Copyright 1992 by The Gerontological Society of America The Cerontologist Vol. 32, No. 5, 641-646

A case-control study compared home health care (HHC) users from the 1984 Supplement on Aging to users of other community services and of no community service, matched on age and gender. Examination of specific activities of daily living (ADLs), instrumental activities of daily living (lADLs), and sociodemographic variables showed that HHC users were significantly more limited than controls in every ADL and I A D L In multivariate analyses, HHC use was significantly associated with three ADLs (dressing, going outside, bathing), two lADLs (shopping, heavy housework), and poor health status. Key Words: ADL, IADL, Home care

Functional Limitations Among Home Health Care Users in the National Health Interview Survey Supplement on Aging1

Home health care (HHC) covers a variety of nursing and physician-supervised services that are provided to individuals at home. Home health care is used both as an alternative to institutionalization by frail elderly, and as a post-hospital service for persons who are unable to independently perform basic activities of daily living (ADLs) at discharge. Home care services, which included both home health care and homemaking services, were used by 19.7% of functionally dependent elderly, as reported in the 1987 National Medical Expenditure Survey, making these the most frequently used community services among elderly persons with limitations in ADLs and instrumental activities of daily living (lADLs) (Short & Leon, 1990). Studies have shown a strong association between functional limitations and home health care use. Home health care (defined as visiting nurse or home health aide services) was used by 6.5% of elderly persons limited in lADLs only, 8.3% of those with one to two ADL limitations, 16.8% of those with three to four ADL limitations, and 38.2% of those with five to seven ADL limitations among respondents to the Supplement on Aging (SOA) to the 1984 National Health Interview Survey (Hing & Bloom, 1990). Similarly, in the 1987 National Medical Expenditure Survey, utilization of home care services increased from 13% among elderly persons limited only in lADLs, to 19.4% among those limited in one to two ADLs, to

36.3% among elderly limited in three or more ADLs (Short & Leon, 1990). Other studies have found that functional limitations were among the strongest and most significant determinants of HHC use (Beland, 1986; Branch et al., 1988; Evashwick et al., 1984; Hughes, Cordray, & Spiker, 1984; McAuley & Arling, 1984; Soldo, 1985). Needing help with at least one of six ADLs was the strongest predictor of incident medical home care use in a 2-year prospective study of elderly community residents (Branch et al., 1988). Impairment in at least one of the Rosow-Breslau Functional Health Scale items and receiving help for IADL impairments were also independently predictive of incident home care use in univariate and multivariate models. Recently, the focus on the association between functional limitations and home health care needs has shifted to the usefulness of ADLs and lADLs in determining eligibility for reimbursement for longterm care services, including home health care services (Kane, Saslow, & Brundage, 1991; Spector, 1991; Stone & Murtaugh, 1990). Several studies have estimated the number of persons who would be eligible for home health care insurance benefits, based on different combinations of ADL and IADL limitations (Spector, 1991; Stone & Murtaugh, 1990). The studies cited above do not address two basic issues regarding the association between functional limitations and home health care use. The first issue is the importance of specific ADL and IADL limitations in determining need for long-term care services. The studies cited above employed counts of ADL and IADL limitations. However, home health care services might be sought for some limitations but not for others. Also, estimates from both the Supplement on Aging (Hing & Bloom, 1990) and the 1984 National Long-Term Care Survey (NLTCS) (Stone & Murtaugh, 1990) show a wide variation in the population-based prevalence of specific functional limita-

1 The authors thank Laura Gardiner for her computer programming assistance. This study was funded by grant # 1 R03 HS06049-01 from the National Center for Health Services Research. department of Family Medicine, University of Maryland School of Medicine, 405 West Redwood St., First Floor, Baltimore, MD 21201. 3 Center on Aging, Georgetown University School of Medicine. 'Division of Health Care Studies, Department of Community and Family Medicine, Georgetown University School of Medicine.

