Printer Friendly

The use of formal and informal home care by the disabled elderly.

The aging of the population has raised questions among policymakers, the elderly, and their families about how to care for the growing numbers of disabled elderly. Most of the elderly prefer care at home rather than in a nursing home, and some see care at home as potentially less costly than nursing home care. Understanding the factors that determine the type and amount of home care used is important for predicting use in the future and developing long-term care policy. This article analyzes the determinants of the hours of formal and informal home care used by the disabled elderly.


A number of previously published studies have presented multivariate analyses of the determinants of the amount of formal and informal home care used. Because data on the receipt of care have been more readily available than data on the amount of care used, many studies have analyzed receipt of care. Studies of the amount of care have focused on either formal or informal care rather than on the total amount of care from all sources. Multivariate analyses of the use of home care can be classified into three groups: those concerning (1) receipt of formal or informal care among the entire population, (2) receipt and amount of formal care among the disabled population, and (3) receipt and amount of informal care among the disabled population.

Studies of Receipt of Care among the Entire Population

Branch et al. (1981); Branch and Jette (1983); Branch, Wetle, Scherr, et al. (1988); Coulton and Frost (1982); Evashwick et al. (1984); and McAuley and Arling (1984) examined the determinants of receipt of formal or informal home care (or in some cases, the number of services received) among the entire noninstitutionalized elderly population. They all found that disability as measured by limitations in activities of daily living (ADL), such as eating, getting in and out of a chair, bathing, and the like, was a strong predictor of receipt of formal care, informal care, or any home care, depending on the study. Other measures of need were also positively related to receipt of formal home care, as was age, which may reflect unmeasured need. Evidence on other factors is limited.(1)

These studies analyze the entire noninstitutionalized elderly population, including the vast majority who do not need care. Consequently, they simultaneously predict both the need for care and the type of care received among those needing it. For some variables, the effects on need for care may dominate any effect on the type of care received.

Studies of Formal Care among the Disabled Population

Other studies restricted their analyses to the disabled elderly. McAuley and Arling (1984), Soldo (1985), Wan (1987), Greene (1983), and Soldo, Wolf, and Agree (1990) have analyzed receipt of formal care or measures closely related to it (number of formal services received and whether the primary caregiver was formal or informal). They all found a significant positive relationship between receipt of formal care and ADL disability. The effects of other measures of need for care were also generally positive. Availability of informal care, measured in a variety of ways, was generally associated with less reliance on formal care. Results for other factors were limited.(2) While useful for understanding the determinants of receipt of formal care, these studies do not provide information on the amount of care used.

Other studies analyzed the amount of home health care used. Manton and Hausner (1987) analyzed a national sample of the elderly using Medicare home health services. Their results suggest that expenditures for home health care are related to disability and medical conditions. Williams et al. (1990) found that the number of home health visits received was significantly related to diagnosis and prognosis. In limiting their analyses to home health care (nursing, therapy, and home health side aide care) these studies excluded personal care (such as help bathing) and supportive services (such as housekeeping), which are extremely important types of home care.

Other studies analyzed expenditures for all formal home care, including personal care and supportive services. Moscovice, Davidson, and McCaffrey (1988) found that expenditures on home care under Minnesota's Alternative Care Grants Program differed significantly by ADL disabilities and across communities. As part of an evaluation of the Channeling demonstration, Corson, Granneman, and Holden (1988) found that expansion of publicly financed home care greatly increased expenditures for formal home care; however, they did not analyze the effects of other variables. Reanalyzing the first six months of follow-up data from the Channeling demonstration, Liu, McBride, and Coughlin (1990) estimated the effect of other variables on expenditures for home health care and for personal care and housekeeping. They found that higher numbers of ADL disabilities, living alone, and a greater availability of home health services were positively associated with expenditures in both types of services; other variables affected the two types of services differently or were not statistically significant.(3)

Studies of Informal Care among the Disabled Population

Analyzing the number of services disabled elderly persons receive from informal caregivers, Greene (1983), in the study mentioned earlier, found that the number of informal services received increased with ADL disability and limitations in psychological and social functioning, and that they decreased with the number of formal services received. Analyzing follow- up data for a small sample of participants in formal service programs, Edelman and Hughes (1990) found that the best predictor of the number of informal services received at follow-up was the number received at baseline. Other variables showing a positive relationship with informal services in some (but not all) of the analyses were disability, being married, living with the caregiver, and receiving fewer formal services.

