HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Li, L. W. Search for Related Content PubMed PubMed Citation Articles by Li, L. W.

Longitudinal Changes in the Amount of Informal Care Among Publicly Paid Home Care Recipients

Correspondence: Address correspondence to Lydia Li, School of Social Work, University of Michigan, 1080 S. University, Ann Arbor, MI 48109-1106. E-mail: lydiali{at}umich.edu

	Abstract

TOP Abstract Methods Results Discussion References

 
Purpose: This study examined how the amount of informal care received by disabled elders changes when they are receiving publicly paid home care, and whether formal service use, disability, caregiving arrangements, and demographic characteristics of older adults predict changes in the amount of informal care. Design and Methods: Hierarchical linear models were estimated, using 3-year data (12 repeated observations) collected from elderly participants (N = 888) in Michigan's Home- and Community-Based Medicaid Waiver Program. Results: The amount of informal care declined in the beginning period when publicly paid home care was received, and then it stabilized. Changes in activities and instrumental activities of daily living and caregiver residence predicted changes in the amount. The living arrangement and age of elders predicted different patterns of change over time. Neither formal service amount nor its change significantly predicted the amount of informal care. Implications: Informal caregivers do not relinquish caregiving when publicly paid home care is available. Expanding community-based long-term care is a means of fostering partnerships between formal and informal caregivers.

Key Words: Home- and community-based care • Substitution • Longitudinal study • HLM • Community-dwelling frail elders • Caregiving • Long-term care

Substitution of Formal for Informal Support
Policy makers' fear that family caregivers will withdraw service when publicly paid home care is available is supported theoretically by the substitution hypothesis, which suggests that formal services replace informal services (Greene, 1983). Therefore, the amount of informal care received by disabled elders should decrease when publicly paid home care is provided. The hypothesis, however, does not clarify whether complete withdrawal or partial reduction in care provision is the likely result. Whereas the former can be interpreted as a relinquishment of the caregiving role, the latter can be interpreted as a respite that may enable informal caregivers to continue caregiving longer.

In spite of intense interest, research to date has not offered a definitive conclusion regarding family caregivers' response to the availability of publicly paid home care. Some cross-sectional studies found a significant negative correlation between formal and informal service use, thereby supporting the substitution hypothesis (e.g., Greene, 1983). Other cross-sectional studies, however, did not find formal and informal service use to be significantly correlated (e.g., Penning, 2002).

A longitudinal design is more appropriate to investigate the question. One comprehensive investigation of the longitudinal relationship between formal and informal care was conducted by Tennstedt, Crawford, and McKinlay (1993). They found that the substitution of formal for informal services occurs for a small proportion of older adults (15–20%) and tends to be transient. Elderly persons who have more severe disabilities and experience the loss of primary caregivers, a change in living arrangement (from living with others to living alone), and increased disability are more likely to experience substitution. These findings are consistent with other longitudinal studies that suggest that the substitution effect is modest and short term (Christianson, 1988; Edelman & Hughes, 1990).

Methodological Limitations of Previous Studies
A limitation of previous longitudinal studies is that the analyses were limited to change in two points in time, even when more points were available (e.g., Christianson, 1988; Edelman & Hughes, 1990; Tennstedt et al., 1993). The dynamics of change that can be revealed in two time points are limited, and the results (e.g., significance and direction of the change) depend on the time points chosen and the interval between them, particularly if substitution was temporary (Tennstedt et al.). In order to detect changes that are brief, frequent assessments are needed. To examine change in the long term, the observation period has to be extended, and the sample size must be sufficiently large. Some longitudinal studies may have limited statistical power to test their hypotheses because of the small sample sizes that they used (e.g., Edelman & Hughes). A primary aim of this study is to examine changes in informal care when publicly funded home care is received, using a design and methodology that overcome the limitations of prior studies.