Vol. 32, No. 5,1992

641

Downloaded from http://gerontologist.oxfordjournals.org/ at University of York on April 16, 2015

Lisa Fredman, PhD,2 Janet A. Droge, MA,3 and David L. Rabin, MD, MPH'

Methods

The 1984 National Health Interview Survey Supplement on Aging contained a section on community services used within the past year (Fitti & Kovar,

1987). Of the 16,148 respondents, 418 (3%) reported using HHC, defined as receiving home health aide or visiting nurse services. OCS users were respondents who reported the use of one or more other community services in the past year. These services included senior center, special transportation, homedelivered meals, meals in senior center, homemaker services, and adult day care {n = 2,340). 10,982 respondents reported using neither HHC nor other community services within the past year. Each HHC user was matched to one OCS and to one NCS user by gender and by age in years, using a random selection design in PC-SAS. These variables were used for matching due to their association with HHC use and functional limitations in the SOA dataset (Stone, 1986) and in previous studies (Berk & Bernstein, 1985; Evashwick et al., 1984; U.S. Department of Health and Human Services, 1986). Seven cases with no exact age matches were matched to controls of the next closest age, by randomly rounding up or down 1 year. ADL activities were bathing/showering, dressing, eating, transferring, walking, going outside, or using the toilet. IADL activities were shopping, preparing meals, using the telephone, doing heavy housework, and doing light housework. These variables were dichotomized into respondents who reported difficulty performing the activity, versus respondents who had no difficulty performing the activity or who did not perform it for a reason unrelated to physical functioning. Other covariables included educational level, marital status (married vs. other), income, living arrangement (alone vs. with others), and self-rated health (poor vs. fair to excellent). The SOA sampling design could result in selecting two people from the same household for the SOA interview, and in proxy responses for persons who were absent or mentally or physically incapable of participating (Fitti & Kovar, 1987). Of the SOA responses, 8.5% were by proxy. Mantel-Haenszel odds ratios and associated 95% confidence intervals were calculated for associations between individual functional limitation, demographic variables, and HHC users compared with OCS and with NCS users (Kleinbaum, Kupper, & Morgenstern, 1982). Multiple logistic regression analyses, using PROC CATMOD (SAS Institute, Inc., 1988), were performed to assess the associations between functional limitation and demographic factors and HHC use. The matched design was not maintained in these analyses due to the unavailability of conditional maximum likelihood methods in the version of the PC-SAS software used. This risked overestimation of the odds ratios (Kleinbaum, Kupper, & Morgenstern, 1982). Separate regression models were performed for ADL and IADL limitations. The regression models compared HHC users with two control groups: OCS alone, and OCS and NCS combined (i.e., all nonHHC users). Backward stepwise elimination of individual, nonsignificant (p > 0.05) variables was used 642

The Gerontologist

Downloaded from http://gerontologist.oxfordjournals.org/ at University of York on April 16, 2015

tions. Both surveys found that bathing limitations, the most frequent ADL problem, were experienced by about 6% of the population, whereas eating limitations, the least frequent ADL problem, were reported by 1-2% of the population. The most frequent IADL limitation, shopping, was estimated in 7% of the SOA population and 11% of NLTCS population, compared with IADL limitations such as telephoning and light housework, which were reported by 2-5% of the population. Thus, assessing the association between HHC use and specific ADL or IADL limitations might be more informative than using the number of ADL or IADL limitations for planning home health care services. The second issue is accounting for other community-based services that might complement or substitute for home health care in meeting people's ADL and IADL needs. Only two of the above studies (Hughes et al., 1987; Weissert, Wan, & Livieratos, 1980) compared home health care users with users of other community-based health and social services for the elderly. Inclusion of such comparison groups is important in order to evaluate the association between functional limitations and HHC use within the context of other community-based health services for the elderly. The current study evaluated the associations between specific ADL and IADL limitations and home health care use. Specific limitations, rather than the number of limitations, were chosen because of the range in the prevalence of individual ADL and IADL limitations and in the different types of assistance that are needed for different limitations. A case-control design was used to select home health care users from the 1984 National Health Interview Survey Supplement on Aging (SOA). HHC users were matched on age and gender to two sets of controls: respondents who reported using other community services (OCS) and respondents who used no community services (NCS). This was done to determine whether the ADL and IADL limitations of HHC users were more similar to OCS users than to NCS users, as an indicator of substituting other community services for HHC to meet participants' ADL and IADL needs. Univariate analyses assessed whether individual limitations in ADLs and lADLs were associated with HHC use compared with OCS and NCS users, respectively. Multivariate analyses were performed to determine which individual ADL and IADL limitations were associated with HHC use, controlling for other covariables. The results would not only give more insight into how home health care meets the physical functioning needs of the elderly, but would also provide additional information on using ADL and IADL limitations to define eligibility criteria for home health care and other long-term care benefits.

