Printer Friendly

Transitions from home to nursing home in a capitated long-term care program: the role of individual support systems.

There is growing recognition that the system of long-term care financing and delivery in this country is inadequate. The current system is "patched together from multiple funding streams" (U.S. General Accounting Office 1994), and each program has its own rules for determining eligibility and services covered. Covered services often are not the ones that individuals want or need. Traditionally, the formal long-term care system in the United States has been concentrated more on nursing home care than on home-and community-based (HCB) services. In thinking about a system that more appropriately meets individual needs and preferences, recognition is growing among policymakers that such a system should include expanded availability of HCB services (Weiner and Illston 1994). While there is little evidence that expanding HCB services will reduce nursing home use or overall long-term care costs, HCB programs have been growing due to the preferences of elderly clients to remain in their homes and due to the belief by some that HCB care can be less expensive than nursing home care (U.S. General Accounting Office 1993). With this in mind, policymakers have a growing need for data on use of services and duration of service use that might be expected in a reformed system that covers a range of care settings and services.

Much of the literature on long-term care utilization has focused on the use of nursing homes. This literature provides empirical evidence that functional, medical, and cognitive factors are strong predictors of the probability of nursing home use (Branch and Jette 1982; McCoy and Edwards 1981; Coughlin, McBride, and Liu 1990; Liu, Coughlin, and McBride 1991). In addition to such need-related characteristics, research has shown the importance of the client's support system in determining whether or not a client is placed in a nursing home (Boaz and Muller 1994; Freedman et al. 1994; Pearlman and Crown 1992). Most of this research, however, is based on a historically fragmented and uncoordinated long-term care system. Different studies have looked at different portions of the overall long-term care system (e.g., for selected services or for selected populations), but relatively little information exists on utilization in a coordinated long-term care system.

In particular, little is known about the factors associated with nursing home use among long-term care clients who have been receiving formal HCB services. Often nursing home and HCB care have been provided under separate programs with little incentive or data to study the flow of clients between the two settings. A recurring theme in the long-term care reform debate, however, is the need for more coordination between care settings and between acute and long-term care (Leutz, Greenlick, and Capitman 1994; Weiner and Illston 1994; American Association of Retired Persons 1994). If, as some propose, long-term care providers or health plans are to become responsible for providing care in a range of long-term care settings, it is important to understand the factors associated with the flow of clients between the two settings. This is particularly true for capitated, managed care models, where the provider or plan has incentives to reduce a client's functional decline and to provide services at the lowest and least costly levels of care. Understanding the factors associated with the transition of clients from formal HCB care to nursing home care can help policymakers and long-term care providers to identify those at risk of nursing home entry and to develop services or programs aimed at reducing that risk.

This study calculates the risk of being placed in a nursing home for clients receiving HCB care in a capitated long-term care system. The primary focus is on the extent to which data on the characteristics of the client's support system obtained at the time of admission to the program can be used to predict the risk of a subsequent nursing home placement. A previous analysis of the same study population found that clients with stronger support systems were significantly less likely to be placed in a nursing home upon admission to the program (Bauer 1995). Clients with informal support, for example, had a probability of initial nursing home placement that was 27 percentage points less than clients without such a source of help. The study reported in this article follows the same population for a period of up to three years after the initial long-term care placement and examines whether the characteristics of a client's support system at the point of his or her initial placement reduces the risk of a subsequent nursing home entry.



The Arizona Long-Term Care System (ALTCS) is Arizona's alternative to a traditional Medicaid program. It is the first statewide Medicaid long-term care program that promotes the use of HCB services as a substitute for nursing home care. Persons eligible for ALTCS are assessed by state screening personnel and determined to be at risk of entering a nursing home. Information collected by the state via a preadmission screening instrument is used to generate a single ordinal index of functional and medical need. To qualify for the program, individuals must have functional and medical scores that meet specified minimum criteria (McCall, Paringer, Weissert, et al. 1991). If a client is determined eligible for the program by the state screening staff and wishes to enroll, he or she is assigned with the long-term care contractor (also called program contractor) in his or her county of residence.

