| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||||
Correspondence: Address correspondence to William D. Spector, Agency for Healthcare Research and Quality, 540 Gaither Road, Rockville MD 20850. E-mail: wspector{at}AHRQ.gov
 |
Abstract |
|---|
 TOP Abstract MethodsResultsDiscussionReferences |
|---|
Key Words: Health care expenditures • Medicare • Medicaid • Long-term care
Between the mid-1980s and the mid-1990s, the provision of home-care services under Medicare expanded dramatically. This was followed by a period of retrenchment associated with a number of federal policy changes, including several provisions enacted in the Balanced Budget Act of 1997 (BBA). Most research on the changing home-care system has focused on home-health use within the Medicare program. These studies found that, after the BBA, the number of Medicare recipients and the value of services received per recipient declined (Bishop, Kerwin, & Wallack, 1999; McCall, Komisar, Petersons, & Moore, 2001; McCall, Petersons, Moore, & Korb, 2003).
Medicare is only one of a number of programs that fund home care, however. The full picture depends on payments from all sources, especially Medicaid, state and local governments, and private expenditures (private insurance and out-of-pocket payments). In this article, we summarize the differences in eligibility and coverage provisions of the various programs that finance home care, review the policy changes that occurred before and after the BBA, and explore whether the decline in home-health expenditures that followed the BBA was associated with lower use of all home-care services, whether there has been a shift in financing from Medicare to other sources, and whether the type and intensity of services received have changed.
Sources of Payment
Medicare reimburses elderly and some disabled nonelderly persons for medical home-health care visits delivered by Medicare-certified home health agencies (HHAs). A physician develops a Medicare service plan and certifies that intermittent skilled nursing care or therapies are needed and the recipient is homebound. Licensed nurses and aides are reimbursed only if they are needed on an intermittent or part-time basis. Personal care is only reimbursed when there is also a need for skilled care or therapies. Medical social services, medical supplies, and equipment are also covered, although there is a copayment for medical equipment. A Medicare managed-care plan must include Medicare home-health services. Medicare also covers hospice care, most of which is provided at home (Centers for Medicare and Medicaid Services [CMS], 2003a).
Under the Medicaid state plan, Medicaid covers home care for individuals who meet state-specific income and asset requirements or, in states with a medically needy program, who meet those requirements after medical expenses have been accounted for. Medicaid provides medically necessary (as certified by a physician) home-health services for individuals entitled to nursing home care. Services are provided by licensed nurses and home-health aides and also include medical supplies, medical equipment, and appliances. Therapies and personal care are optionally reimbursed under the Medicaid state plan. All services under the plan cannot be denied based on diagnosis, illness, or condition, and coverage rules must be the same for the entire state.
Medicaid funds home care under 1915c waiver authority to provide services not usually covered by Medicaid for persons who would otherwise be institutionalized. These services may include personal care, adult day, respite care, and services for individuals with chronic mental illness and can be targeted to subpopulations in the state, in contrast to services covered in the state plan. All states have waivers, and states often operate several waiver programs, each offering a separate service package to a different group of individuals, although enrollment may be limited by the number of slots available. Targeted case management is an option under the state plan to assist recipients in gaining access to services, but it can be targeted to specific groups of Medicaid recipients (Smith et al., 2000).
Other federal sources fund only a small portion of the home-care bill. The Department of Veterans Affairs pays for medical health care for veterans delivered by its hospital-based home-care units. The Indian Health Service provides home-care services to Native Americans. Some personal care services, homemaker, and chore services are provided by programs funded under federal social service block grants administered by the states as well as programs funded under the Older Americans Act administered by Area Agencies on Aging.