to determine the most powerful and parsimonious model to predict HHC use (Kleinbaum, Kupper, & Morgenstern, 1982). The baseline model contained all ADL or IADL variables and all the demographic and health covariables. Although about 20% of the HHC users lacked income data, no adjustment was made in the regression models because the income variable was not statistically significant. Thus, the final regression models included all but 10 of the respondents. None of the analyses used weighted data because the main goal was to compare the functional limitation characteristics of HHC users and controls, rather than to produce national population estimates. PCSAS was used for all analyses.

Results

Vol.32, No. 5,1992

643

Discussion This study found, in univariate analyses, that HHC users were significantly more impaired in every ADL and IADL limitation, whether the comparison group was OCS or NCS users. This is consistent with other studies that found associations between the number of ADL and IADL limitations and HHC use (Berk & Bernstein, 1985; Branch et al., 1988; Evashwick et al., 1984; Soldo, 1985; U.S. Department of Health and Human Services, 1986). It also supports the results of other studies in which HHC use was more strongly associated with ADL and IADL limitations and perceived health status than with sociodemographic factors (Branch et al., 1988; Evashwick et al., 1984; Hughes, Cordray, & Spiker, 1984; McAuley & Arling, 1984; Soldo, 1985). Moreover, it supports the utility of ADL criteria in the long-term care eligibility models (Spector, 1991; Stone & Murtaugh, 1990): each of the five ADL limitations considered in those models — bathing, transferring, toileting, dressing, and eating — was independently associated with HHC use. The logistic regression analyses showed that HHC users were more likely than OCS and NCS users to be impaired in the following specific ADL and IADL limitations, while controlling for age, gender, and significant covariables: eating, going outside, dressing, bathing, and shopping, performing housework, and using the telephone. Separate regression models were performed for ADL and IADL limitations. This was done because many studies have separated ADL

Downloaded from http://gerontologist.oxfordjournals.org/ at University of York on April 16, 2015

Table 1 shows that the selection procedures resulted in similar age and gender distributions among HHC, OCS, and NCS users: 30% of all groups were male and 60% were over age 75. HHC users were significantly more likely than OCS users to live with others and to have incomes of $20,000 or more. When compared with NCS users, these relationships were reversed but did not reach statistical significance. HHC users and OCS users were significantly less likely than NCS users to be married. HHC users rated their health as poor seven times more than either OCS or NCS users. The most prevalent ADL limitations for the total sample were walking (37%), going outside (27%), and bathing (26.5%), and the most prevalent IADL limitations were heavy housework (43%) and shopping (30%), as derived from Table 2. These limitations were the most prevalent among HHC users as well. The proportion of HHC users who were limited in each ADL were 64% in walking, 55% in going outside, 54% in bathing, 40% in both dressing and transfer, 31 % in toileting, and 14% in eating. The proportion of HHC users who were limited in each IADL were 72% in heavy housework, 57% in shopping, 43% in both cooking and light housework, 29% in money management, and 23% in using the telephone. As shown in Table 2, every ADL and IADL limitation was significantly more prevalent in HHC users, whether compared with OCS or with NCS users. When comparing HHC with OCS users, eating was the ADL limitation most strongly associated with HHC use (odds ratio [OR] = 23.21), and transferring was the ADL limitation least associated with HHC use (OR = 5.17). Analyses comparing ADL limitations among HHC users with those among NCS users showed that bathing limitations were most strongly associated with HHC use (OR = 8.29) while transferring was least associated with HHC use (OR = 5.43). The odds ratios between ADL limitations and HHC use were generally similar, whether OCS users or NCS users were the comparison group. Eating limitations was the one exception, where the odds ratio for HHC versus OCS users was 23.21 and that for HHC versus NCS users was 7.62; this result might have