The program contractors receive prepaid, capitated payments from the state to provide all ALTCS-covered acute and long-term care services to enrolled beneficiaries, and they are expected to develop a provider network that encompasses the full range of such services. Program contractors receive a fixed monthly payment, prospectively determined, for each enrollee. The amount is based on the following four components: (1) projected nursing home costs, (2) projected HCB costs, (3) the expected mix of nursing home and HCB clients, and (4) other costs (including case management, acute care, and administration). Except for ventilator-dependent clients, the contractors receive the same capitation rate for each enrollee. No adjustments to the rate are made for functional or need levels that may result in different levels of utilization and cost. The prepaid, capitated financing provides an incentive to the program contractors to manage carefully the care delivered to enrolled beneficiaries in order to be able to provide all required services within the budget determined by the sum of their capitation payments. While all clients admitted to the program have been determined by the state to be at risk of institutionalization, the capitated payment methodology provides an incentive for contractors to substitute HCB care for more costly nursing home care when appropriate. (For a more detailed description of the capitation rate-setting methodology, see McCall, Korb, Crane, et al. 1994). Initial results from the HCFA-funded evaluation of ALTCS have shown that the program has been successful in providing HCB services as a cost-effective alternative to nursing home care (McCall, Crane, Bauer, et al. 1992).

A key feature of the ALTCS program, for the purpose of this study, is that both nursing home care and HCB care are covered services, thus enabling us to assess the factors that are associated with the transition between care settings. The sociodemographic and health and functional status data used in this study are collected upon application to the program by the state screening staff using a standardized preadmission screening instrument.


Of the 5,789 noninstitutionalized elderly and physically disabled (EPD) clients who had a long-term care placement in the ALTCS program during a three-year period-January 1989 through December 1991-about half (n = 2,866) were initially placed in a nursing home and the other half (n = 2,923) in HCB care. The analysis excluded developmentally disabled clients and clients who entered the program already residing in a nursing home. We followed those clients initially placed in HCB care to assess the occurrence of a subsequent nursing home placement. We tracked the program experience of these HCB clients from the date they entered the program until the end of December 1992. Clients entered the program over a three-year period, and thus the period of observation was not equal for each client.

Survival analysis was used to examine the length of time from initial HCB placement until nursing home admission. Clients who were still in their initial HCB placement at the end of the study period (December 1992) were considered "censored" observations. Clients who died, and "discontinued" clients, that is, clients who lost their eligibility for the program or who did not have continuous placement history data in the ALTCS administrative database, were also considered censored because they were removed from the study before the failure event (nursing home entry) was observed.(1)

The first step in the analysis was to estimate the distribution of failure time (i.e., time until nursing home entry). The cumulative distribution function (CDF) is used as a general description of the cumulative hazard of leaving the community due to a nursing home entry. The CDF for a particular risk evaluated at time t is the probability that a client from the study population will have left the community placement due to that risk by time t. The cumulative hazard was estimated using the SAS LIFETEST procedure.

A Cox proportional hazards regression model was then used to assess the individual factors associated with the risk of nursing home entry during any time interval (t, t + [Delta]t) for persons who were still in the community at time t (the risk set). The estimates are conditional upon the client remaining in the risk set at time t. Positive and significant coefficient estimates imply that higher values of the variable are associated with shorter survival times (higher hazard rates). This translates into a shorter duration of survival in the community before nursing home entry. The proportional hazards model was estimated using the SAS PHREG procedure. All explanatory variables were tested for proportionality by examining how the ratio of the hazard functions for different strata of the explanatory variable changed over time.

