Impact of Growing Numbers of the Very Old on Medicaid Expenditures for Nursing Homes: A Multi-State, Population-Based Analysis

Transcription

1 Impact of Growing Numbers of the Very Old on Medicaid Expenditures for Nursing Homes: A Multi-State, Population-Based Analysis WAYNE A. RAY, PHD, CHARLES F. FEDERSPIEL, PHD, DAVID K. BAUGH, MA, AND SUZANNE DODDS, MS Abstract: We utilized Medicaid data from five states which account for 39 per cent of Medicaid expenditures to study the impact of the near-trebling of persons age 85 and older (the very old) projected to occur by the year 2012 upon Medicaid nursing home expenditures. We found a one-year prevalence of Medicaid-covered nursing home residence of 20 per 100 among the very old. If this rate continues, with no changes in current levels of Medicaid nursing home payments, and if population forecasts are accurate, increasing Introduction In the past five years, intense interest has been focused upon the population 85 years of age and older, the very old. 1' This group is projected to be the most rapidly growing population segment for the next 25 years8'9 with a forecasted increase from 2.5 million in 1982 to 6.9 million by the year These numbers may underestimate the actual growth; previous projections have fallen short because ofgreater than anti9ipated declines in mortality rates.`13 Because the health needs of the very old are extensively funded by the public sector (primarily Medicare and Medicaid), this trend will shape the allocation of resources for public financed health care over the next 25 years. The greatest impact on health care financing may be an attendant increase in demand for nursing home care. The 1977 National Nursing Home Survey'l found an estimated 22 per cent of persons age 85 and older residing in a nursing home on any given day, over four times the rate for all elderly persons (age 65 and older). Thus, the rapidly increasing numbers of the very old can be expected to intensify concern over the expanding nursing home population9""'2 and rising nursing home expenditures.9" 1-13 The fact of this trend is well known and its potential impact upon overall public expenditures for the elderly, particularly for Social Security and Medicare, has been explored. '-4'7 However, the possible effects upon Medicaid expenditures have been less well documented. Medicaid pays for nearly half of all nursing home care'4 and expends 40 per cent of its budget for nursing home care.'4 Furthermore, Medicaid serves other disadvantaged and vulnerable populations: low-income single mothers and their dependent children, and the adult disabled and medically needy. Therefore, unless Medicaid funding is increased, strong growth in expenditures for Medicaid-financed nursing home services could reduce the resources available for these groups, which in turn could lead to poorer levels of health. We utilized data from five state Medicaid programs, which together account for 39 per cent of total national Medicaid expenditures,'5 to study the possible impact of the Address reprint requests to Wayne A. Ray, PhD, Department of Preventive Medicine, and Director, Division of Pharmacoepidemiology, Vanderbilt University School of Medicine, Nashville, TN Dr. Federspiel is also affiliated with the Department of Preventive Medicine at Vanderbilt; Mr. Baugh is with the Office of Research and Demonstration, Health Care Financing Administration; Ms. Dodds is with SysteMetrics, Inc. This paper, submitted to the Journal June 12, 1986, was revised and accepted for publication December 16, c 1987 American Journal of Public Health /87$1.50 numbers of the very old will generate an additional $6.3 billion (1982 dollars) annually of Medicaid nursing home payments by 2012: an increase of 280 per cent from 1982 levels. The stress this trend will place upon societal ability to check growth in public expenditures for medical care while maintaining basic services for other low income populations will be an important force shaping public health policy in the next 25 years. (Am J Public Health 1987; 77: ) growing numbers of the very old upon Medicaid expenditures. First, we calculated the one-year prevalence of Medicaid-covered nursing home residence for the very old in California, Georgia, Michigan, New York, and Tennessee during Then, we applied this rate to compute national estimates of the growth in Medicaid expenditures for nursing homes that would occur if current population forecasts prove correct, if the prevalence of Medicaid-covered nursing home residence does not change, and if current long-term care reimbursement levels continue. While we recognize that some changes in these factors are a virtual certainty, we suggest that the projections developed can stimulate discussions of the public health implications of increasing Medicaid expenditures for the very old in nursing homes. Methods Medicaid Data Medicaid is ajoint federal-state program