Home care is also provided by state and local governments and is typically funded from general revenues; in recent years some states have also used lottery and tobacco-settlement revenue. Almost all states have state-funded programs. In 1997, more than $1.2 billion was spent by programs primarily serving older adults (Kassner & Williams, 1997). These are generally multiservice programs providing personal care, homemaker, chore, respite, and case-management services. They typically serve individuals who do not meet the Medicaid eligibility requirements, and, in some states, they also supplement care not covered under Medicaid. Large state programs exist in California, Illinois, Massachusetts, Indiana, and Pennsylvania (Coleman, Fox-Grage, & Folkemer, 2002; Kassner & Williams, 1997).
Private-funding sources include commercial and private health insurance, including managed-care organizations (MCOs), Medigap, long-term care insurance, and the Civilian Health and Medical Program of the Uniformed Services (CHAMPUS). Worker's Compensation is often included in either private or government funding, because it is a mandated benefit for participating enterprises. Private health insurance generally pays for some medical home-health care, but it may include some personal care and hospice care. Medigap policies cover the copayment for durable medical equipment and may include care for recovery from an illness or injury. MCOs generally limit coverage to medical home-health care and hospice care.
Private long-term care insurance typically covers personal care at home after a waiting period. Growth in the number of long-term care insurance policies sold has accelerated in recent years; however, the proportion of the population with policies remains quite small, and the policies are generally not applicable to nonelderly disabled persons because of pre-existing condition clauses. In 2001, approximately 3.5 to 4 million Americans had long-term care insurance policies in effect (Cohen, 2003).
CHAMPUS covers hospice care, skilled nursing care, and other professional medical home-care services for dependents of active military personnel as well as military retirees and their dependents and survivors. Worker's Compensation covers medically necessary home-care services as a result of injury on the job.
Historical Growth
Home-care expenditures grew rapidly from the end of the 1980s until the mid-1990s, in large part because of the growth in spending under the Medicare program (U.S. Health Care Financing Administration, 1998). There were a number of different reasons for this growth. For example, the institution of the prospective payment system in hospitals in the early 1980s created an incentive to discharge people earlier to lower levels of care, such as home-health care. In addition, stringent controls on eligibility and coverage restricted the growth in home health until eligibility and coverage criteria were liberalized in 1989 as part of a settlement in a beneficiary suit (Bishop et al., 1999). Over the next 8 years the number of visits per user nearly doubled, and many more Medicare beneficiaries used the benefit (Komisar & Feder, 1998; McCall et al., 2001).
Also during this period, intermediaries denied fewer claims, and new technologies were being used that enabled patients to be treated at home (Komisar & Feder, 1998; U.S. Government Accounting Office [U.S. GAO], 1996). States also instituted "Medicare maximization" policies to encourage agencies to treat Medicare as the primary payer for individuals with dual eligibility (U.S. GAO, 1996).
The Interim Payment System
In the first half of the 1990s, Medicare paid for services with few restraints on cost. The main constraint was on cost per visit. The limit for each agency was the sum of the national average cost per visit multiplied by the number of visits for each type of visit for six visit types. Above-average-cost visits could be offset by below-average-cost visits in the same category or in other categories. In addition, although costs per visit were limited, revenue could be increased by raising the number of visits per recipient. A prospective payment system was proposed to reduce the number of visits per beneficiary as well as reduce the average cost per visit. In anticipation of the development of a prospective payment system for Medicare home-health care, an interim payment system (IPS) was instituted as part of the BBA. The IPS changed the reimbursement method by making the cap on average cost-per-visit type more stringent and adding a per-beneficiary cap. Consequently, the interim system discouraged increases in the number of visits per beneficiary but potentially reduced access for high-cost clients.
The BBA also changed the definition of "part-time or intermittent" for coverage to mean fewer than 8 hr a day of skilled nursing and home-health-aide care limited to a total of 28 hr/week. Previously, more than 8 hr/day was permissible under the 28-hr limit, and 8 hr/day for up to 21 consecutive days was also permissible. The definition of intermittent for eligibility was expanded to include individuals who needed care for fewer than 7 days a week. Prior to the BBA, eligible persons could only have care on fewer than 5 days a week for unlimited duration and could get care for 7 days for 2 to 3 weeks (Komisar & Feder, 1998; McCall et al., 2001). In addition, under the BBA, individuals who only needed venipuncture but did not need skilled care were no longer eligible for coverage. Changes were also made to hospital payments that reduced the financial advantage for hospitals to discharge short-stay patients who were in specified diagnostic-related groups to a postacute setting.