Variation Among Disabled Elders
Given the diversity among community-dwelling disabled elders and their informal caregivers, a variation in the response to publicly paid home care provision should be expected. According to the substitution hypothesis, the amount of formal service use should be negatively related to the amount of informal care received. Therefore, elderly persons receiving more formal services should have a steeper decline in informal care. Disability has been shown to be the strongest predictor of formal and informal care (Freedman, Aykan, Wolf, & Marcotte, 2004). As mentioned, Tennstedt and colleagues (1993) found that the severity of disability increases the odds of substitution. One reason for this may be that informal caregivers of elders with more disability have a greater need for respite.

Elders' caregiving arrangements, including caregivers' residence and relationship to the elder, may affect how informal care changes over time. Coresidence with informal caregivers has been found to decrease the likelihood of substitution (Christianson, 1988; Edelman & Hughes, 1990; Tennstedt et al., 1993). For example, Christianson found a significant reduction in informal services by caregivers living apart from their care recipients, but not by coresiding caregivers at the 6-month follow-up among participants of the National Channeling Demonstration project. The importance of the relationship between the caregiver and the elder was incorporated into Cantor's (1975) hierarchical compensatory model, in which spouse and adult child caregivers should be less likely than other informal caregivers to withdraw their care. Using the Channeling Demonstration project data to provide further support for Cantor's model, Christianson found some withdrawal from caregiving on the part of friends, neighbors, and other relatives, but not spouses or adult children.

The effects of demographic characteristics on changes in informal care following the receipt of publicly paid home care are less clear. Miner (1995) found that White elders are more likely than Black elders to experience a decrease in informal care when formal services increase. Age, gender, and education were often used as control variables in previous studies, and these were generally found to have modest effects on informal and formal care, after disability and caregiver relationship were controlled for (Kemper, 1992; Penning, 2002).

Changes in Formal Service Use, Disability, and Caregiving Arrangement
In addition to investigating whether the pattern of change in the amount of informal care over time varies by the amount of formal service use, disability, and caregiving arrangement, this study also examines whether and how changes in these characteristics within individual affect the amount of informal care that one receives. According to the substitution hypothesis, an increase in formal services should lead to a decrease in informal services (Greene, 1983). Increased disability has been found to increase the amount of informal care received (Freedman et al., 2004), although it is less clear whether this is the case when publicly paid home care is available. Other research suggests that changes in disability is associated with substitution, along with changes in living arrangement (Tennstedt et al., 1993).

Research Questions and Hypotheses
This research addresses three questions with related hypotheses.

First, how does the amount of informal care received by community-dwelling disabled elders change when publicly funded home care is provided? Hypothesis 1: The amount of informal care received by elderly persons is expected to decline, on average, with the steepest decrease occurring during the initial period of receiving publicly paid home care, and then leveling off.

Second, does the pattern of change in amount of informal care vary by caregiving arrangements, demographic characteristics, the amount of formal service use, and disability of elderly persons? Hypothesis 2: Older persons who receive more formal services, are more disabled, have caregivers who live apart (vs coresiding caregivers) and have nonimmediate family members (vs immediate family) as caregivers, and are White (vs Black) are expected to experience a steeper decline in the amount of informal care received.

Third, do changes in caregiving arrangements, formal service use, and disability predict changes in the amount of informal care received? Hypothesis 3: An increase in formal service use is expected to predict a decrease in informal care. Experiencing an increased disability, changing arrangements to live with the caregiver, and changing caregivers to having immediate family as caregivers are hypothesized to predict an increase in the amount of informal care that one receives.

	Methods

TOP Abstract Methods Results Discussion References

 
Data
Data for this analysis are obtained from elderly participants of Michigan's Home- and Community-Based Medicaid Waiver program. The Medicaid Waiver program was designed to maintain low-income individuals who are at risk of nursing home placement in the community by providing comprehensive home- and community-based services, such as home health aide, homemaking, personal duty nursing, personal emergency response systems, and environmental modifications (Michigan Department of Community Health, 2004). Applicants to the Medicaid Waiver program are screened for eligibility, usually by phone, and the eligibility criteria include (a) a financial criterion that the person's income must be at or below 300% of Supplemental Security Income and assets must not exceed the limit set for Medicaid; and (b) a functional criterion that individuals must require nursing facility levels of care. Those deemed eligible are then interviewed by a social worker and a nurse (i.e., baseline assessment), who use a standardized instrument (Minimum Data Set—Home Care; MDS-HC) to assess the applicant's health status, functional levels, informal support, and basic demographics. As part of the care-planning process, a reassessment is conducted every 3 months, using the same instrument for those admitted to the Medicaid Waiver program.