which provides medical care for qualifying poor.15 Not all of the poor qualify for Medicaid; a recent estimate was that only 52 per cent of those with an income below the poverty level had Medicaid coverage.'6 While Medicaid is financed by both the federal government and the states, detailed data files are maintained by the individual states. Recently, the Health Care Financing Administration (HCFA) has employed these files to create a uniform, person-based Medicaid data set for policy and research analysis.'7 This database, which currently encompasses the states of California, Georgia, Michigan, New York, and Tennessee, was utilized for this study. In 1981, the time of this study, there were notable differences among these states in the Medicaid eligibility criteria for the elderly.'5 Four of the states automatically extended Medicaid eligibility to Supplemental Security Income recipients; however, New York did not. Unlike the other states, Georgia had no medically needy program. Among the four states with medically needy programs, the income level for an aged couple was between $5,000 and $6,000 for California, Michigan, and New York but was $1,600 for Tennessee (1982 dollars). The study population consisted of persons ages 65 and older who were Medicaid enrollees and had Medicaidcovered nursing home residence during calendar year 1981 in the five states. It excluded the very small number of elderly Medicaid Aid to the Blind recipients and the few Medicaid enrollees (at the time of the study) who were part of a health maintenance organization. We also excluded stays in facilities for the mentally retarded (ICF-MR). 699

2 RAY, ET AL. Projections and Rates We employed the United States Bureau of the Census "middle-series" population projections.8 National average annual Medicaid expenditures were calculated from HCFA Medicaid statistics for 1982,'" when there were a reported 1.3 million Medicaid recipients receiving nursing home services and reported nursing home payments of $9.4 billion, for an average of $7,100 per nursing home recipient. 5 It is possible that the denominator was somewhat inflated by doublecounting of recipients; however, our estimates are conservative in this circumstance. The projections used the age and sex-specific prevalence rates of Medicaid-covered nursing home residence in the study states for calendar year Each rate was computed as the number of persons in a nursing home at any time during 1981 (numerator) divided by the size of that specific demographic subgroup (denominator),18,19 yielding a one-year prevalence rate. A person was counted only once in the numerator, regardless of the number of nursing home admissions during the year. The rates reported by the National Nursing Home Survey,'0 often cited in other works, are computed by dividing the number of persons in a nursing home on a given day (numerator) by the population size (denominator) yielding a point-prevalence rate. One-year prevalence rates are higher because the numerator includes persons who enter and exit the nursing home during a one-year period but who would be missed in a single-day census. This rate can depict more accurately nursing home utilization than does the point prevalence, which is less sensitive to the large numbers of persons whose stay in the nursing home is short.20 The one-year prevalence rate is also more suitable for projections of Medicaid expenditures, as it parallels the annual HCFA fiscal reporting conventions. Because the study was descriptive and populationbased, no tests of statistical significance were performed. Results Characteristics of Medicaid Nursing Home Residents There were 257,449 Medicaid enrollees age 65 and older who resided in a nursing home in the five states during 1981 and met the other study criteria. Of these, 75 per cent were female and 45 per cent were very old, 85 years of age and older. As expected, those 85 years of age and older were more likely to be female: 82 per cent of very old residents were female, as opposed to 69 per cent of persons ages 65 through 84. There was little interstate variation in these demographic characteristics. Rates of Medicaid-covered Nursing Home Residence In 1981, the estimated elderly population of the five states studied was 6.7 million (Table 1). The very old, ages 85 and older, comprised an estimated 9 per cent of all elderly. These census estimates were the denominators for the annual prevalence rates of Medicaid-covered nursing home residence during 1981 (Table 1) for the five states. Among all persons 65 and older, the one-year prevalence of Medicaidcovered nursing home residence was 3.8 per 100, with little interstate variation. The very old had a nearly 10-fold greater one-year prevalence of nursing home residence covered by Medicaid than did those of ages 65 to 84 (Table 1: 20 per 100; as opposed to 2.3 per 100). Women had greater one-year prevalence rates of Medicaid-covered