Other Policy Changes
Other factors may have contributed to declines in Medicare home-care expenditures after the BBA. Operation Restore Trust was implemented by the Office of the Inspector General and the Health Care Financing Administration to identify fraud and abuse in HHAs as well as nursing homes and medical equipment suppliers. By 1998 it had been instituted in 18 states. This prompted the involuntary closing of hundreds of HHAs that did not meet the integrity standards. Operation Restore Trust also established civil liabilities for physicians who knowingly falsified eligibility certification for beneficiaries. This made many physicians reluctant to refer Medicare patients for services (Medicare Payment Advisory Commission, 2003; U.S. GAO, 2000). A 6-month moratorium on certification of HHAs and the institution of policies in 1998 that slowed down the payment of claims also may have contributed to the decline in Medicare payments.
By January 1, 1999, 14% of HHAs had closed, but the total number of agencies remaining in existence was about equal to the 1996 level. Forty percent of the closures were in three states—Louisiana, Oklahoma, and Texas. These states had above-average rates of people served and numbers of visits per beneficiary. Most closures were in urban areas and in areas with a high concentration of HHAs. As a consequence, findings by the U.S. GAO suggested that closures did not affect overall access, although utilization in 1998 had declined to 1994 levels, and some HHAs indicated a reluctance to take expensive patients. Nevertheless, the variability of utilization rates by counties had narrowed (U.S. GAO, 1999). Following these policy changes, Medicare-funded home-health care use declined. Between fiscal year 1997 and fiscal year 1999, the proportion of enrollees using home health fell by one fifth and the number of visits per user fell by two fifths. Spending per enrollee fell by 52% (Komisar, 2002).
At the same time that Medicare home health was being transformed, many states were attempting to expand Medicaid and state and local home-care programs to help reduce Medicaid expenditures for nursing home care by increasing access to home-care services to encourage substitution. These states also tried to encourage substitution by restraining nursing home use, such as by establishing certificate-of-need requirements and moratoria on new bed construction, establishing incentives to take nursing home beds out of circulation, and expanding mandatory screening of applicants for nursing home use to all applicants irrespective of income (Coleman, 1998).
Many states also expanded their participation in Medicaid Home and Community-Based Services Waiver programs. Although limited to persons eligible for institutionalization, waivers can be targeted to special populations, in contrast to the home-health and personal care provisions that need to have the same eligibility requirements and service availability across the state. The majority of waiver expenditures go to the nonelderly population—individuals with mental retardation or developmental disability or physical disabilities; children with special problems; individuals with brain or head injury; and those with mental health problems. Three quarters of waiver expenditures go to individuals with mental retardation and developmental disabilities (Coleman et al., 2002). Waivers have been the fastest growing component of Medicaid long-term care, whereas growth for personal care and home health has been relatively flat. Waivers totaled $8.1 billion in 1997 and grew by 45% between 1996 and 1997 and 15% between 1997 and 1998. During this period, overall Medicaid long-term care grew in the 5% to 6% range. Medicaid long-term care remains mainly institutional care, with home care representing only 27% of total expenditures (Burwell, 1999). A report by the U.S. GAO indicates that Medicaid waiver services for elderly persons vary greatly across states (U.S. GAO, 2002).
|
Methods |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
Each survey consists of three main components: (a) a household survey, (b) a survey of the medical providers used by those households, and (c) a survey of employers, who are the primary sources of health insurance in the United States. To collect detailed information on individuals and their use of health-care services, each survey used a series of in-person interviews with respondents knowledgeable about health care received by household members. Recall periods in both surveys were determined on the basis of methodological studies and averaged approximately 4 to 6 months. These studies indicated that recall bias remained low, using recall periods of up to 6 months (Cohen & Burt, 1985).