Others have analyzed the determinants of the hours of informal care received. Stoller (1983) analyzed care given by a small sample of sons and daughters who were primary caregivers of a disabled parent. She found that hours of care increased with the parent's level of disability and age, and decreased if the parent was married. (The parent's age and marital status were not statistically significant for sons.) Married children and sons who were employed or had children under age six gave less care. In the study noted earlier, Moscovice, Davidson, and McCaffrey (1988) also analyzed the relation between the hours of informal care received and the characteristics of the disabled person and the primary caregiver. They found that disability in activities of daily living (ADL) and instrumental activities of daily living (IADL), cognitive impairment, and male sex of the care recipient increased the hours of informal care received; however, none of the other characteristics of the care recipients or of the primary caregiver was statistically significant. Dwyer and Miller (1990) found that, in general, hours of care provided by the primary informal caregiver increased with ADL and IADL disability. The other independent variable included in their analysis, number of unpaid helpers, was significantly related to hours of care only in rural areas. In addition to these three studies, Christianson (1988) found as part of the evaluation of the Channeling demonstration that expansion of publicly financed home care did not substantially reduce the amount of informal care received.

Place of this Study in the Literature

This investigation builds on previous research by using a large sample of disabled elderly persons to analyze the determinants of the amount of home care used, including personal care and supportive services as well as home health care. It differs from studies published previously in two respects. First, it focuses on all home care, including both formal and informal care. Three types of home care are analyzed separately: formal care, informal care from caregivers not living with the disabled elder, and informal care from resident caregivers. The results of these three analyses are also combined to analyze the determinants of the total amount of care used from all sources, formal and informal combined. Second, in analyzing the amount of home care used, determinants of the receipt of each type of care are distinguished from determinants of the amount of care used by those receiving it. As described later, this is done using the two-part model refined by Duan et al. (1983).


A disabled elderly person who lives at home and needs help with such basic activities as cooking, bathing, or eating can rely on care purchased in the formal market or on care provided informally by family and friends (or on both). Economic theory suggests that the amount of formal and informal care used depends on five factors. (See Kemper 1990 for the economic model of the demand for various types of home care that motivated this framework.) The greater the need for care, the more of both formal and informal care the disabled elderly are expected to use.

A higher price of formal home care is expected to decrease use of formal care and increase use of informal care. As a practical matter, however, lack of data on prices prevents testing these hypotheses, forcing the empirical analysis to test instead for the effect of several proxies for the availability of home care.

Higher income is expected to lead to use of more formal care and -- because higher income permits the "purchase" of more informal caregiver time for activities other than care giving, such as paid work, home production, or leisure--use of less informal care.

Finally, greater availability of family is expected to lead to use of less formal care and more informal care. The empirical analysis uses whether the person is married or not and has children or not as an indicator of the availability of family, although information on the opportunity cost of potential caregivers' time, their attitudes toward caregiving, and other characteristics would also be desirable.

In addition to these variables, tastes undoubtedly play a role in the use of home care. Sociodemographic characteristics are included to control for differences in tastes. However, these same variables are also likely to be proxies for unmeasured variables; for example, age is likely to be a proxy for unmeasured frailty and loss of friends and family as persons age.

Living arrangement is not included as an independent variable even though it is clearly related to care arrangements. A disabled elderly person can live alone (with the caregiver visiting to provide care) or with a caregiver. Sharing households makes it easier to give care and saves time. When living together, no time is spent traveling back and forth between two households, and there are economies in home production when two households are combined--for instance, cleaning one house, preparing one meal, and so on, instead of two. Consequently, the amount of formal and informal care used and living arrangement are jointly determined; any factor that affects one also affects the other. Because living arrangement is endogenous, it is not included as an independent variable in the reduced-form model estimated here.

Because the empirical analysis uses cross-section data, however, living arrangements may not be in equilibrium. For example, persons with a recent increase in their need for care are less likely to rely on resident informal care than persons with similar needs who have had time to move in with a caregiver. The analysis therefore includes, as explanatory variables, recent changes that might lead to disequilibrium in living arrangements.

Based on this framework, the empirical analysis estimates the relationship between hours of home care and need for care, availability of formal care (proxies for price), income, availability of family, demographic characteristics, and any recent changes in need that might affect living arrangements. The dependent variables are the hours of three types of home care used: formal care provided by paid caregivers living outside the household (or in a few cases by helping organizations using volunteers); informal care provided by family or friends living outside the household; and informal care provided by family or friends residing with the disabled elderly person. Due to the lack of hours data, this article does not analyze hours of resident formal care (that is, care provided by paid, live-in helpers).


Data were collected as part of the Channeling experiment, a ten-site test of whether public financing of home care would reduce long-term care costs by substituting care at home for care in nursing homes. (See Kemper, Brown, Carcagno, et al. 1988 for a description of the demonstration, its evaluation, and findings; Carcagno, Applebaum, Christianson, et al. 1986 for analysis of Channeling's implementation; and Phillips, Stephens, and Cerf 1986 for detailed documentation of the data collection process.)

The sample consisted of disabled elderly persons who had applied to Channeling or were referred by hospitals, home health agencies, or other health or social service providers and who met two main eligibility criteria: disability and unmet need. For disability, one of three specific criteria had to be met: (1) moderate disability in two or more activities of daily living (bathing, dressing, toileting, transfer, and eating -- plus continence); (2) three severe impairments in instrumental activities of daily living (housekeeping, shopping, meal preparation, taking medicine, transportation, telephoning, and managing finances); or (3) two severe IADL impairments and one severe ADL disability. Cognitive or behavior problems affecting the ability to perform activities daily could count as one of the severe IADL impairments. For unmet need, either the applicant had to have at least two unmet needs for ADL or IADL help expected to continue for at least six months, or the informal caregiver system had to be sufficiently "fragile" that family and friends were no longer able to continue to give the amount of help that they had been giving. Applebaum (1988) estimates that about 5 percent of the elderly population would meet the Channeling eligibility criteria.

The eligibility determination process included a screening interview with the applicant, family members, or care providers. This short interview (usually carried out by telephone) covered, among other things, disability, income, living arrangement, cognitive impairment, behavior problems, and demographic characteristics. About a week and a half later, on average, eligible applicants were assessed in person. If an applicant was not capable of responding to the interview because of cognitive impairment or illness, or for some other reason, the interview was conducted with a proxy respondent. Proxies were the sole respondents in about 30 percent of the cases, and they helped the disabled elderly person answer questions in an additional 30 percent of the cases. In addition to more detail on many of the same variables, the baseline assessment contained information on (1) the hours of care received regularly in the home from formal and informal caregivers who did not live with the disabled elderly person and (2) whether care was provided by anyone living in the household.

Separate baseline interviews were conducted with the primary informal caregivers of a random subsample of those disabled elderly persons who had been screened and found eligible for Channeling between November 1982 and May 1983. The primary caregiver was defined as the family member or friend who, according to the disabled person, helped the most to take care of him or her or to do things around the house. The informal caregiver interview obtained, among other things, information on the hours spent on care by informal caregivers living in the same household. To distinguish time spent helping the disabled elderly person from that which would be spent anyway, the informal caregiver was asked to estimate the percentage of time spent helping with IADL that was "extra time over and above what you would have spent if the person helped weren't ill or disabled." Estimating this percentage is clearly difficult, especially when living arrangements or other circumstances have changed so that no frame of reference exists for estimating the amount of time that is extra. Moreover, respondents may differ with respect to what they consider extra time. For example, male and female caregivers may differ about whether time spent preparing meals or doing housework is extra time or if the time would be spent pursuing those activities whether the family member was disabled or not. Consequently, resident informal hours are undoubtedly subject to measurement error. To the extent that errors in estimates of the percentage of time that is extra are systematically related to the independent variables in the analysis, the conclusions identifying the factors associated with the amount of resident informal care received could be biased. To the extent that measurement error is random, however, such error simply increases the variance of the estimates of the effects on resident informal hours presented later in this article.

The analysis used the three sources of baseline data to conduct a cross-section analysis at the time of enrollment. Table 1 defines the variables used in the analysis. Because the data are for the period before the experimental intervention took place, no distinction is made between the treatment and control groups.

The initial sample of persons who passed the eligibility screen numbered 6,326. From this sample, observations were excluded whenever information on a dependent variable was not present for any of four reasons: (1) the disabled elderly person or a proxy respondent could not be located or refused to respond to the baseline interview (N = 700); (2) the sampled person lived in a personal care home or with a paid live-in caregiver (N = 76) (because the interview did not collect information on hours of care received from staff of personal care homes or paid live-in caregivers); (3) the sampled person had been in a hospital or nursing home for more than two months before the interview (N = 153) (because the interview did not collect hours data in this case); or (4) items necessary to construct the dependent variables obtained from the baseline interview (receipt of formal, visiting informal, and resident informal care, and hours data of formal and visiting informal care) were missing (N = 247). The resulting sample size for the analysis of the probability of receiving each of the three types of care was 5,150 (81 percent of those screened). Of these, 3,117 received formal visiting care and 3,339 received visiting informal care, and these cases were used in the analyses of hours received among those receiving each type of care.
Table 1: Definitions of Variables
Variable Definition
Formal hours Hours per week of in-home assistance
 with medical treatments and ADL and
 IADL limitations provided regularly
 by helpers who do not live with the
 disabled elderly person and who help
 as part of their paid or volunteer
 work (Based on reports of the
 disabled elderly person or their
Informal visiting hours Same as formal hours except provided
 by friends, neighbors, or family
Resident informal hours Hours per week of extra help with
 medical treatments and ADL and IADL
 limitations due to illness or
 disability, that is provided
 regularly by family members or
 friends living with the disabled
 elderly person (Based on reports of
 the primary informal caregiver)
ADL disability The number of activities
 of daily living with which the
 disabled elderly person received help
 from another person (including
 staying in the room in case help is
 needed). These activities include
 eating, getting out of bed or chair,
 getting to the bathroom or using the
 toilet, dressing, bathing, or none
Incontinence Received help from another person in
 changing a catheter or colostomy bag;
 or if no help was received, patient
 accidentally wet or soiled self
 during the past week
Cognitive impairment Displayed disorientation, confusion,
 impairment of judgment, or memory
 loss that affected ability to perform
 activities nearly every day or
 necessitated supervision to ensure
 personal safety (Excludes those who
 also display inappropriate behavior)
Inappropriate behavior Displayed inappropriate behavior that
 affected ability to perform
 activities nearly every day or
 necessitated supervision to ensure
 personal safety (Includes those who
 also have cognitive impairment)
Home medical treatment Regularly received help with medical
 treatment such as changing a
 dressing, taking vital signs,
 providing physical therapy, etc. at
 home from a formal or informal
State home care program Patient lived in one of four sites
 that had a state home care program
 at the time of baseline interview
Medicaid eligible Patient was eligible for Medicaid
 based on interview report confirmed
 by Medicaid agency records
City size Residence by size of city: large
 (250,000 or over), medium
 (50,000-250,000), small (less than
 50,000), or rural
Income (logarithm) Natural logarithm of dollars of
 individual income + 25. A married
 couple's income was divided by the
 ratio poverty income for elderly
 couples to that of singles (1.26).
 Income of a married couple not living
 together was divided in half
Family availability Elderly person has a surviving spouse
 (excluding spouses not living with
 the disabled elderly person) or
Age Age in years (approximated as
 midrange of five-year age categories)
Female Female
Race African American (non-Hispanic),
 Hispanic, or white/other
Completed high school Completed at least a high school
Recently hospitalized Hospitalized during the past two
Health worsened Disabled elderly person's illness or
 health condition first became a
 problem or became much worse during
 the past year
Loss of caregiver Reason for disabled elderly person's
 referral was permanent loss of
 caregiver, or disabled elder was
 widowed, separated, or divorced during
 the past year
Time frame discrepancy Interview was conducted while the
 disabled elderly person was in a
 hospital or nursing home. In these
 cases, hours information refers to the
 time before admission; in all other
 cases the information is for the time
 of the interview

One evident problem posed by the sample design of the evaluation is that resident informal hours are available only for the subsample that also has data from interviews with informal caregivers. To make maximum use of the data on the full sample, the analysis distinguishes between resident and visiting informal care and estimates separate relationships for them. Only the analysis of hours of resident informal care among those receiving that type of care was restricted to the smaller subsample (see the discussion of estimation methodology further on).

Of the applicants eligible for Channeling who were in the subsample selected to have their primary informal caregiver interviewed, a total of 2,484 had a complete disabled elderly person baseline interview that was analyzed in the Channeling evaluation. Of these cases, observations were excluded for any of four reasons: (1) the disabled elderly person lived in a personal care home or with a paid live-in caregiver (N = 32); (2) the disabled elderly person was in a hospital or nursing home for more than two months before the interview (N = 52); (3) information on receipt of resident informal care was missing in the disabled elderly person's baseline interview (N = 1); or (4) the disabled elderly person reported receiving resident informal care, but the caregiver information on resident informal care was inconsistent or missing (due to interview or item nonresponse) (N = 369). Of the remaining 2,030 observations, 1,039 persons received care from an informal caregiver living in the same household. (The others received visiting informal care and, in some cases, formal visiting care.) These 1,039 cases were used in the analysis of hours of resident informal care among those receiving it.

The analysis had to address two data problems: item nonresponse and a time frame discrepancy. Item nonresponse affected less than 2.6 percent of the analysis sample for all independent variables except education, which was missing 6.6 percent of the time. Missing items were assigned the modal or mean value of the known cases, except for income, which was imputed using a hot deck imputation procedure. Dummy variables indicating that the variable had been imputed were included in the multivariate models. Few differences between cases that had been imputed and those that had not were statistically significant. (These coefficients are not reported in the tables.)

The second problem concerned the time to which questions referred. In addition to the basic disability and unmet need criteria for eligibility, applicants had to be living in the community, or if they were hospitalized or in a nursing home, certified as likely to be discharged within the next three months. If the baseline interview was conducted in a hospital or nursing home, the time frame for the dependent variables was before admission rather than at the time of the interview, as it was for the rest of the sample. Eleven percent of the sample had such a time frame discrepancy. The models estimated further on included a dummy variable indicating when such a time frame discrepancy existed, to control for this discrepancy. It was associated, as expected, with significantly lower probabilities of using formal care and of using resident informal care, and with using more hours of formal and informal visiting care among those receiving each of these types of care. (These results are not shown in the tables.)


Table 2 shows the means and standard deviations of the independent variables. As expected for applicants eligible for a home care program, the sample exhibited a high level of need for care. The eligible applicants were old and frail -- the average age was 80, and 87.3 percent had some disability in the activities of daily living of bathing, dressing, toileting, transferring in or out of bed or chair, and eating. Indeed, 22.5 percent were so seriously disabled that they needed help with all five activities. Almost half of the sample had been hospitalized in the last two months. The sample was poor: average income was just under $500 per month. Over two-thirds were not married, and one-fifth had neither a spouse nor a child as a potential caregiver.

Because persons in the sample had applied to a home care program, they presumably were more likely to use home care and less likely to have equilibrium care and living arrangements than similar people who had not applied. In addition, given Channeling's eligibility criteria, eligible applicants may have had more unmet needs for care than is typical of persons with similar disabilities. For these reasons, the sample should not be expected to be nationally representative. Indeed, Applebaum (1988) found that, compared with the national elderly population with similar disabilities, the Channeling sample was less likely to be married and was much more likely to live alone. It was also much more likely to have had a recent hospital stay and to be receiving formal home care than was the national reference group. Thus, although the results of the analysis are presumably generalizable to eligible applicants of other home care programs like Channeling, they cannot be generalized to persons who meet the eligibility criteria but do not apply to the program, or to the entire disabled elderly population.

Table 3 presents the means and standard deviations of the three dependent variables. Consistent with their need for care, virtually all persons in the sample received at least one of the three types of home care. About three-fifths of the sample received visiting formal care, a slightly higher proportion received visiting informal care, and three-fifths received resident informal care. Many received combinations of care-- indeed, 93.1 percent received some form of informal care from those either in or outside the household, and about one-fifth received all three types of care (not shown). As stated earlier, the selected nature of the sample implies that receipt of formal care is greater and receipt of resident informal care is less than for the elderly nationwide with similar disabilities.
Table 2: Means and Standard Deviations of Independent Variables
(Proportions, except for Income and Age)
 Mean Deviation
ADL Disabilities(*)
Five .225 .418
Four .240 .427
Three .116 .320
Two .134 .340
One .158 .364
Accident, last week .477 .500
Catheter or colostomy .099 .299
Cognitive/Behavior problem(*)
Cognitive impairment .310 .463
Inappropriate behavior .157 .364
Home medical treatment .407 .491
State home care program .419 .494
Medicaid eligible .224 .417
City size(*)
Large city/suburb .658 .474
Medium city/suburb .160 .367
Income (dollars per month)(**) 482 271
Availability of family(*)
Married, has child .249 .432
Married, no child .065 .246
Not married, has child .486 .500
Age (years) 80 8
Female .715 .452
African American .223 .416
Hispanic .036 .187
Completed high school .416 .476
Recently hospitalized .482 .500
Health worsened .823 .379
Loss of caregiver .063 .242
Time frame discrepancy .110 .313
* Omitted categories are: "none" under ADL disability,
"continent" under incontinence, "none" under cognitive or
behavior problem, "not married, no children" under family
availability, "small town or rural" under city size, and "white
or other" under race.
** The natural logarithm of income plus 25 was used in the
multivariate analysis.


The amount of care received was large, 46.4 hours per week on average. Most of the care was provided informally by family and friends, 27.0 hours by caregivers living in the same household and 11.9 hours by those living apart. Although formal care accounts for less than one-sixth of total hours, it is large absolutely; the average, 7.3 hours, amounts to almost a day a week. Among those receiving each type of care, the average amount, of course, is even greater. Those receiving visiting informal care, for example, get the equivalent of over two eight-hour days a week on average.

The variation around the mean is extremely high. For the entire sample, the ratio of the standard deviation to the mean ranges from 0.9 for total hours to 2.4 for formal hours. Like those of other forms of health care, the distributions of the amounts of care used are highly skewed. The top 10 percent of users of formal care accounted for 63 percent of all the formal care used. The same was true of visiting informal care. For resident informal care, the top 10 percent used 41 percent of all care used. Finally, for total hours, the top decile used 31.7 percent of the care (not shown).


The analysis uses the two-part method developed by Duan et al. (1983) in the RAND Health Insurance Experiment for distributions that, like these, are skewed and have many zero values. First, a probit model is used to estimate the probability of using each of the three types of care as a function of the independent variables:

Prob(H |is less than~ O) = F(X|beta~) (1)

where F is the cumulative standard normal distribution, H is hours of care, X is the vector of independent variables, and |beta~ is the vector of probit coefficients to be estimated. Second, for cases with positive hours, ordinary least squares regression is used to estimate the natural logarithm of hours as a function of the same variables:

1n(H) = |X.sub.|gamma~ + |epsilon~ (2)

where |gamma~ is the vector of regression coefficients to be estimated.

Thus, a pair of equations is estimated for each of the three types of home care, one estimating the probability of receiving care and the other estimating the logarithm of hours of care conditional on receiving care. Because the amounts of the three types of care are jointly determined, the errors are correlated across the three types of care. This will lead to consistent but inefficient parameter estimates compared with joint estimation for all three types of care. No adjustment has been made to correct for this inefficiency, because the sample size is relatively large and joint estimation of the two-part model for three types of care would be cumbersome. Consequently, significance levels will be somewhat underestimated.

To simplify interpretation of the results, the coefficients estimated for Equation 1 and Equation 2 are transformed into natural units (probabilities and hours). First, the derivative of the probability of using the type of care with respect to any independent variables, X,, is given by:

|Mathematical Expression Omitted~

where f is the standard normal density function. This is calculated forach observation using its value for the vector of independent variables,|X.sub.i~, together with the estimated probit coefficients, and is then averaged acrosomes into conflict with the values of many people and poses difficult implementation problems. Demonstrations and state programs typically have not used the availability of informal care as an eligibility criterion. Instead, they have made case managers responsible for determining benefit levels for individual clients. Yet this simply shifts the responsibility for making these equity judgments to the case manager where the decisions are less visible. In allocating long-term care, in short, it is impossible to avoid judgments about how to treat differences in the availability of informal care.


The author is grateful to Randy Brown for econometric advice; Jon Christianson, Dean Farley, Tom Grannemann, Peter Gottschalk, Chris Murtaugh, Jaana Muurinen, Pam Short, Bill Spector, Brenda Spillman, Robyn Stone, and the anonymous reviewers for helpful comments on an earlier draft of this article; Herman Liau, Beth MacDougall, and Sandy Smoot of Social and Scientific Systems, Inc. of Bethesda, MD for programming support; and Mary Seidenberg for secretarial support.


1. Higher income was associated with more receipt of formal services (Branch, Jette, Evashwick, et al. 1981) and less receipt of informal care (Branch and Jette 1983). Evashwick et al. (1984) found that those who are Medicaid eligible are more likely to use formal care. Availability of informal care was associated with greater probability of receiving any care (McAuley and Arling 1984) and informal care (Branch and Jette 1983), but the evidence of its effect on receipt of formal care was mixed.

2. Soldo, Wolf, and Agree (1990) found a positive effect of income, and McAuley and Arling (1984) found a positive effect of education. Soldo, Wolf, and Agree (1990) found a greater probability of receiving formal care among those eligible for Medicaid, and McAuley and Arling (1984) and Soldo (1985) found a greater probability to exist in urban areas.

3. Age and severe or moderate cognitive impairment were both associated with greater expenditures on personal care and housekeeping and smaller expenditures on home health care. Variables associated only with greater expenditures on personal care and housekeeping were being female, monthly income of $500-$1,000, having no assets, being in the Channeling treatment group, and greater nursing home bed supply; paralysis was associated with smaller expenditures. Variables associated only with greater expenditures on home health care were Medicaid eligibility, cancer, stroke; senility was associated with smaller expenditures.

4. Much of the formal home care that the Channeling sample received was covered by Medicare. Medicare paid for approximately two-thirds of the control group's formal care during the first six months of follow-up. See Corson et al. (1986, Table V.2).

5. The effect of higher income of future cohorts could, of course, be offset by other factors. For example, despite greater incomes at retirement, greater longevity of future cohorts could result in lower incomes late in life when disability is most likely. Or, rising relative prices of formal home care could at least partly offset the effect of rising incomes.


Applebaum, R. A. "Recruitment and Characteristics of Channeling Clients." Health Services Research 3, no. 1 (April 1988): 51- 65.

Branch, L., A. Jette, C. Evashwick, M. Polansky, G. Rowe, and P. Diehr. "Toward Understanding Elders' Health Service Utilization." Journal of Community Health 7, no. 2 (Winter 1981): 80-92.

Branch, L. G., and A. M. Jette. "Elders Use of Informal Long- Term Care Assistance." The Gerontologist 23, no. 1 (February 1983): 51-56.

Branch, L. G., T. T. Wetle, P. A. Scherr, N. R. Cook, D. A. Evans, L. E. Hebert, E. N. Masland, M. E. Keough, and J. O. Taylor. "A Prospective Study of Incident Comprehensive Medical Home Care Use among the Elderly." American Journal of Public Health 78, no. 3 (March 1988): 255-59.

Christianson, J. B. "The Effect of Channeling on Informal Caregiving." Health Services Research 3, no. 1 (April 1988): 99-117.

Carcagno, G. J., R. Applebaum, J. Christianson, B. Phillips, C. Thornton, and J. Will. The Evaluation of the National Long Term Care Demonstration: The Planning and Operational Experience of the Channeling Projects. Princeton, NJ: Mathematica Policy Research, July 1986.

Corson, W., T. Grannemann, and N. Holden. "Formal Community Services under Channeling." Health Services Research 23, no. 1 (April 1988): 83-98.

Corson, W., T. Grannemann, N. Holden, and C. Thornton. Channeling Effects on Formal Community Based Services and Housing. Princeton, NJ: Mathematica Policy Research, May 1986.

Coulton, C., and A. K. Frost. "Use of Social and Health Services by the Elderly." Journal of Health and Social Behavior 23, no. 4 (December 1982): 330-39.

Duan, N., W. G. Manning, Jr., C. N. Morris, and J. P. Newhouse. "A Comparison of Alternative Models for the Demand for Medical Care." Journal of Business and Economic Statistics 1, no. 2 (April 1983): 115-26.

Dwyer, J. W., and M. K. Miller. "Differences in Characteristics of the Caregiving Network by Area of Residence: Implications of Primary Caregiver Stress and Burden." Family Relations 39, no. 1 (January 1990): 27-37.

Easterlin, R. A., C. Macdonald, and D. J. Macunovich. "Retirement Prospects of the Baby Boom Generation: A Different Perspective." The Gerontologist 30, no. 6 (December 1990): 776- 83.

Edelman, P., and S. Hughes. "The Impact of Community Care on Provision of Informal Care to Homebound Elderly Persons." Health Services Research 19, no. 3 (August 1984): 357-82.

Greene, V. L. "Substitution between Formally and Informally Provided Care for the Impaired Elderly in the Community." Medical Care 21, no. 6 (June 1983): 609-19.

Kemper, P. Living Arrangement and the Demand for Home Care: An Illustrative Theoretical Model. Division of Long Term Care Studies Program Note. Rockville, MD: Agency for Health Care Policy and Research, April 1990.

Kemper, P., R. S. Brown, G. J. Carcagno, et al. Spesial Issue: "The Evaluation of the National Long Term Care Demonstration." Health Services Research 23, no.1 (April 1988): 1-174

Liu, K., T. D. McBride, and T. A. Coughlin. "Costs of Community Care for Disabled Elderly Persons: The Policy Implications." Inquiry 27 (Spring 1990): 61-72.

Manton, K. G., and T. Hausner. "A Multidimensional Approach to Case Mix for Home Health Services." Health Care Financing Review 8, no. 4 (Summer 1987): 37-54.

McAuley, W. J., and G. Arling. "Use of In-home Care by Very Old People." Journal of Health and Social Behavior 25, no. 4 (March 1984): 54-64.

Moscovice, I., G. Davidson, and D. McCaffrey. "Substitution of Formal and Informal Care for the Community-Based Elderly." Medical Care 26, no. 10 (October 1988): 971-81.

Phillips, B. R., S. A. Stephens, and J. J. Cerf. The Evaluation of the National Long Term Care Demonstration Survey Data Collection Procedures. Princeton, NJ: Mathematica Policy Research, March 1986.

Rivlin, A. M., and J. M. Wiener. Caring for the Disabled Elderly: Who Will Pay Washington, DC: The Brookings Institution, 1988.

Soldo, B. J. "In-home Services for the Dependent Elderly: Determinants of Current Use and Implications for Future Demand." Research on Aging 7, no. 2 (June 1985): 281-304.

Soldo, B. J., D. A. Wolf, and E. M. Agree. "Family, Households, and Care Arrangements of Frail Older Women: A Structural Analysis." Journal of Gerontology: Social Sciences 45, no. 6 (November 1990): S238-S249.

Stoller, E. P. "Parental Caregiving by Adult Children." Journal of Marriage and the Family 45, no. 4 (November 1983): 851-58.

Wan, T. T. H. "Functionally Disabled Elderly: Health Status, Social Support, and Use of Health Services." Research on Aging 9, no. 1 (March 1987): 61-78.

Williams, B. C., E. K. Phillips, J. C. Torner, and A. A. Irvine. "Predicting Utilization of Home Health Resources." Medical Care 28, no. 5 (May 1990): 379-91.
COPYRIGHT 1992 Health Research and Educational Trust
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1992 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Kemper, Peter
Publication:Health Services Research
Date:Oct 1, 1992
Previous Article:Determinants of hospital tax-exempt debt yields: corrections for selection and simultaneous equation bias.
Next Article:Home care for the disabled elderly: predictors and expected costs.

Terms of use | Copyright © 2016 Farlex, Inc. | Feedback | For webmasters