been due to the small number of respondents with those limitations. With regard to IADL limitations, shopping limitations were most strongly associated (OR = 8.78) and money management limitations were least strongly associated (OR = 5.73) with HHC users compared with OCS users. When HHC users were compared with NCS users, heavy housework was the IADL limitation most associated with HHC use (OR = 6.87), and telephone limitations were least associated with HHC use (OR = 4.07). Table 3 shows that the ADL limitations remaining significantly associated with HHC versus OCS users in the final regression model were eating, going outside, dressing, and bathing. Poor health, living with others, and being unmarried were significant covariables. In the final regression model comparing HHC users with OCS and NCS users combined, limitations in going outside, dressing, bathing, and poor health were significantly associated with HHC use. In the final logistic regression model of IADL limitations reported by HHC versus OCS users, shopping, light housework, heavy housework, poor health, and living with others remained significantly associated with HHC use (Table 4). In contrast, shopping, heavy housework, using the telephone, and poor health remained significant in the final regression model of IADL limitations associated with HHC use versus OCS and NCS use combined.

Table 1 . Demographic Characteristics Associated with Users of Home Health Care (HHC) Compared with Users of Other Community Services (OCS) and No Services (NCS)

Odds ratio OCS (95% C.I.)

Odds ratio NCS (95% C.I.)

OCS (n)

NCS (n)

Gender Male Female

127 291

127 291

127 291

Age 60-64 65-69 70-74 75 +

19 67 81 251

19 67 81 251

19 67 81 251

Education =5 12 years > 12 years

345 61

355 56

336 71

0.89 (0.60-1.32)

1.20 (0.82-1.74)

Living arrangement Alone With others

147 271

223 195

122 296

0.47 (0.36-0.63)

1.32 (0.98-1.76)

Marital status Married Unmarried

161 256

162 256

203 212

0.99 (0.75-1.31)

0.66 (0.50-0.87)

Income < $20,000 > $20,000

258 76

301 43

258 93

0.49 (0.32-0.73)

1.22 (0.86-1.73)

Perceived health Poor Good

179 239

37 378

40 374

7.63 (5.18-11.24)

6.99 (4.78-10.20)

Table 2. Functional Limitations Associated with Users of Home Health Care (HHC) Versus Users of Other Community Services (OCS) and No Services (NCS)

NCS (n)

Yes No

223 192

56 359

51 364

7.45 (5.40-10.26)

8.29 (5.98-11.49)

Dress

Yes No

165 250

28 387

31 383

9.12 (6.21-13.41)

8.15 (5.60-11.88)

Eat

Yes No

60 355

3 412

9 406

23.21 (10.17-52.99)

7.62 (4.10-14.19)

Toilet

Yes No

128 285

14 401

22 393

12.86 (8.00-20.69)

8.02 (5.24-12.28)

Walk

Yes No

264 151

99 316

96 319

5.58 (4.17-7.47)

5.81 (4.33-7.79)

Transfer

Yes No

165 250

47 368

45 370

5.17 (3.67-7.28)

5.43 (3.84-7.67)

Go outside

Yes No

226 188

48 367

63 352

9.19 (6.61-12.79)

6.71 (4.91-9.19)

Yes No

235 179

54 361

80 335

8.78 (6.36-12.11)

5.50 (4.07-7.43)

Cook

Yes No

178 236

37 378

49 366

7.71 (5.39-11.02)

5.63 (4.02-7.89)

Money

Yes No

118 2%

oo ro

VI OO

Odds ratio NCS (95% C.I.)