Both the 1987 NMES and the MEPS provide annual information on home-care use and expenditures. Utilization estimates are based on the household survey, which collects detailed information on health-care services received by each household member, whereas expenditure estimates are based primarily on the more detailed and accurate data collected about the reported services from providers. All organizations providing home-care services to sampled persons (including both freestanding and hospital-based HHAs) are included in the medical provider component. This component collects detailed information on the services provided, charges, and sources and amounts of payment associated with each month of home-care services provided to each individual in the household survey. Response rates for home-care organizations in the medical provider component average approximately 88%. For independent providers, both the utilization and expenditure information come from the household respondent (Machlin & Taylor, 2000; Tourangeau & Ward, 1992). Independent providers account for less than 20% of total home-care expenditures. In the small proportion of cases in which expenditure data could not be obtained from the providers or the household respondents, expenditures were imputed by using a hot deck method (Cox & Cohen, 1985).
The NMES and MEPS are the most appropriate sources of data for this study because they allow analyses of home-care use and expenditures at the individual level for persons of all ages and include a comprehensive set of home-care services as well as all sources of payment for care. Other sources of home-care data are not as comprehensive. The Medicare Current Beneficiary Survey, for example, includes only Medicare beneficiaries and does not collect expenditure data for most non-Medicare-covered home-care services. The National Health Accounts (NHA) contains aggregate data on home-care spending but cannot be disaggregated by user characteristics. In addition, the NHA builds on estimates of revenues received by providers classified by type of establishment, whereas the NMES and MEPS build on event-level expenditure flows that are classified by service type. Thus, unlike NMES and MEPS, the NHA does not include hospital-based HHAs or self-employed independent providers in their home-health expenditure estimates. The NHA also includes in its expenditure estimates revenue that would not be included in the NMES and MEPS, such as revenue from nonpatient care activities (CMS, 2003b). A comparison of home-care expenditure estimates from the MEPS and the NHA for 1996 showed that, after differences in scope and definitions of services were adjusted for, overall estimates from the two sources were within 10%. Medicare and Medicaid estimates were within 1% of each other (Selden et al., 2001). Medicare home-health expenditure estimates from the MEPS and the MCBS for 1996 were also very similar. Finally, the MEPS estimate of home-care expenditures from state and local sources in 1997 was very similar to a 1997 estimate from the AARP that was based on a survey of state-funded home and community-based care programs for older persons (Kassner & Williams, 1997).
Annual home-care use and expenditure estimates in NMES and MEPS are based on data collected for each month of the year. The MEPS and NMES are very similar in their methods of collecting use and expenditure data. Respondents are first asked a set of questions about the types of care anyone in the household may have received during the recall period. The respondents are asked in this set about visits to the home because of a health problem. In MEPS, they are prompted with a show card containing a list of home-care services to consider, which includes the following categories of services, along with examples of each: skilled medical care, personal care, household chore services, companionship, and any other type of home care.
For each individual with home-care use, for each provider, and for each month in which the individual received home care, two types of questions are asked. The first is asked only about agency services, whereas the second is asked about agency and independent providers. These questions are as follows. First, "during the visit month, what types of home-care workers from (provider) provided (person) with home-care services"? Types of workers include registered nurses and nurse practitioners, home-health aides, certified nurse assistants, personal care workers, hospice-care workers, social workers, therapists, dieticians, and companions. Respondents are asked about each agency and prompted with show cards listing the possible types of workers. Second, "during the visit month, did someone from (provider) help (person) by providing (service type)"? Types of services include services associated with a hospitalization before and after the visit, medical treatments or therapies, help with daily activities or personal care tasks, and companion services.
Medical treatments are defined as "things like changing bandages, wound care, giving medications, taking blood pressure, or giving shots or injections." Therapies include "physical, occupational, and speech therapy." Daily activities and personal care tasks include "using the telephone, paying bills, shopping, driving, doing housework, preparing meals, bathing, dressing, using the toilet, getting in or out of bed or chair, walking or eating." Examples of companion care include "reading, watching TV, playing games, going for a walk or to a restaurant, or just being there."
Estimation Approach
Both NMES and MEPS use complex clustered-sampling methodologies. To appropriately account for these methodologies and for survey nonresponse, estimates are provided by using population weights and standard errors calculated with SUDAAN, using the strata and primary sampling unit information. When expenditures are compared, all years are adjusted to 1999 prices by use of the overall consumer price index. This allows us to examine changes in the amounts spent on home care relative to changes in the overall price level in the economy (i.e., the opportunity cost of these services).
Overall samples sizes for the NMES and MEPS range from 22,000 to 35,000 individuals, depending on year. Sample sizes of home-care users are 973 for the 1987 NMES, and for MEPS they are 556 in 1996, 588 in 1998, and 394 in 1999. Because of sample-size constraints on the number of individuals receiving home care, for some comparisons across years the data in 1998 and 1999 were pooled, and prior years' data were compared with the mean expenditures during those 2 years. Where sample sizes permit, we make separate estimates for the elderly and the nonelderly population. Estimates with a relative standard error greater than 30% are indicated with an asterisk. All estimates discussed in the text are statistically significant at the.05 level or better.
Payment Sources
We aggregated payment sources into the following categories: Private funding, which included out-of-pocket spending, private and employer insurance (including managed care), CHAMPUS, Medigap, and long-term care insurance. Government funding included Medicare, Medicaid (including home care provided under waivers), state and local funding, and other government funding. Other government funding included funding from the Older American's Act, Indian Health Service, and the Department of Veteran Affairs. Following the practice used in the NHA, we included Worker's Compensation with other government-funded programs. Medicare included funds paid from fee-for-service and Medicare managed-care organization policies.
Service Use
Service-use change was calculated as the difference in the mean number of months a service was provided in 1996 and 1998. Because the funding mix varies so much for nonelderly and elderly persons, comparisons were made separately for individuals aged under 65 and 65 or older (for persons alive in the community at the end of the year). Variables were constructed based on the two questions shown earlier. A number of variables were constructed to determine if the mix and intensity of services changed. We calculated the difference in the mean months of care for total home care and for subtypes of care. Skilled care included care from physicians, nurse practitioners, therapists, and social workers. Therapists included persons who provide respiratory, speech, physical, occupational, intravenous, and vision therapies. Unskilled care included services provided by home-health aides, nurse aides, companions, and homemakers or personal care workers. We calculated the mean number of months with any unskilled care, and, because unskilled care is often accompanied by skilled care, especially under Medicare, we calculated months of "unskilled care only" to determine if individuals were receiving unskilled home care not combined with skilled care. Thus, skilled care and unskilled care only are mutually exclusive categories. They do not add up to the total because the total includes several small categories of service that we did not classify as skilled or unskilled, such as durable medical equipment and dietary services. We also calculated the mean number of months that (a) care was associated with a hospitalization, (b) medical treatments or therapies were provided, (c) a person was helped with daily activities or personal care, and (d) a person received companion services.
|
Results |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
|
|
In 1987, government funds financed less than half of home-care expenditures (Table 3). By 1996, the government funded more than three quarters and Medicare alone paid for more than half (compared with 19% in 1987). Thus, there was a shift from private to public spending over this period for both the elderly and nonelderly population. Most of the decline in private spending was in out-of-pocket expenses; the share paid by private insurance did not change (not shown). The precipitous fall in Medicare-funded home-care expenditures after the BBA was accompanied by a dramatic change in the share of funding from government sources but not an increased share for government as a whole. The overall Medicare share fell to a low of 25% in 1999. For the elderly population, it declined from 61% in 1996 to 32% in 1999. The overall share grew for Medicaid and state and local programs, however. Although the level of Medicaid expenditures for home care remained stable, Medicaid's share rose from 17% in 1996 to almost 33% by 1999. The most dramatic increase, however, was among state and local programs, which had a very minor share in 1996 (1.7%), but funded over a quarter of expenditures in 1999. This increase was particularly large for nonelderly persons.