nursing home residence than did men. For the very old, women were nearly twice as likely as men to be both Medicaid enrollees and reside in a nursing home during 1981 (22.8 per 100 versus 11.7). Within each specific subgroup, there was little interstate variation in these annual rates of Medicaid-covered nursing home residence. Indeed, TABLE 1-Population Size and One-Year Prevalence of Medicaid-Covered Nursing Home Residence by Age, Sex, and State, 1981 Califomia Georgia Michigan New York Tennessee Total States Population Size Total 2,268, , ,400 1,994, ,300 6,101,700 Male 949, , , , ,900 2,496,400 Female 1,318, , ,300 1,200, ,400 3,605,300 Total 225,100 40,800 84, ,500 42, ,200 Male 66,100 11,000 26,000 58,800 13, ,000 Female 159,000 29,800 58, ,700 29, ,200 Medicaid Nursing Home Residents Total 51,297 15,257 19,581 42,111 14, ,586 Male 16,837 4,647 6,286 12,092 4,442 44,304 Female 34,460 10,610 13,295 30,019 9,898 98,282 Total 42,451 8,461 16,270 39,460 8, ,863 Male 7,154 1,441 3,077 7,253 1,582 20,507 Female 35,297 7,020 13,193 32,207 6,639 94,356 One-Year Prevalence (per 100) Total Male Female Total Male Female

3 MEDICAID NURSING HOME PAYMENTS FOR THE VERY OLD for women 85 years of age and older, the group with the highest prevalence, the rates for the five states were all between 22.2 and 23.6 per 100. Projections The projections are restricted to those 85 years of age and older, the very old. We applied the sex-specific, one-year prevalence rates of Medicaid-covered nursing home residence in the five states during calendar year 1981 to the Census Bureau forecasts of the numbers of the very old for the next 25 years8 to estimate the number of very old, Medicaid nursing home residents. The projected national Medicaid expenditures were computed by multiplying the national average annual Medicaid expenditure per nursing home resident in 1982 ($7, 100)15 by the projected number of Medicaid-covered nursing home residents. Although Medicaid enrollees in the nursing homes of the five study states do not constitute a random sample of Medicaid-covered, nursing home residents, we believe these estimates of the prevalence of Medicaid nursing home residence will model national experience for three reasons: * First, the Medicaid programs of the five states accounted for 39 per cent of total national Medicaid expenditures in ; therefore, trends in these states will have a significant impact on national Medicaid expenditures. * Second, the point-prevalence rates of nursing home residence in the entire United States were slightly higher than those in California and New York (4.8 per 100 for the US versus 4.5 per 100 for California and New York,'0 the two largest states in our study, suggesting that our projections are smaller than would be those based upon national data. * Third, aggregate data show that among those age 65 and older, the one-year prevalence of Medicaid-covered residence in a nursing home at any time during 1981 was 4.4 per cent for the entire US,14 while the comparable rate in this study was 3.8 per cent for the five states, suggesting our rates are underestimates. Thus, although our projections would not be appropriate for detailed, short-term actuarial forecasts, they illustrate the potential long-term impact of the unchecked continuation of current trends. Table 2 shows the near trebling of the number of the very old forecasted to occur by the year 2012: from 2.5 million in 1982 to a projected 6.9 million by Among those age 65 and older, the proportion of the very old will nearly double, rising from 9.3 per cent in 1982 to 16.6 per cent by The relative proportion of females among the very old also is projected to increase during this period as well, rising from 234 females per 100 males in 1982 to nearly 276 females per 100 males by During this same period, if the one-year prevalence of Medicaid-covered nursing home residence for the very old is as estimated in this study and remains constant, the annual number of such Medicaid-covered, nursing home residents will increase nearly three-fold, from an estimated 489,000 per year in 1982 to 1.4 million per year by If the expenditures per resident do not change, Medicaid expenditures (in 1982 constant dollars) for nursing home services for the very old will rise during this period from $3.5 billion in 1982 to $9.8 billion by 2012, a projected increase of $6.3 billion (1982 dollars.) Discussion The very old have an impact upon Medicaid programs which is strikingly disproportionate to their numbers. In this study, persons 85 years and older accounted for 1 per cent of the states' populations and 9 per cent of all elderly. However, they comprised 20 per cent of elderly Medicaid enrollees and 45 per cent of all elderly Medicaid nursing home residents. Nearly one in four was both a Medicaid enrollee and in a nursing home at some time during the study year. Because of the anticipated rapid growth in