Particpants
The present analysis was based on a cohort of elderly Medicaid Waiver participants who were admitted to the program in 1999 (January–December). Only those aged 65 or older, either Black or White, were selected, for a total of 3,161 individuals. This cohort was observed for 3 years following baseline assessment. At the end of the 3 years, 29% (n = 902) had died, 25% (n = 779) had been institutionalized and exited the program, and 19% (n = 592) had dropped out for various reasons such as moving out of the area or becoming ineligible. Approximately 28% (n = 888) stayed in the program for at least 3 years. These "survivors" comprise the sample for this analysis. During the 3-year period, the study sample had been assessed on average 12 times, with a range of 7–15 times.

Variables and Measures
Outcome
The amount of informal care was the outcome variable, which was indicated by number of hours per week that the elder received help with basic activities of daily living (ADL; e.g., dressing, bathing, transferring) and instrumental activities of daily living (IADL; e.g., shopping, housekeeping, managing finances) from informal caregivers. It was based on a question in the MDS-HC that asks about the extent of help the elderly person received from family, friends, and neighbors with ADL and IADL tasks over the past 7 days. Because there was a skewed distribution, I transformed this variable by using natural log.

Predictors
I examined four sets of predictors: formal service use, disability, caregiving arrangements, and sociodemographic characteristics. I defined the amount of formal service use as the total number of formal service hours received per week in the following categories: home health aides, visiting nurses, homemaking, meals, volunteers, adult day care, and in-home social workers. I chose these seven service categories because of their potential to replace the services provided by informal caregivers. The MDS-HC records the hours of service received by the older adult from each category in the past 14 days. I summed the hours from each category and divided by 2 to indicate the amount of formal services received per week. I transformed the variable by using natural log to reduce skew.

I used three indicators to measure disability: limitations in ADL, limitations in IADL, and cognitive impairment. The MDS-HC includes eight ADL items that are rated on a 5-point subscale from 0 (independent) to 4 (totally dependent). Similarly, the MDS-HC records clients' difficulty in performing seven IADLs, rated on a 3-point scale from 0 (no difficulty) to 2 (great difficulty). I summed both ADL and IADL scale scores across the respective items. The former ranged from 0 to 32 and the latter from 0 to 14, with higher scores indicating more severe disability. Cronbach's alpha reliabilities of the ADL and IADL scales in this sample were =.92 and =.78, respectively. I measured cognitive impairment by using an adaptation of the MDS Cognitive Performance Scale, which is highly predictive of Folstein's Mini-Mental State Exam scores (Morris, Fries, Mehr, Hawes, & Phillips, 1994). Because most study participants were cognitively intact and few were severely impaired, I classified them into two categories: no (0) or some (1) cognitive impairment.

Caregiving arrangements included the caregiver's relationship to the elder and residence. The former referred to whether the primary caregiver was the elder's immediate family (spouse or child, coded 1) or not (other relative, friend, or neighbor, coded 0); the latter referred to whether the caregiver lived with (coded 1) or apart (coded 0) from the elder.

Sociodemographic characteristics of the elder included race, age, gender, and education. Race was coded as Black (1) or White (0). Age was measured in years. Gender was dichotomized (1 = female), and education was dummy coded (1 = high school or more, 0 = less than high school) because few participants had more than a high school education.

A preliminary analysis showed that formal service use, disability, and caregiving arrangements were not constant but rather changed significantly during the 3-year study period. Their time-varying characteristics provided an opportunity to test whether changes in these variables predict changes in the outcome within individual. In addition, the sample varied in the aggregate values of these time-varying variables; therefore, I could use differences between elders in the mean of the time-varying variables to predict variation in the trajectory of the amount of informal care. Hence, formal service use, ADL limitations, IADL limitations, cognitive impairment, caregiver relationship, and residence each had two measures. One was time invariant, indicated by the mean of the time-varying variables for each participant (e.g., average amount of formal service use during the 3-year study period). Another was time varying, which I used to examine the effects of within person change (e.g., change in the amount of formal service use) on changes in the amount of informal care.

There were missing data for all study variables except race and age, with the ADL scale having the highest percentage (9.1%) missing. I undertook multiple imputations that involved a regression approach and a data-augmentation algorithm to impute missing values, using the software developed by Schafer (1999), and I analyzed three imputed data sets. I obtained a single-point estimate by averaging across the estimates from the three data sets, and I used a formula to calculate the standard errors, taking into account both within- and between-sample variance (Schafer & Olsen, 1998).

Data Analysis
I used hierarchical linear modeling (HLM) based on maximum likelihood estimation to analyze the longitudinal data, using the software developed by Raudenbush, Bryk, Cheong, and Congdon (2000). Compared with other methods of analyzing change, HLM has two major advantages that are particularly relevant for this study. First, the approach can accommodate unbalanced designs, allowing investigators to proceed when the number and spacing of assessments vary across respondents. Although every study participant was supposed to be assessed once every 3 months, the true intervals often varied because of scheduling and other problems. Hence, the number of assessments in the 3-year study period was not the same for all participants, and the interval between assessments was not consistent over time. Second, the HLM approach is quite flexible in modeling individual change; for example, the change may be characterized by linear or polynomial functions. Such flexibility helps to identify changes that are more complicated and often overlooked by other methods. Covariates that change over time (i.e., time varying) can also be incorporated in the model easily (Raudenbush & Bryk, 2002).

From the standpoint of HLM, the repeated measures of the outcome variable (i.e., informal care amount) are nested within individuals and thus the analysis of change has two levels: within persons (Level 1) and between persons (Level 2). At Level 1, the outcome is conceived as varying within individuals and is a function of person-specific change parameters plus error. At Level 2, these person-specific change parameters are multivariate outcomes that vary across individuals. These Level 2 outcomes could be modeled as a function of background or environmental characteristics that vary between individuals, plus an error associated with the individual. Characteristics that change over time can be specified as time-varying covariates at Level 1. Combining Levels 1 and 2 results in a mixed model with fixed and random effects.

The HLM analysis in this study proceeded in three stages. First, I examined the functional form that best described individual change in the amount of informal care over time, with time measured in years. I estimated three models of change (linear, quadratic, and cubic functions of time). I used deviance statistics to assess model fit, and I used likelihood ratio tests to compare the three nested models.

Next, I examined variation between elders in the pattern of change in the amount of informal care received by entering the time-invariant covariates into Level 2 submodels. Then, I entered the time-varying predictors into the Level 1 model as covariates. The time-varying variables were group-mean centered; their estimates, therefore, can be interpreted as representing the effects of within-person change in the time-varying variables on changes in the amount of informal care. (See Raudenbush & Bryk, 2002, pp. 181–183, about centering and interpretation of Level 1 covariates.) I used a level of <.05 for all statistical tests.

	Results

TOP Abstract Methods Results Discussion References

 
Descriptive Findings
The sample averaged 77 years old at the time of application to the Medicaid Waiver program, with most being female and White and having a low level of education (see Table 1).

Individual and Mean Change in Amount of Informal Care
The first HLM analysis examined how the amount of informal care changes over time when publicly funded home care is received. I estimated three models in which the function of time was linear, quadratic, and cubic. Table 2 presents the comparison of the three nested models. The likelihood ratio tests indicate that a cubic model fit the data significantly better than the other two; however, the fixed effect of the cubic term of time was not significant, and the random effects of the cubic term were highly correlated with that of the linear and quadratic terms (not shown). Therefore, I selected a quadratic change model, which fit the data significantly better than a linear model, to represent individual change in the amount of informal care received over time.

View larger version (20K): [in this window] [in a new window]   Figure 1. Informal care amount trajectories of elders with publicly paid home care

Interindividual Differences in the Trajectory of Informal Care Amount
The next HLM analysis tested the hypothesis that the pattern of change over time in the outcome varies by formal service use, disability levels, caregiving arrangements, and elders' demographic characteristics. To test differences between individuals, I entered the average amount of formal service use, disability, and caregiving arrangements across the 3-year study period, as well as race, age, gender, and education, into Level 2 submodels. I examined each variable separately first and then simultaneously, and I dropped nonsignificant variables. Consequently, I omitted the amount of formal service use, race, gender, and education from further analyses because they had neither significant main effects, nor interaction effects with linear or quadratic rates, on the outcome. Model 2 in Table 3 presents the results with time-invariant covariates added.

The results suggest that older persons with more ADL and IADL limitations, cognitive impairment, and immediate family as caregivers received a greater amount of informal care. Caregiver residence and elder age significantly moderated the rates of change in the amount of informal care. Figure 1 illustrates their interaction effects with time. Regarding caregiver residence, two trajectories are displayed; one represents elders who lived with their caregivers throughout the study period, and the other represents elders who lived apart from their caregivers throughout. The latter group of elders had a steeper decline initially and received less informal care throughout compared to the former. Regarding age, Figure 1 depicts the adjusted change curves of elders aged 70 and 85 years old, respectively. The former compared with the latter received more informal care at baseline, but had a steeper decrease in the amount over time. Hence, by the end of the study period, the gap in the amount of informal care between younger and older elderly individuals was almost nonexistent. Together, the time-invariant covariates explained 39.2% of the variance in the intercept (final status of the outcome), but only a small percentage of the variance in linear (0.6%) and quadratic (1.5%) rates.

Effects of Within-Person Changes
I entered the time-varying variables into the Level 1 model as covariates. To test Hypothesis 3, I used a stepwise procedure to remove insignificant variables (see Model 3 in Table 3).

The results suggest that changes in ADL limitations, IADL limitations, and caregiver residence significantly predicted changes in the outcome. Specifically, an increase in ADL and IADL limitations predicted an increase in informal care; also, when those elders who lived apart from caregivers then lived with them, there was an increase in the amount of informal care. The linear and quadratic rates of change were reduced after the time-varying covariates were included, suggesting that the change in ADL limitations, IADL limitations, and caregiver residence explained some of the observed decline in the amount of informal care. (Note that changes in the amount of formal service use were not significant in predicting changes in the amount of informal care.)

	Discussion

TOP Abstract Methods Results Discussion References

 
During the first 3 years that community-dwelling disabled elders received publicly paid home care, the amount of informal care received by them declined, on average. The decrease occurred mostly in the first year after the elders received Medicaid Waiver services. Then the amount of informal care became more stable, with almost no change during the third year. Throughout the time when publicly paid home care was provided, informal services received by elderly persons remained at a level of 18–25 hr per week. These findings suggest that informal caregivers do not relinquish their caregiving activities when publicly paid home care is available.

The decline in the amount of informal care that elders received during the first year when publicly funded home care was provided appears to support the substitution hypothesis; however, neither the amount of formal service use nor changes in formal service use correlated with the amount of informal care and its change. This finding is contrary to the substitution hypothesis and suggests that formal services do not replace informal care.

What accounts for the decline of the amount of informal care? The analysis shows that changes in ADL and IADL abilities and caregiver residence explain some of the observed decline. In particular, the change from living with the caregiver to living apart strongly predicts a decrease in informal care. About 16% of the sample experienced change in living arrangement during the 3-year study period, with more participants changing from living with to living apart from the caregiver. Additionally, although mean disability levels were fairly stable over time, significant within-person change in disability was detected. Research has shown that both paid and unpaid care hours respond to shifts in ADL and IADL disability, but unpaid care hours decrease in response to improvement, especially in IADL, more so than paid care hours (Freedman et al., 2004), which may explain the observed decline in informal but not formal service amount in this study.

Other reasons for the declining trend may be that, as a result of the support provided by the Medicaid Waiver program, informal caregivers become more efficient in providing care. For instance, informal caregivers may need less time to perform some caregiving tasks because of environmental modifications or the use of assistive devices that may increase elder self-care (Penning, 2002). In addition, if the conditions that triggered application to the Medicaid Waiver program were linked to an increase of informal care (e.g., an injury experienced by the elder), the baseline measure may represent the amount of informal care during a crisis, which should decrease over time as it is resolved. Finally, the decline may simply be due to the passing of time. Without a comparison group, it is not known whether the decline in informal care also occurs for elders who do not receive publicly paid home care, particularly when their disability is fairly stable as in this sample.

Study findings indicate that substantial individual differences exist in the change of informal care amount, which implies that the mean downward trajectory does not apply to all. For instance, elders of more advanced age are less likely to have a decline in informal care when receiving publicly paid home care. However, most characteristics of the elders investigated here were not significant in predicting change. Future studies should explore other factors such as characteristics of caregivers. Male caregivers, caregivers in poor health and with competing demands, and caregivers who feel highly burdened with little satisfaction may be more likely to reduce their care when publicly paid home services are available.

This study is apparently the first that included frequent assessments and a relatively long period of time to examine changes in the amount of informal care received by community-dwelling disabled older adults after publicly paid home care is provided. The trajectories of the amount of informal care shown here should quell fears that family caregivers withdraw from caregiving in response to the availability of publicly funded home care. It is well recognized that informal caregivers are the cornerstone of long-term care. However, given the demographic trends, which include smaller family size, increased rates of marital disruption, and increased labor force participation of women, the availability of family caregivers in the future is in question (Tennstedt et al., 1993). The findings here suggest that informal caregivers are committed to caring for their elderly relatives and friends, and publicly paid home care may provide some relief to informal caregivers but does not replace their services. The findings thus suggest that partnerships between formal and informal sectors in providing long-term care to relatively severely disabled elderly individuals in the community are workable and their benefits for elders and caregivers should be examined.

This study has several limitations. The regional nature and the vulnerability of the sample limit the generalizability of the findings. The Medicaid Waiver program in Michigan may not represent other community-based long-term care programs, and its operation may affect how informal caregivers respond to the intervention. The study sample included quite disabled elders, and their disability levels were relatively stable throughout the study period, which may produce a bias toward stable trajectories of informal care. In addition, questions have been raised about the type of data used in this study because the data were collected primarily for administrative and care-planning purposes (Teresi & Holmes, 1992). More recent studies, however, tend to conclude that this type of data is valid and reliable, and that it is appropriate for research purposes (Landi et al., 2000; Lawton et al., 1998).

The issue of sample attrition is important because it excluded elderly persons who were institutionalized, died, or dropped out during the study period, which may lead to biased estimation. I conducted additional HLM analyses that compared the trajectories of informal care amount during the first two years of Medicaid Waiver participation for those who died (n = 176), were institutionalized (n = 128), or dropped out (n = 92) between the second and third year of the study period, and the study sample (n = 888). Findings showed no difference in changes in the amount of informal care between the study sample and the other three groups, which suggests that attrition bias is minimal in this study (results available on request).

	Footnotes

 
This study was supported by an award to L. Li from the Hartford Geriatric Social Work Faculty Scholars Program, funded by the John A. Hartford Foundation and administered by the Gerontological Society of America. An earlier version of this paper was presented at the 8th annual conference of the Society for Social Work and Research in New Orleans, LA. I thank the Michigan Department of Community Health and Brant Fries for making the data available, Pablo Aliago and Natalya Verbitsky for technical assistance, and Aloen Townsend, David Biegel, Roland Zullo, and Elizabeth Essex for commenting on earlier drafts of the manuscript.

¹ School of Social Work, University of Michigan, Ann Arbor.

Decision Editor: Linda S. Noelker, PhD

Received for publication August 23, 2004. Accepted for publication February 16, 2005.