|
This shift in payment-source burden was accompanied by changes in the mix and intensity of services received by the elderly and nonelderly population (Table 4). Between 1996 and 1998, for elderly recipients, the mean number of months a home-care recipient received paid home-care services during the year declined from 6.0 to 5.0. The decline was mainly in skilled services: care from nurses, care involving medical treatments and therapies, and care associated with a hospital visit. No differences occurred for months of unskilled care only, any unskilled care, or companion care. These results are consistent with a decline in Medicare funding, which mainly funds skilled care. Most state and local programs, which grew in the aggregate, focus on personal care.
|
In summary, between 1996 and 1998–1999, use of paid home-care services declined, with the greatest declines experienced by the elderly population. Expenditures also fell after 1996, falling back to 1987 levels (in 1999 dollars). The large decline in Medicare spending was partially offset by increases in state and local funds, and the share of total expenditures was greater after the BBA for Medicaid and state and local programs, increasing the burden to states.
|
Discussion |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
During 1998 and 1999, states experienced large increases in tax revenues. Tax revenues had quarterly growth rates that peaked in mid-2000 at 11% (The Kaiser Commission on Medicaid and the Uninsured, 2002). This enabled expansions in state-funded services during this period. However, by the first quarter of 2001, state revenues were actually declining by 8%. As a consequence, many states initiated or were planning cuts in services, including home-care services. States were also attempting to cut Medicaid, which on average represents 15% of state budgets (The Kaiser Commission on Medicaid and the Uninsured, 2002). This dramatic change in the fiscal health of states may result in declines in both Medicaid and state and local home-care services. This is important, because the growth in state and local funding after the BBA partially counteracted the overall decline in home-care spending, which in large part resulted from declines in Medicare spending on home health. In this environment, it may be difficult for state and local governments to maintain their increased shares of Medicaid and state and locally funded home care, however. The increase in nonskilled services for the nonelderly population may also be in jeopardy. Attempts at raising state taxes to maintain these kinds of services have been mixed. For example, recently a referendum in Oregon for a temporary income tax surcharge to fund social services was defeated, resulting in cuts in state-funded home-care services (Gustafson, 2003). Despite these financial constraints, both the Supreme Court's Olmstead decision and the federal government's New Freedom Initiative encourage states to provide services in community-based settings rather than in institutions.
This article has highlighted the role of federal and state policies in determining the use and distribution of payments for home care. The findings show that while federal funding was declining, there was an increase in state and local spending for home care. They also show a decline in private spending between 1987 and 1996 when Medicare was increasing, but no significant change after the BBA, even though Medicare spending declined dramatically. These results stimulate more questions for future research. What is not clear is how state and federal policy making interact with each other to affect overall home-care use, as well as how private and government funding interact, and how the change in mix of services received as a result of this funding reallocation affected unmet need and quality of care. Nor is it clear what will happen to the overall level of home-care use and spending, given current economic conditions. The amount of home care and the types of services that will be provided in the future and the impact on elderly and disabled populations will depend on how both federal and state governments respond to a difficult fiscal environment.
|
Footnotes |
|---|
1 Agency for Healthcare Research and Quality, Rockville, MD.
2 Department of Community & Preventive Medicine, University of Rochester, NY.
Decision Editor: Linda S. Noelker, PhD
Received for publication March 25, 2003. Accepted for publication October 1, 2003.
|
References |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  W. J. McAuley, W. D. Spector, J. Van Nostrand, and T. Shaffer The Influence of Rural Location on Utilization of Formal Home Care: The Role of Medicaid Gerontologist, October 1, 2004; 44(5): 655 - 664. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