OCS (n)

OJ

Odds ratio OCS (95% C.I.)

HHC (n)

35 380

5.73 (3.80-8.64)

Use phone

Yes No

97 317

19 396

29 386

6.38 (4.01-10.15)

4.07 (2.68-6.19)

Light work

Yes No

179 233

34 379

54 360

8.56 (5.95-12.33)

5.12 (3.68-7.12)

Heavy work

Yes No

300 114

126 289

115 300

6.04 (4.51-8.07)

6.87 (5.12-9.20)

Variable ADL limitations Bathe

IADL limitations Shop

644

4.33 (2.94-6.37)

The Gerontologist

Downloaded from http://gerontologist.oxfordjournals.org/ at University of York on April 16, 2015

Variable

HHC (n)

Table 3. Multiple Logistic Regression Results: ADL Limitations Associated with Users of Home Health Care (HHC) Versus Users of Other Community Services (OCS) and No Community Services (NCS)

HHCvs .OCS Variables Eat Go outside Dress Bath Unmarried (vs. married) Poor health Lives with others (vs. alone)

Odds ratio

95% C.I.

4.09 3.17 1.94 1.64 1.67

1.18-14.18 2.00-5.03 1.11-3.39 1.01-2.65 1.04-2.70

2.76 2.27

1.96-3.85 1.42-3.60

HHCvs. OCS + NCS Odds ratio 95% C.I. 2.62 2.12 2.23

1.80-3.82 1.37-3.28 1.50-3.34

2.63

1.98-3.50

Table 4. Multiple Logistic Regression Results: IADL Limitations Associated with Users of Home Health Care (HHC) Versus Users of Other Community Services (OCS) and No Community Services (NCS)

HHCvs. OCS Variables Shop Heavy work Light work Use phone Poor health Lives with others (vs. alone)

HHCvs. OCS + NCS

Odds ratio

95% C.I.

Odds ratio

95% C.I.

2.68 2.29 1.84 — 2.54 1.97

1.67-4.32 1.55-3.39 1.08-3.16

2.45 2.67 — 1.75 2.54

1.72-3.47 1.92-3.71

1.81-3.56 1.41-2.75

1.13-2.71 1.90-3.38

limitations from IADL limitations (Branch et al., 1988; Beland, 1986; Evashwick et al., 1984; Hing & Bloom, 1990; Soldo, 1985), and federal proposals for reimbursement of home care services have focused on ADL limitations (Stone & Murtaugh, 1990). Future analyses could include both ADL and IADL limitations to evaluate whether the presence of lADLs changes the association between ADL limitations and HHC. Some of the ADL and IADL limitations that were most strongly associated with HHC use in univariate analyses (e.g., toileting and cooking limitations) were not statistically significant in the multivariate analyses. This might have been due to collinearity with other ADL or IADL limitations. About 60% of HHC users had two or more ADL or IADL limitations, compared with 20% of the OCS and NCS users. Specifically, about 10% of HHC, OCS, and NCS users had one ADL limitation. However, 12% of HHC users were limited in two ADLs, 9% were limited in three, four and five ADLs, respectively, 13% were limited in six, and 12% were limited in seven ADLs. Similarly, 18% of HHC users were limited in one IADL, 12% were limited in two, 9% in three, 16% in four, 8% in five, and 15% in six lADLs. Thus, multiple ADL or IADL limitations among HHC users, or small numbers of persons with individual limitations, might have influenced these results. These analyses corroborated results of other cross-sectional analyses of SOA respondents with functional limitations (Droge, Fredman, & Rabin, Vol. 32, No. 5,1992

645

Downloaded from http://gerontologist.oxfordjournals.org/ at University of York on April 16, 2015