the number of very old over the next 25 years, our findings projected an additional $6.3 billion in annual Medicaid nursing home payments for the very old by 2012, an increase over 1982 total Medicaid long-term care expenditures and trebling of such expenditures for the very old. There are four primary sets of parameters we estimated which determine the accuracy of our projections: the number of very old; among the very old, the rate of nursing home residence; among these very old nursing home residents, the rate of Medicaid coverage; and Medicaid nursing home expenditures per covered, very old resident. TABLE 2-National Prolections for the Very Old (85+): Population, Medicaid-Covered Nursing Home Residents, and Expenditures by Sex, 1982 to 2012 Medicaid Nursing Homes Projected Population (1000s)a Residents (1000s)b Expenditures Year Total Male Female Total Male Female (1982 Dollars, Millions)c , , , , , , ,796 1,052 2, , ,634 1,267 3, , ,472 1,479 3,994 1, , ,313 1,691 4,623 1, ,054 8, ,922 1,841 5,081 1, ,158 9,754 asource: Siegel JS, Davidson M.8 bsource: Application of sex-specific one-year prevalence rates for Medicaid nursing home residence from five-state data (see Table 1). The rate for males 85+ was.117; that for females.228. CComputed assuming annual Medicaid payments of $7,100 per recipient ever in a nursing home, which was the case for 1982 (Source: Sawyer D, Ruther M, Pagan-Berlucchi A, Muse DW.)15 No adjustments for inflation, change in nursing home costs, or service mix were made. 701

4 RAY, ET AL. Number of Very Old This estimate is the most likely to be conservative. Those people who will be 85 years of age and older during the next 25 years are already born; indeed, they were at least 55 years old in The accuracy of population projections for this age group hinges upon the accuracy of anticipated age-specific mortality rates for persons 55 and older. Even if age-specific mortality rates show no further decline from 1976 rates, the number of the very old is expected to increase to 4 million by However, in the recent past, mortality rates among older adults have decreased even more than anticipated. If this continues, the number of very old could be substantially larger than those shown in our projections. For example, the "high series" Census Bureau projections forecast as many as 8.5 million persons 85 years of age and older by 2012,8 1.6 million more than the "middle series" projections we used. Given current trends, this translates into an annual excess of 312,000 more very old nursing home residents than we projected, generating an additional $2.2 billion (1982 dollars) of annual Medicaid expenditures by the year Rates of Nursing Home Residence Studies of determinants of nursing home residence4'21-24 show two primary factors which drive demand: poor health (often expressed as need for assistance in ambulation or activities of daily living), and lack of viable alternatives to the nursing home. Being very old is strongly related to both of these factors. There are estimates that 70 per cent of those 85 years of age and older require assistance with some of the activities of daily living,2 that 40 per cent need help with walking,2 and that the incidence of dementia is highest among the very old.5 Persons age 85 and older often have outlived their support systems: spouses have died and children or other relatives who provided care previously are no longer able to do so.24 At the present time, it is not clear whether future medical advances will reduce or increase the numbers of the very old with chronic disabilities. Societal trends which are thought to increase the use of the nursing home, such as "rising income, the increased propensity to live alone, higher mother-todaughter ratios, and higher labor force participation rates by young and middle-aged women,"22 appear to be intensifying rather than weakening. Nursing home bed shortages in some states with limitations on new construction attest to the strength of nursing home demand. Increased use of community-based care has been proposed as a means of reducing nursing home utilization. However, attempts to do this have had limited success.23 Indeed, community-based care can be as or more expensive than nursing home care. 13 Furthermore, there are substantial numbers of impaired elderly maintained in the community now: some authors have estimated that for every impaired person in a nursing home, there are between one and three equally impaired persons in the community.'3'24 Supply of nursing home beds is another factor which influences nursing home utilization.25 Supply can be influenced by economic conditions (cost of construction, interest rates), Medicaid reimbursement rates, Medicaid enrollment regulations (which, for example, control the "spend down" process, by which private-pay nursing home residents can convert to Medicaid), and governmental regulations, such as state "Certificate