While the other reasons for leaving the community (i.e., death or being discontinued from the program) are important for calculating the expected total survival probability, the purpose of this analysis was to assess the relative importance of support system variables in helping clients avoid entering a nursing home. Thus, the probability of survival without nursing home entry, while giving an overestimate of the true survival in the community (because clients would also be leaving due to death or program discontinuation), is a suitable measure to assess these relative effects. It is recognized that complete independence between nursing home entry and death is not likely. Previous research suggests that disabilities and medical conditions increase both the likelihood of death and the demand for long-term care. Headen (1992) suggests, however, that conditional independence between the two survival events can be achieved by including in the model a set of variables that control for heterogeneity in the health status of the study population. Such variables include functional limitations, medical conditions, and age.(2)

Description of Variables

The dependent variable in the survival analyses is the length of time spent in the community without a nursing home placement after admission to the program, [LOS.sup.c]. [LOS.sup.c] was not completely observed for all clients. Each client was facing two competing risks that would affect the calculation of his or her ultimate length of stay in the community without a nursing home entry. These were death and being discontinued from the program. Thus, the definition of the dependent variable is:

[LOS.sup.c] = min ([T.sup.nha] - [T.sup.0], [T.sup.d] - [T.sup.0], [T.sup.i] - [T.sup.0], [T.sup.e] - [T.sup.0])

where [T.sup.0] is the date of entry into the program, [T.sup.nha] is the date of nursing home admission, [T.sup.d] is the date of death, [T.sup.i] is the date of ineligibility or discontinuation from the program, and [T.sup.e] is the end of the study period.

Explanatory variables (Table 1) were created based on measures that have been shown in the literature to be predictors of nursing home use. The variables of primary interest for this study were those related to the client's support system. Covariates to control for other differences in the study population were also included. The explanatory variables were categorized into the following four groups: (1) demographics, (2) support system, (3) client need, and (4) market characteristics.

Demographic variables included age, gender, racial/ethnic categories, and home ownership status. Support system variables included a number of variables that describe the potential availability of support and caregiving [TABULAR DATA FOR TABLE 1 OMITTED] for the client. In general, it is expected that a stronger client support system reduces the amount of formal services needed to maintain the client in the community, thus making it more likely that an HCB placement falls within the program contractor's cost parameters for a community placement.

Different configurations of support system variables were included in two separate Cox proportional hazards models. Model 1 included the following three support system variables: married, receives paid help, and receives informal care for ADLs. Some research has shown, however, that the composition of the informal care network is an important factor in predicting nursing home use, and that a simple measure of presence or absence of informal care is inadequate. Pearlman and Crown (1992) suggest the importance of using multiple measures of social support, rather than a simple measure of its presence or absence, to investigate the relationship between social support and nursing home use. Several studies have found that having a spouse for a caregiver significantly lowers the risk of nursing home entry (Freedman et al. 1994; Boaz and Muller 1994; Pearlman and Crown 1992). Other studies have found that adult children as caregivers are important in reducing the risk of nursing home entry (Boaz and Muller 1994; Pearlman and Crown 1992).

Model 2 included indicator variables for the type of the client's primary caregiver (spouse or parent was omitted as a reference group) and replaced the informal care variable with indicator variables for the following non-mutually exclusive indicators of the types of people that make up the client's support network: daughter, son, other relative, or other helper (spouse as a helper was omitted due to its strong correlation with the spouse as a primary caregiver variable).

Client need factors included functional status, medical conditions, cognitive impairment, incontinence, and prior use of hospital services. Functional status was measured by the extent of client limitations in performing activities of daily living (ADLs). The following seven activities were considered ADLs on the ALTCS preadmission screening instrument: toileting, bathing, dressing, grooming, eating, mobility, and transferring. Each activity was assessed during the client's initial preadmission screening and was scored using a four-point scale: 1 = independent, 2 = minimum assistance, 3 = moderate assistance, 4 = maximum assistance. For this analysis, clients were considered to be limited in a certain activity if they were found always to need human assistance in performing the activity (i.e., they had a score of 3 or 4 for an activity). Indicator variables represented the presence of selected medical conditions among the study population at the time of the preadmission assessment. Other client need measures included an assessment by state screening personnel of whether the client was exhibiting behavior detrimental to himself or herself, whether the client was totally bladder incontinent, and whether the client had had an acute hospitalization within six months prior to the date of the preadmission assessment. An indicator variable for rural counties and a measure of county nursing home bed supply were included in the model to capture any differences in the supply of long-term care services across the counties.