1990). In those analyses, ADL limitations in dressing, bathing, going outside, and eating, and IADL limitations in shopping, light housework, using the telephone, and preparing meals were significantly associated with HHC use compared with OCS and NCS use. Regardless of the comparison group, HHC appeared to be used for specific functional limitations. The focus on specific ADL and IADL limitations and HHC use added more detail to the current literature on functional dependencies and HHC use. It quantified the associations between specific ADL and IADL limitations and HHC use. These results suggest that knowing the specific functional limitation gives more precise information on the likelihood of using HHC services than knowing the number of functional limitations, especially if a person is limited in only one activity. Using two comparison groups, OCS users and NCS users, resulted in slightly different associations between functional limitations and HHC use in the multivariate analyses. However, there was no consistent pattern in the differences of associations obtained using OCS users versus OCS and NCS users combined. Although the final models differed by the presence of one ADL or IADL limitation, the odds ratios in the two models were of comparable magnitude. Moreover, the contrasts between the univariate results using separate OCS and NCS comparison groups was not as marked as expected: if persons with functional limitations were more likely to use either HHC or other community services, then the odds ratios would be smaller using the OCS comparison group than the NCS comparison group. Yet, similar odds ratios were obtained with both comparison groups, which suggested that the OCS users were more similar to NCS users than to HHC users. Therefore, other community services did not appear to substitute for HHC in meeting respondents' ADL and IADL needs. A potential limitation of this study is that minimal information was available on when HHC use occurred. It could not be determined whether HHC preceded or followed the self-reported recent hospitalizations. Furthermore, there was no information about length of HHC use, which prevented assessing whether functional limitations were associated with short- or long-term utilization. Such information would have illuminated further the association between ADL and IADL limitations and HHC use. The HHC variable combined visiting nurse and home health aide services in order to increase the sample size, and because both of these services are used by homebound elderly persons with functional limitations. However, visiting nurse services might be associated with ADL limitations that require nursing care, while home health aide services might be associated with different ADL or IADL limitations. Using a heterogeneous variable precluded investigation into these potential differences. These analyses did not include the SOA items on cognitive impairment. In a prospective study of elderly community residents, cognitive impairment and functional limitations were the strongest predic-

tors of incident HHC use (Branch et al., 1988). Recent studies have evaluated the effect of combining cognitive impairment with ADL and IADL limitations to determine eligibility for long-term care benefits (Kane, Saslow, & Brundage, 1991; Spector, 1991). While one study found that adding information on cognitive impairment to ADL and IADL limitations had a large impact on determining the number of persons in the eligibility pool (Spector, 1991), another concluded that adding information on need for supervision to the ADL and IADL information would account for most of the eligible persons with cognitive impairment (Kane, Saslow, & Brundage, 1991). Clearly, future analyses on HHC use and functional limitations should evaluate the role of cognitive impairment. This study used a matched case-control design. This design was chosen because the small proportion of HHC users in the SOA dataset conformed to a "rare-disease" model. Matching HHC users to controls was an efficient way to select the subsample from the SOA dataset and eliminated the need to control for age and gender as potential confounders. However, due to selecting the sample on the basis of HHC use, the prevalence and odds ratio estimates will not be exactly the same as in other reports of the SOA data that were based on the total sample and adjusted for the sampling design. Furthermore, these results do not provide national populationbased estimates. The results of the multivariate analyses suggest that basing eligibility for home care reimbursement on specific ADL and IADL limitations might be considered as an alternative to basing eligibility on the number of ADL limitations. Of the five ADL limitations suggested for determining reimbursement eligibility (Stone & Murtaugh, 1990), dressing and bathing were consistently associated with HHC use. Determining which limitations to use for eligibility should be based on similar analyses within national datasets, such as the total SOA sample or the NLTCS, so that they are determined by more representative, population-based information than was the present study.

and would indicate the effect of HHC services in meeting specific ADL and IADL impairment needs.

References

646

The Gerontologist

Downloaded from http://gerontologist.oxfordjournals.org/ at University of York on April 16, 2015

Finally, these results suggest that analyzing individual ADL and IADL limitations could have merit in determining HHC service need and in predicting long-term care service utilization by the elderly. Such analyses would provide more detail than a count or summary score of ADL and IADL limitations

Beland, F. (1986). The clientele of comprehensive and traditional home care programs. The Gerontologist, 26, 382-388. Berk, M. L , & Bernstein, A. (1985). Use of home health services: Some findings from the National Medical Care Expenditure Survey. Home Health Care Services Quarterly, 6, 13-23. Branch, L. C , Wetle, T. T., Scherr, P. A., Cook, N. R., Evans, D.A., Hebert, L. E., Masland, E. N., Keough, M. E., & Taylor, J. O. (1988). A prospective study of incident comprehensive medical home care use among the elderly. American Journal of Public Health, 78, 255-259. Droge, J. A., Fredman, L , & Rabin, D. L. (1990). Specific ADUIADL limitations and other characteristics of home health care users. Manuscript submitted for publication. Evashwick, C , Rowe, C , Diehr, P., & Branch, L. (1984). Factors explaining the use of health care services by the elderly. Health Services Research, 19, 357-382. Fitti, J., & Kovar, M. G. (1987). The Supplement on Aging to the 1984 National Health Interview Survey. Vital and Health Statistics, 7(21). Hing, E., & Bloom, B. (1990). Long-term care for the functionally dependent elderly. National Center for Health Statistics. Vital and Health Statistics, 73(104). Hughes, S. L , Cordray, D. S., & Spiker, V. A. (1984). Evaluation of long-term home care program. Medical Care, 22, 460-475. Hughes, S. L , Manheim, L. M., Edelman, P. L , & Conrad, K. J. (1987). Impact of long-term home care on hospital and nursing home use and cost. Health Services Research, 22, 19-47. Kane, R. L, Saslow, M. C , & Brundage, T. (1991). Using ADLs to establish eligibility for long-term care among the cognitively impaired. The Gerontologist, 31, 60-66. Kleinbaum, D. C , Kupper, L. L , & Morgenstern, H. (1982). Epidemiologic research, principles and quantitative methods. Belmont CA: Lifetime Learning Publications. McAuley, W. J., & Arling, C. (1984). Use of in-home care by very old people. Journal of Health and Social Behavior, 25, 54-64. SAS Institute, Inc. (1988). SAS Procedures Guide. Cary, NC: Author. Short, P. F., & Leon, J. (1990). Use of home and community services by persons age 65 and older with functional difficulties. Agency for Health Care Policy and Research (AHCPR) Report # 90-23. (DHHS Publication No. (PHS) 90-3466). National Medical Expenditure Survey Research Findings 5, Agency for Health Care Policy and Research. Rockville, M D : Public Health Service. Soldo, B. (1985). In-home services for the dependent elderly, determinants of current use and implications for future demand. Research on Aging, 7, 281-304. Spector, W. D. (1991). Cognitive impairment and disruptive behaviors among community-based elderly persons: Implications for targeting long-term care. The Gerontologist, 31, 51-59. Stone, R. (1986). Aging in the eighties, age 65 years and over — use of community services. Preliminary data from the Supplement on Aging to the National Health Interview Survey: United States, January-June 1984. Advance Data from Vital and Health Statistics, 124. Stone, R. I., & Murtaugh, C. M. (1990). The elderly population with chronic functional disability: Implications for home care eligibility. The Gerontologist, 30, 491^96. U. S. Department of Health and Human Services. (1986). National Medical Care Utilization and Expenditure Survey. Health care visits with nurses by place of visit, United States, 1980. Series B, Descriptive Report #9. Rockville, M D : Public Health Service. Weissert, W. C , Wan, T. T. H., & Livieratos, B. B. (1980). Effects and costs of day care and homemaker services for the chronically ill: A randomized experiment. NCHSR Research Report Series. DHEW Publication (PHS) 79-3258. Washington, DC: U.S. Government Printing Office.

Functional limitations among home health care users in the National Health Interview Survey Supplement on Aging.

A case-control study compared home health care (HHC) users from the 1984 Supplement on Aging to users of other community services and of no community ...
779KB Sizes 0 Downloads 0 Views