of Need" programs25 which require preauthorization for construction of new beds. If supply restraint slows growth in nursing home utili- 702 zation, there may be increased utilization in other sectors, particularly home health services. We did not include home health care in our projections because, at the time of our study (1981), Medicaid home health expenditures were only one-twentieth of those for nursing homes'5;thus, they would have changed the projections by no more than 5 per cent. However, the annual compound rate of growth in Medicaid payments for home health care was 39.1 per cent between 1973 and 1982,15 which suggests that in the future home health care payments may account for a substantial portion of Medicaid expenditures for the elderly. Medicaid Enrollment Rates Medicaid coverage of nursing home care frequently has been cited as a force in creating nursing home demand.2122 Medicaid enrollment is determined by income and asset levels among the very old and Medicaid eligibility criteria, which are set by government regulation. It has been noted that, in recent years, the overall rate of poverty among the elderly has decreased.2"22'26 Yet, these studies grouped the very old with other elderly persons; studies which focus specifically upon those age 85 and older do not necessarily suggest that rates of poverty will decline appreciably for the very old in the near future.7 Furthermore, the high, long-term costs of nursing home care qualify many of the very old for Medicaid through medically needy provisions. Although insurance coverage of long-term care may eventually shield against this form of impoverishment, it is unlikely that such programs can have a substantial impact in the next 25 years, because those at highest risk of needing nursing home care are no longer in the work force and would have difficulty paying large insurance premiums. The trends of the next 25 years in the other determinant of Medicaid enrollment rates, criteria set by government regulation, are unpredictable. Future income criteria and asset tests may exclude progressively larger proportions of the very old from Medicaid eligibility. Per Capita Medicaid Nursing Home Expenditures The level of Medicaid payments per nursing home resident is another highly variable determinant of future total Medicaid nursing home expenditures. Our projections assume that the 1982 national Medicaid payments per nursing home resident accurately reflect those for recipients 85 years and older and that such payments will not change. Per-person expenditures for the very old may be higher than average because length of stay may increase with increasing age (the result of greater use of nursing homes as "long-term care" facilities and less as convalescent facilities following hospitalization). Three other factors could markedly affect nursing home expenditures per resident: * First, growth in nursing home costs may be different than the rate of increase in the cost of living. Nursing home costs have tended to increase more rapidly than the consumer price index'4; however, cost-containment measures could reverse this trend. * Second, the quality of care delivered by nursing homes may change, requiring adjustment of expenditures. There is currently considerable dissatisfaction with the care provided'3; increasing the quality of care could substantially increase expenditures per resident. * Finally, the case mix of residents could change, either increasing or decreasing the need for professional services, with a corresponding impact on expenditures.

5 MEDICAID NURSING HOME PAYMENTS FOR THE VERY OLD Public Health Implications The Medicaid program, although it covers only selected poor, is the primary source of health care financing for low-income, single mothers and their children, the disabled poor, and the aged poor. Medicaid expenditures for these three groups have increased rapidly since the program's inception in 1967 and there is currently substantial pressure to hold constant or even reduce the level of Medicaid expenditures.'6'27 With such a policy, an increase in expenditures for one group served by Medicaid implies a decrease in spending for the other groups, which must be achieved by lowering reimbursement to providers, reducing the number of persons eligible for services, or decreasing the services per eligible. The latter two alternatives can potentially lead to poorer health outcomes Indeed, it has been suggested that the apparent end ofthe reduction in the perinatal death rate30 is in part attributable to recent reductions in Medicaid and other programs which provide health care for pregnant women. 