	References

TOP Abstract Methods Results Discussion References

 

• Cantor, M. H., (1975). Life space and the social support system of the inner city elderly of New York. The Gerontologist, 15, 23-27.[Medline]
• Christianson, J., (1988). The effect of channeling on informal caregiving. Health Services Research, 23, 99-117.[Medline]
• Edelman, P., & Hughes, S., (1990). The impact of community care on provision of informal care to homebound elderly persons. Journal of Gerontology: Social Sciences, 45, S74-S84.
• Freedman, V., Aykan, H., Wolf, D., & Marcotte, J., (2004). Disability and home care dynamics among older unmarried Americans. Journal of Gerontology: Social Sciences, 59B, S25-S33.[Abstract/Free Full Text]
• Greene, V., (1983). Substitution between formally and informally provided care for the impaired elderly in the community. Medical Care, 21, 609-619.[Medline]
• Kemper, P., (1992). The use of formal and informal home care by the disabled elderly. Health Services Research, 27, 421-451.[Medline]
• Landi, F., Tua, E., Onder, G., Carrara, B., Sgadari, A., & Rinaldi, C., et al (2000). Minimum Data Set for home care: A valid instrument to assess frail older people living in the community. Medical Care, 38, 1184-1190.[Medline]
• Lawton, M. P., Casten, R., Parmelee, P. A., Van Haitsma, K., Corn, J., & Kleban, M. H., (1998). Psychometric characteristics of the Minimum Data Set II: Validity. Journal of the American Geriatrics Society, 46, 736-744.[Medline]
• Liu, K., Manton, K. G., & Aragon, C., (2000). Changes in home care use by older people with disabilities: 1982–1994. Washington, DC: AARP Public Policy Institute.
• Michigan Department of Community Health. (2004). Choices for older or disabled persons who may need help caring for themselves. Retrieved December 22, 2004, from http://www.michigan.gov/mdch/.
• Miner, S., (1995). Racial differences in family support and formal service utilization among older persons: A nonrecursive model. Journal of Gerontology: Social Sciences, 50B, S143-S153.[Abstract]
• Morris, J., Fries, B., Mehr, D., Hawes, C., & Phillips, C., (1994). MDS cognitive performance scale. Journal of Gerontology: Medical Sciences, 49, M174-M182.
• Olson, L., (2003). The not-so-golden years: Caregiving, the frail elderly, and the long-term care establishment. Lanham, MD: Rowman & Littlefield.
• Penning, M., (2002). Hydra revisited: Substituting formal for self- and informal in-home care among older adults with disabilities. The Gerontologist, 42, 4-16.[Free Full Text]
• Raudenbush, S., & Bryk, A., (2002). Hierarchical linear models: Applications and data analysis methods (2nd ed.). Thousand Oaks, CA: Sage.
• Raudenbush, S., Bryk, A., Cheong, Y., & Congdon, R., (2000). HLM5: Hierarchical and nonlinear modeling. Lincolnwood, IL: Scientific Software International.
• Schafer, J., (1999). NORM: Multiple imputation of incomplete multivariate data under a normal model [Computer software]. University Park: Pennsylvania State University, Department of Statistics.
• Schafer, J., & Olsen, M., (1998). Multiple imputation for multivariate missing-data problems: A data analyst's perspective. Multivariate Behavioral Research, 33, 545-571.
• Tennstedt, S., Crawford, S., & McKinlay, J., (1993). Is family care on the decline? A longitudinal investigation of the substitution of formal long-term care services for informal care. The Milbank Quarterly, 71, 601-625.[Medline]
• Teresi, J., & Holmes, D., (1992). Editorial: Should MDS data be used for research? The Gerontologist, 32, 148-151.[Medline]

This article has been cited by other articles:

				D. L. Roth, W. E. Haley, V. G. Wadley, O. J. Clay, and G. Howard Race and Gender Differences in Perceived Caregiver Availability for Community-Dwelling Middle-Aged and Older Adults Gerontologist, December 1, 2007; 47(6): 721 - 729. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Li, L. W. Search for Related Content PubMed PubMed Citation Articles by Li, L. W.