Table 2 shows the percentage and number of HCB clients whose initial placement ended with a transition to a nursing home or with one of the censor events. The table presents the transition and censor experience separately by the year of the initial long-term care placement and by the client's initial placement setting. For all cohorts combined, about one-third of the clients initially placed in the community were observed to have left their initial placement due to nursing home entry, but only 16 percent were observed to have left their initial placement due to death. When compared across the placement cohorts, Table 2 shows that the earlier cohorts experienced a larger percentage of deaths as of December 1992 than the later cohorts. Accordingly, the percentage of clients who remained in their initial placement at the end of the study period is much lower in the earlier cohorts. The dynamics in the client population and the varied experience across cohorts highlights the importance of using analytical techniques that can account for differences in observed time in the study population.

Figure 1 plots the cumulative hazard rates in 30-day increments from the date of initial placement in the community for each of the "failure events." The hazard estimates represent the cumulative probability that a client from the study population will exit the community due to the failure event during the interval ([t.sup.i-1], [t.sup.i]), given that they are still in the risk set at time t, where t is a measure of the number of days from the date of the client's initial long-term care placement ([t.sup.0]). The probability of nursing home entry was generally greater than that of either death or discontinuation from the program. At 360 days from initial HCB placement, clients faced an approximately 25 percent chance of having been admitted to a nursing home, a 12 percent chance of having died, and an 18 percent chance of having been discontinued from the program, given that they were still in the risk set up until that point in time. At 720 days, 39 percent of those initially placed in the community would have experienced a nursing home admission, 21 percent would have died, and 33 percent would have been discontinued from the program.
Table 2: Percentage and Number of HCB Clients by Reason for End
of Initial Placement, by Year of Placement

Reason for End of
Initial Placement Percent Number

All, 1989-1991 100.00 2,923

Death 16.01 468
Transition 32.16 940
Discontinued 24.97 730
None 26.86 785

1989 100.00 1,078

Death 18.92 204
Transition 32.00 345
Discontinued 27.55 297
None 21.52 232

1990 100.00 824

Death 15.53 128
Transition 32.77 270
Discontinued 28.64 236
None 23.06 190

1991 100.00 1,020

Death 13.33 136
Transition 31.86 325
Discontinued 19.31 197
None 35.49 362

An analysis of transition rates by age group revealed a significantly lower risk of transition for the nonelderly physically disabled (those less than 65 years of age) than for the elderly (those age 65 and over). Only about 18 percent of the nonelderly group had made the transition into a nursing home by the end of the study period, whereas nearly 40 percent of the group over age 65 had. Tables 3 and 4 present the estimated coefficients and standard errors for the explanatory variables included in the two different Cox proportional hazards models. To accommodate the different transition rates, each of the models is run separately for the elderly and the nonelderly. While the two groups were similar along most of the client characteristics, the nonelderly group had a higher percentage of male clients (48 percent versus 26 percent for the elderly) and lower rates for most of the medical conditions, except for bladder incontinence. Twenty-seven percent of the nonelderly were totally bladder incontinent, compared with only 19 percent of the elderly group.

In Model 1 (Table 3), having paid help at the time of admission to the program was the only statistically significant support system variable; however, the effect of this variable was the opposite for elderly versus nonelderly. For the under-age-65 group, those who indicated receiving paid help at the time of the preadmission screening were significantly less likely to enter a nursing home than those without paid help. For the elderly, however, those with paid help had a 33 percent higher risk of nursing home entry than other clients. For these latter clients, it is uncertain whether the presence of paid help captures an unmeasured degree of client frailty or illness, is associated with a lesser ability to rely on informal sources of help, or perhaps represents some other unmeasured aspect of the client's support system. Contrary to our initial hypothesis, those who were married and/or receiving informal care upon admission to the program did not have a significantly lower risk of nursing home entry after their initial placement in HCB care.

In each of the two groups, strong effects on the risk of nursing home entry were observed for those who were older and for those clients with Alzheimer's disease. For the elderly, those who were Hispanic, African American, or Native American were observed to have a significantly lower risk of a subsequent nursing home placement than all other clients (omitted as the reference group). For the nonelderly, only the Native American ethnic group [TABULAR DATA FOR TABLE 3 OMITTED] was observed to have a significantly lower risk of nursing home entry. Men in the elderly group were found to have a 23 percent greater risk of nursing home entry than women. No effect for gender was detected for the nonelderly.

The results for Model 2 are shown in Table 4. Replacing the support system variables with indicators for the client's primary helper and for the composition of the client's informal support system did not produce significantly different results than those seen in Model 1. In Model 2 we see that for the nonelderly, none of the variables that describe the composition of the support team is significantly associated with the risk of nursing home entry. For the elderly, the only significant results were observed for the variables indicating a son as a source of informal help or a nonrelative ("other") as a source of help. The regression results show that those with a son as a caregiver [TABULAR DATA FOR TABLE 4 OMITTED] have a 28 percent higher risk of nursing home entry than those without a son as a caregiver. While previous research has found that wives and daughters are more likely to be caregivers than sons and husbands (Stone, Cafferata, and Sangl 1987), it is not possible to determine from these data why the presence of a son in the support network increases the risk of nursing home entry. It could be that sons are only called upon to help as a last resort, but are then unable to sustain a major caregiving role. This relationship warrants further study. Those with an "other" source of help were found to be significantly less likely to enter a nursing home. There were few significant differences in the regression results between Model 1 and Model 2 for the other variables in the models.

The support system variables were measured at the time of the initial preadmission assessment and do not capture subsequent changes to the support system. Because such changes are less likely in the short time interval after the initial placement, we conducted a separate analysis of clients admitted to a nursing home within 90, 180, and 360 days from their initial HCB placement to see if the support system variables were more strongly predictive of nursing home entry. The results of that analysis, however, did not show an increased effect for the support system variables (results not reported here).


Policymakers concerned with long-term care reform need to have information on use of services and duration of service use in order to plan and budget for any new long-term care program. This study documents the rate at which clients who are newly admitted to HCB care in a capitated long-term care system move to a nursing home, die, or leave the program. We saw that nearly one-third of the initial HCB placements were, at some point during the study period, placed in a nursing home, but only 16 percent of the HCB clients had left their initial placement due to death. The percentage of HCB clients who undergo transition to a nursing home was fairly consistent across the three placement cohorts (1989, 1990, and 1991); thus, it is not clear that this percentage would increase if the cohorts could be followed over a longer period of time. Differences between the HCB placement cohorts were primarily in the percentage who had died or left the program.

Given the large percentage of clients who move from HCB to nursing home care (40 percent for the elderly and 18 percent for the nonelderly), the objective of this study was to use data collected upon admission to the program to identify the factors associated with a greater risk of subsequent nursing home entry. Understanding the factors associated with transitions can help policymakers and providers to identify clients at risk of entering nursing home care and to develop services or programs aimed at reducing or postponing that risk. We had hypothesized that clients with stronger support systems would have a lower risk of a subsequent nursing home entry. This hypothesis was formulated from the literature and from the findings of a previous analysis of the same study population that found that clients with stronger support systems were significantly less likely to be placed in a nursing home upon admission to the program (Bauer 1995). In that previous study, those who were married were 9 percentage points less likely to be initially placed in a nursing home, those with paid help were 14 percentage points less likely to be placed in a nursing home, and those with informal help were 27 percentage points less likely to be placed in a nursing home upon being newly admitted to the program.

In looking at the relative effects of client characteristics, the results of this study suggest that nursing home entry may be affected more by factors related to frailty (e.g., age and health conditions) than to characteristics of a client's support network. A particularly strong effect was also observed for the racial/ethnic indicator variables. Among the elderly, Hispanics, African Americans, and Native Americans had significantly less risk of nursing home entry than the omitted group, consisting primarily of white clients. While a higher risk of nursing home use has been observed for whites in the literature, it cannot be determined from these data whether the observed effects represent different cultural preferences or whether they are the result of unmeasured differences in access to nursing home care.

The lack of significant effects for the support system variables could be due to a number of factors. One possibility is that data collected at the time of a client's preadmission screening for the program are not sufficient to capture the changes in client status, such as changes in health, functional, or social support status, that may be important in predicting subsequent nursing home use. Although we attempted to overcome this limitation by conducting a separate analysis of clients who had moved to a nursing home shortly after admission to the program, actual measures of functional decline, acute episodes, or support system changes may be needed to predict subsequent nursing home entry. Such factors have been found to be precursors of nursing home admission in other studies, with changes in health status being the most important type of reason cited (Arling and McAuley 1983; Wetle, Walker, De Matteo, et al. 1994). Further study is needed of the factors that lead to a transition from the community to the nursing home, and of whether data collected at the time of admission to the program can be used to predict subsequent nursing home use.

The lack of effect of support system variables may also be due to the length of time that the study population was observed. No client was followed for more than four years. Because those initially placed in the community were typically younger and less sick than those placed in nursing homes, it is possible that support system factors could have become more important in preventing nursing home use when observed over a long period of time, as the clients moved into the oldest age categories and highest levels of frailty.

It should also be noted that these data are limited to the experience of one state that has a unique long-term care program. Results from this study should be compared with the experience of other states having different programs for Medicaid beneficiaries before these findings can be generalized to other states or programs.

Caveats aside, the results suggest that the composition of a client's support network once the client is placed in HCB care does not appear to be strongly associated with subsequent nursing home use. Some have proposed that greater attention be given to helping bolster community support in the effort to keep people out of nursing homes (Boaz and Muller 1994). How to accomplish this goal most appropriately warrants further study. This analysis suggests that such efforts may be more productive if they focus on the point at which clients are first assessed for placement into the ALTCS program. Once in HCB care, subsequent risk of nursing home placement may be related more to the client's health and frailty than to support system factors.


The author is grateful to Nelda McCall and to Teh-wei Hu for their guidance and substantive comments throughout this study.

This article is based on work funded by the National Institute on Aging (NIA)-funded Research Training Program in Economics of Aging and Health Services at the University of California at Berkeley and by Health Care Financing Administration (HCFA) Contract # 500-89-0067, Evaluation of the Arizona Health Care Cost Containment System. The analysis and conclusions contained herein are solely those of the author and do not express any official opinion of or endorsement by the NIA or HCFA.


1. Of the 2,923 clients who were initially placed in HCB care, 730 were classified as "discontinued" due either to loss of program eligibility (n = 76) or to lack of continuous placement history data in the ALTCS administrative database (n = 654). Chi-square tests were conducted to test for statistically significant differences in the characteristics of those clients who were classified as discontinued and those who were not discontinued. Discontinued clients were younger, were more likely to be male, were more likely to be Native American, had fewer ADL limitations, had fewer medical conditions, and were more likely to live in a rural county.

2. An investigation was undertaken to assess the robustness of the results when death is treated as a censoring event. The results indicated that treating death as a censoring event does not significantly affect the Cox regression results.


American Association of Retired Persons (AARP). 1994. "Integrating Acute and Long-Term Care: Advancing the Health Care Reform Agenda." Conference Proceedings, AARP, Washington, DC.

Arling, G., and W. J. McAuley. 1983. "The Feasibility of Public Payments for Family Caregiving." The Gerontologist 23 (3): 300-306.

Bauer, E. Jones. 1995. "The Role of Individual Support Systems in Determining Home or Nursing Home Placement in a Capitated Long-Term Care System." Unpublished manuscript. Research Training Program in Economics of Aging and Health Services, University of California at Berkeley, April.

Boaz, R. F., and C. F. Muller. 1994. "Predicting the Risk of 'Permanent' Nursing Home Residence: The Role of Community Help as Indicated by Family Helpers and Prior Living Arrangements." Health Services Research 29 (4): 391-414.

Branch, L. G., and A. M. Jette. 1982. "A Prospective Study of Long-Term Care Institutionalization among the Aged." American Journal of Public Health 72 (12): 137-379.

Coughlin, T., T. McBride, and K. Liu. 1990. "Determinants of Transitory and Permanent Nursing Home Admissions." Medical Care 28 (7): 616-31.

Freedman, V. A., L. F. Berkman, S. R. Rapp, and A. M. Ostfeld. 1994. "Family Networks: Predictors of Nursing Home Entry." American Journal of Public Health 84 (5): 843-45.

Headen, A. E. 1992. "Time Costs and Informal Social Support as Determinants of Differences Between Black and White Families in the Provision of Long-Term Care." Inquiry 29 (4): 440-50.

Leutz, W. N., M. R. Greenlick, and J. A. Capitman. 1994. "Integrating Acute and Long-Term Care." Health Affairs (Fall): 58-74.

Liu, K., T. Coughlin, and T. McBride. 1991. "Predicting Nursing Home Admissions and Length of Stay: A Duration Analysis." Medical Care 29 (2): 125-41.

McCall, N., J. Korb, M. Crane, C. W. Wrightson, S. Pollack, L. Paringer, and W. Weissert. 1994. Fourth Implementation and Operation Report: Evaluation of the Arizona Health Care Cost Containment System Demonstration. San Francisco, CA: Laguna Research Associates.

McCall, N., L. Paringer, W. Weissert, P. Turner, M. Crane, C. W. Wrightson, and E. Jones Bauer. 1991. First Implementation and Operation Report: Evaluation of the Arizona Health Care Cost Containment System Demonstration. San Francisco: Laguna Research Associates.

McCall, N., M. Crane, E. Jones Bauer, W. Weissert, L. Paringer, and C. W. Wrightson. 1992. Second Implementation and Operation Report: Evaluation of the Arizona Health Care Cost Containment System Demonstration. San Francisco: Laguna Research Associates.

McCoy, J. L., and B. E. Edwards. 1981. "Contextual and Sociodemographic Antecedents of Institutionalization among Aged Welfare Recipients." Medical Care 19 (9): 907-21.

Pearlman, D. N., and W. H. Crown. 1992. "Alternative Sources of Social Support and Their Impacts on Institutional Risk." The Gerontologist 32 (4): 527-35.

Stone, R., G. Cafferata, and J. Sangl. 1987. "Caregivers of the Frail Elderly: A National Profile." The Gerontologist 27 (5): 616-26.

U.S. General Accounting Office (GAO). 1993. Long-Term Care Case Management: State Experiences and Implications for Federal Policy. Report to the Honorable Ron. Wyden, House of Representatives.

-----. 1994. Long-Term Care: Demography, Dollars, and Dissatisfaction Drive Reform, Testimony before the Special Committee on Aging, U.S. Senate, Statement of J. L. Ross, Associate Director, Income Security Issues, Health Education, and Human Services Division. Publication No. GAO/T-HEHS-94-140. Washington, DC: Government Printing Office. 12 April.

Weiner, J. M., and L. Hixon Illston. 1994. "Health Care Reform in the 1990s: Where Does Long-Term Care Fit In?" The Gerontologist 34 (3): 402-408.

Wetle, T., L. Curry Walker, G. De Matteo, L. Gruenberg, and C. Gruman. 1994. The Transition from Community to Nursing Home: A Survey of Recently Admitted Nursing Home Residents. Reprint Paper #6-94. Hartford: Partnership for Long-Term Care, Office of Policy and Management, State of Connecticut.

Address correspondence and requests for reprints to Ellen Jones Bauer, Ph.D., Senior Research Associate, Institute for Health Services Research, University of Minnesota, Box 729, 420 Delaware Street SE, Minneapolis, MN 55455-0381. This article, submitted to Health Services Research on July 12, 1995, was revised and accepted for publication on January 5, 1996.
COPYRIGHT 1996 Health Research and Educational Trust
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1996 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Bauer, Ellen Jones
Publication:Health Services Research
Date:Aug 1, 1996
Previous Article:Chronic disease, functional health status, and demographics: a multi-dimensional approach to risk adjustment.
Next Article:The effects of ownership and ownership change on nursing home industry costs.

Terms of use | Copyright © 2018 Farlex, Inc. | Feedback | For webmasters