16'30 The large increases in demand for publicly financed nursing home care which our findings suggest are likely to occur in the next 25 years may further exacerbate this intergenerational conflict.31'32 If future expenditure trends described by our study actually occur and if there is no relaxation of current political constraints upon governmental health-care budgets, unpleasant alternatives must be faced. For example, the projected $6.3 billion annual increase in Medicaid nursing home expenditures for the very old, occurring by 2012, is nearly as much as the entire amount expended by Medicaid for the Aid to Families with Dependent Children program in Clearly, the conflict between the demographic fact of the rapidly increasing numbers of the very old, who are at high risk of both having low incomes and requiring nursing home care, and the societal desire to cap public expenditures for health care will be a most important factor in determining public health policy for the next 25 years. ACKNOWLEDGMENTS This research was supported by HCFA Contract An earlier version of this paper was presented at the 113th annual meeting of the American Public Health Association, November 1985, Washington, DC. REFERENCES I. Suzman R, White-Rily M: Introducing the "oldest old". Milbank Mem Fund Q 1985; 63: Rosenwaike I: A demographic portrait of the oldest old. Milbank Mem Fund Q 1985; 63: Manton KG, Soldo BJ: Dynamics of health changes in the oldest old: New perspectives and evidence, Milbank Mem Fund Q 1985; 63: Soldo BJ, Manton KG: Health status and service needs of the oldest old: current patterns and future trends. Milbank Mem Fund Q 1985; 63: Rowe J, Minaker KL: Health and disease among the oldest old: A clinical perspective. Milbank Mem Fund Q 1985; 63: Torrey BB: Sharing increasing costs on declining income. The visible dilemma of the invisible aged. Milbank Mem Fund Q 1985; 63: Atkins GL: The economic status of the oldest old. Milbank Mem Fund 1985; 63: Siegel JS, Davidson M: Demographic and Socioeconomic Aspects of Aging in the US. Bureau of the Census, Current Population Reports, Series P-23, No Washington, DC: Govt Printing Office, Rice, DP, Rosenberg HM, Curtin LR, Hodgson TA: Changing mortality patterns, health services utilization, and health care expenditures: United States, DHHS Pub. No. (PHS) Washington, DC: GPO, Foley DJ: Nursing home utilization in California, Illinois, Massachusetts, New York, and Texas. DHHS Pub. No. (PHS) Washington, DC: GPO, Russell LB: An aging population and the use of medical care. Med Care 1981; 19: Rice DP, Feldman JJ: Living longer in the United States: demograpic changes and health needs of the elderly. Milbank Mem Fund Q 1983; 61: Kane RA, Kane RL: The feasibility of universal long-term-care benefits: ideas from Canada. N Engl J Med 1985; 312: Burwell B: Institutionalized recipients. In: Short Term Evaluation of Medicaid: Selected Issues. Baltimore, MD: Health Care Financing Administration, Grants and Contracts Reports, October 1984; Sawyer D, Ruther M, Paga-Berlucchi A, Muse DW: The Medicare and Medicaid Data Book, HCFA Pub. No , Mundinger MO: Health service funding cuts and the declining health of the poor. N Engl J Med 1985; 313: Pines PL, Baugh DK, Howell EH, Dodds S: The Medicaid tape-to-tape project: Empirical use of a uniform data base. In: Proceedings 9th Annual Symposium on Computer Applications in Medical Care, Baltimore, MD, November 10-13, 1985; US Bureau of the Census: Provisional projections of the population of states by age and sex: Current Population Reports, Series P-25, No Washington, DC: Govt Printing Office, US Bureau of the Census: Estimates of the population of the states by age. July 1, Current Population Reports, Series P-25, No Washington, DC: Govt Printing Office, Liu K, Manton KG: The characteristics and utilization pattern of an admission cohort of nursing home patients (II). Gerontologist 1984; 24: Etheredge L: An aging society and the federal deficit. Milbank Mem Fund Q 1984; 62: Fuch VR: "Though much is taken": Reflections on aging, health, and medical care. Milbank Mem Fund Q 1984; 62: Branch LG, Jette AM: A prospective study of long-term care institutionalization among the aged. Am J Public Health 1982; 72: Vladeck B: Unloving care: The nursing home tragedy. New York: Basic Books, Swan JH, Harrington C: Estimating undersupply of nursing home beds in the states. Health Services Res 1986; 21: Burwell B: Recent trends in the Medicaid program. In: Short term evaluation of Medicaid: Selected Issues. Baltimore, MD: Health Care Financing Administration, Grants and Contracts Reports, October 1984; Iglehart J: Medical care of the poor-a growing problem. N Engl J Med 1985; 313: Lurie N, Ward NB, Shapiro MF, Gallego C, Vaghaiwalla R, Brook RH: Termination of Medi-Cal benefits: A followup study one year later. N Engl J Med 1986; 314: Wise PH, Kotelchuck M, Wilson ML, Mills M: Racial and socioeconomic disparities in childhood mortality in Boston. N Engl J Med 1985; 313: Miller CA: Infant mortality in the US Sci Am 1985; 253: Preston SH: Children and the elderly: Divergent paths for America's dependents. Demography 1984; 21: Preston SH: Children and the elderly in the US. Sci Am 1984; 251: