursing home are positively associated with larger percentages of long-stay residents.
Place of Discharge
Two trends in place of discharge from Florida-freestanding SNFs are notable. First, discharges from death declined by one third over the period, from 37% of all discharges in 1989 to just over 24% in 1998. Second, discharges to home increased from approximately 22% in 1989 to 32% of all discharges in 1997. These discharge patterns are consistent with an increasing percentage of Medicare posthospital admissions.
The correlations between admission source and place of discharge show that SNFs admitting a large proportion of residents from hospitals have smaller percentages of residents who die and larger percentages of residents who are discharged to home or to a hospital. Nursing homes with large proportions of residents admitted from home, assisted living facilities, or from different nursing homes are significantly correlated with higher percentages of discharge from deaths.
|
Conclusions |
|---|
|
TOPAbstractFederal and State Health...MethodsConclusionsReferences |
|---|
The response of nursing homes to the reimbursement gap among payers was not unilateral across all segments of the SNF industry. Rather, for-profit and nonprofit SNFs appear to have responded differently to the opportunities that more generous Medicare and Insurance/HMO reimbursement rates provided. As the gaps in revenue increased, for-profit facilities tended to reduce their percentage of Medicaid-paying residents, whereas nonprofit facilities were left to fill the demand by increasing their Medicaid admissions.
The changes in reimbursement streams also created nursing home populations that differed in later years from the beginning of the period. Because many "nontraditional" SNF residents were admitted from hospitals for subacute rather than chronic care, SNFs had an increasingly diverse resident population, with more residents whose physical status was compromised, and who required greater assistance and more complex levels of care. Nursing home residents admitted to SNFs were also younger, stayed for shorter periods, and were more likely to recover and return home than were traditional nursing home residents of the past.
If the objective of long-term care policy is to improve quality, then trends that alter the mix of nursing home residents could benefit all residents, even those receiving traditional long-term custodial care. Residents who are recovering from an acute illness have more visitors than chronically ill residents and have more ties to the community. Higher visibility in the community could provide the impetus to improve nursing home care across the board. Alternatively, an increase in residents admitted from hospitals could increase the workload for staff and thus reduce the quality of care, without associated staffing increases (Kanda & Mezey, 1991).
If the policy objective is to contain costs, then the trends described in this paper are decidedly mixed. On the one hand, Florida appears to have achieved its objective to constrain Medicaid spending for SNFs, both in absolute and relative (to other states) terms, as the rate of growth in Medicaid SNF spending slowed significantly over the period. On the other hand, what Medicare savings were achieved under hospital PPS may have been swallowed by spiraling costs in the Medicare SNF benefit. That is not to say that this is an inappropriate usage of funds. Because studies suggest that Medicare patients left hospitals "quicker and sicker" after implementation of the PPS, then the rise in SNF admissions likely reflects not only SNF strategies to maximize revenues, but also the legitimate needs of Medicare beneficiaries for extended care.
Instead of a reduction in costs, the policy changes described previously have resulted in a shift in revenue streams. Florida SNFs, especially for-profit SNFs, increased their share of more lucrative residents whose care was reimbursed by Medicare and Insurance/HMOs, and decreased their share of residents covered by Medicaid and private payment. The combined effect of state and federal SNF policies was to increase the share of publicly financed nursing home care in Florida by nearly 35%. Thus, an unintended consequence of changes in Medicare policy has been to give states greater capacity to shift part of their long-term care costs to the federal government.
The increasing burden for long-term care costs absorbed by the Medicare program during the 1990s captured the attention of federal policy makers. Congress initiated Medicare SNF benefit payment reform under the Balanced Budget Act (BBA) of 1997. The BBA payment reforms changed the financial incentives that were central to the former cost-based SNF payment system in an attempt to control rapid spending growth for Medicare-covered services. Previously, SNFs benefited from furnishing more ancillary services to Medicare patients, without considering the necessity or the price of the service. The BBA built on successful Medicare hospital PPS cost-constraining measures by creating PPS for SNFs. Under PPS, SNFs would receive fixed, predetermined rates for each day of care provided to Medicare residents, adjusted for case-mix (Emmer, 2000; General Accounting Office, 2000). The intent of the PPS was to create incentives for providers to control their daily costs and deliver more efficient, appropriate care to Medicare nursing home residents.
The change to PPS for Medicare SNF benefits lowered average revenue per resident by nearly 20% between 1997 and 1999 (Medicare Payment Advisory Commission [MedPAC], 2002). Despite Medicare program cost savings, implementation has not been without its problems. Critics of PPS contend that the implementation was flawed. The SNF industry claimed that Medicare reimbursement was too low to cover the care for medically complex postacute beneficiaries and threatened to bankrupt the industry (General Accounting Office, 2000). In 1999, Congress revised nursing home PPS per diem payments imposed by BBA 1997 upward; the Center for Medicare and Medicare Services (formerly HCFA) continues to struggle to define appropriate case-mix adjustors for Medicare SNF reimbursement (MedPAC, 2002). Although the rate of growth of Medicare SNF spending has slowed since 1997, it is too early to assess fully the consequences of SNF prospective payment for nursing home care. The changes legislated in 1999 were intended to improve access to and quality of care for the frailest elderly Americans who are not adequately covered by the PPS (Emmer, 2000, p. 451). However, PPS reforms may have the effect of reducing the quality of care, if nursing homes reduce services or staff levels in response to lower reimbursement levels.
Medicare PPS implementation for SNFs may also have other unintended consequences, if SNFs' policy response is to again attempt to shift revenue streams from newly constrained Medicare benefits to other sources. This could rebalance the share of SNF revenues from Medicare and other payers (like private sources or Medicaid) to more closely resemble the experiences of the early 1990s. This would require resources from individuals or state governments to reimburse what Medicare recently paid. Yet there is no evidence, at least in Florida, that the state is willing to expand its role in SNF reimbursement. Although evidence of the effects of recent Medicare SNF policy changes is just now emerging, it is too soon to conclude that Medicare PPS for skilled nursing care will meet its objectives to deliver cost-effective, high-quality care to all beneficiaries who need it. What is apparent in our review of the Florida case is that the consequences of public policy decisions often differ dramatically from the stated intent of policy makers. It would be no surprise if this is the case for Medicare PPS, as well.
|
Footnotes |
|---|
1 Pepper Institute on Aging and Public Policy, Florida State University, Tallahassee, FL.
Decision Editor: Laurence G. Branch,, PhD
Received for publication May 14, 2002. Accepted for publication October 1, 2002.
|
References |
|---|
|
TOPAbstractFederal and State Health...MethodsConclusionsReferences |
|---|
| ||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
