Enhancing Quality of Life of Families Who Use Adult Day Services: Short- and Long-Term Effects of the Adult Day Services Plus Program

Transcription

1 The Gerontologist Vol. 46, No. 5, Copyright 2006 by The Gerontological Society of America Enhancing Quality of Life of Families Who Use Adult Day Services: Short- and Long-Term Effects of the Adult Day Services Plus Program Laura N. Gitlin, PhD, 1 Karen Reever, MSG, MPA, 2 Marie P. Dennis, PhD, 1 Esther Mathieu, MSW, LCSW, 2 and Walter W. Hauck, PhD 3 Purpose: This study examined the short- and long-term effects of Adult Day Services Plus (ADS Plus), a low-cost care management intervention designed to enhance family caregiver well-being, increase service utilization, and decrease nursing home placement of impaired older adults enrolled in adult day care. Design and Methods: We used a quasi-experimental design, to recruit 129 caregivers from three adult day centers. Two centers offered adult day services and ADS Plus (n=67); the third center (n=62) offered only routine adult day services. ADS Plus involved a staff social worker who provided care management and support to family caregivers through face-to-face and telephone contact. Education, counseling, and referral services were offered for 12 months. Results: At 3- month follow-up, ADS Plus participants (n = 106) reported less depression (p =.016), improved confidence managing behaviors (p =.013) and enhanced well-being (p =.001) compared with controls. Long-term effects analyses (n = 69) showed that, compared with controls, ADS Plus participants continued to report less depression (p =.005) and enhanced confidence managing behaviors (p =.007). Also, intervention participants used adult day The research reported in this article was supported in part by Grant #90CG2557 from the Administration on Aging. We thank Nancy Chernett, MA, program manager, of the Center for Applied Research on Aging and Health at Thomas Jefferson University for her project assistance. Address correspondence to Laura N. Gitlin, PhD, Center for Applied Research on Aging and Health, Thomas Jefferson University, 130 South 9th Street, Suite 513, Philadelphia, PA laura.gitlin@ jefferson.edu 1 Center for Applied Research on Aging and Health, Thomas Jefferson University, Philadelphia, PA. 2 Mid County Senior Services, Main Line Health System, Newtown Square, PA. 3 Division of Biostatistics, Department of Pharmacology and Experimental Therapeutics, Thomas Jefferson University, Philadelphia, PA. services on average 37 days more than controls (p =.003) and had fewer nursing home placements (n=8) than controls (n= 17). Implications: ADS Plus is cost efficient and is easy to implement within adult day centers. It affords important clinical and public health benefits, including enhanced caregiver well-being, greater adult day service use, and fewer nursing home placements. Key Words: Care management, Caregiver burden, Depression Families continue to be the greatest national resource for helping frail older adults live a high quality of life at home. Most impaired elders receive help from family members, who provide more than 80% of long-term-care services in the United States (Family Caregiver Alliance, 2001; U.S. General Accounting Office, 1994). Providing care to frail older adults can be stressful and result in a range of well-documented negative consequences, including clinical depression, health problems, social isolation, financial difficulties, and mortality (Schulz & Beach, 1999; Schulz, O Brien, Bookwala, & Fleissner, 1995). This is particularly the case for family caregivers of individuals with dementia, who experience greater distress in comparison to individuals caring for nondemented older adults, and who tend to be the most frequent users of adult day services (Ory, Hoffman, Yee, Tennstedt, & Schulz, 1999). Despite the proliferation of caregiver intervention research (see reviews by Schulz et al., 2002; Sorensen, Pinquart, & Duberstein, 2002), existing aging network or long-term-care services have yet to integrate evidence-based approaches to supporting families in their roles as long-term care providers. One service that potentially benefits both impaired persons and family members is adult day services (ADS). ADS, 630 The Gerontologist

2 a community-based program, is designed for adults living at home who need considerable supervision and hands-on assistance due to impairments from chronic disease or disability. Additionally, ADS provides respite opportunities for families, a service need of particular import to caregivers of individuals with dementia. Only a few studies have evaluated the effects of ADS for family caregivers, and findings have been mixed (Zarit, Stephens, Townsend, & Greene, 1998). Zarit and colleagues showed that dementia caregivers reported less overload and depression following three months of ADS use in comparison to caregivers who did not use these services. Other studies, however, showed small improvements in select areas (such as burden), but not in other areas of well-being (such as depression; Cox, 1997; Gottlieb & Johnson, 1995). Only one study evaluated the benefits of an add-on service to ADS that targeted family caregivers. Droes, Meiland, Schmitz, and van Tilburg (2004) found that families who used a combined service involving ADS for individuals with dementia plus systematic support to families had reduced behavior problems and less depression compared with families who used ADS only. In summary, studies of ADS tend to show high dropout rates in the first few months of use, continued underutilization of the service, and that those families who do utilize the service still report significant distress (Zarit & Leitsch, 2001). Furthermore, there is little evidence that ADS delays nursing home placement, a primary objective of the service. The purpose of the present study was to test the short- (3-month) and long-term (up to 12- months) effects of an innovative intervention, the Adult Day Services Plus (ADS Plus) program. Although the 3- month follow-up represented the main endpoint for this study, we also evaluated long-term effects on caregiver well-being, ADS use, and nursing home placement. ADS Plus integrates care management within usual adult day center services in order to address the specific concerns and needs of family caregivers of impaired adults who are enrolled in ADS. The intervention systematically helps family caregivers develop problem-solving and coping skills, improve social and instrumental support, and enhance perceived competence in managing the difficult behaviors of their impaired elder. It utilizes existing ADS staff and resources, thus requiring a minimal increase in staff social work time such that there is nominal financial and administrative burden. Although most ADS centers offer occasional support groups or educational workshops for families, contact with families is typically not scheduled, systematic, or purposeful, nor are structured care management services offered. ADS Plus systematically targeted the multiple needs and concerns of caregivers and involved brief, structured, ongoing caregiver contact for a period of 12 months. We used a two-group quasi-experimental design for this study. We hypothesized that, compared with caregivers whose relatives were enrolled in ADS only, caregivers whose relatives were enrolled in ADS Plus (ADS for the relative plus family caregiver care management) would report less depression and burden, report greater confidence in managing troublesome behaviors, and engage in more preventive health behaviors. We also hypothesized that over the course of 12 months, intervention participants would use ADS for more days and would have a lower rate of nursing home placement compared with families in the control group. Methods Sample and Recruitment We conducted the study at three adult day centers operated by Main Line Health, a large health system in the western suburbs of Philadelphia. Approximately 350 families use Main Line Health ADS annually; 200 at any one point in time, with 150 new enrollees annually. Families using the Main Line Health ADS centers are from diverse ethnic and socioeconomic backgrounds and care for persons with a wide range of diagnoses. Dementia is the primary diagnosis for the majority (58%) of enrollees, with many individuals with other primary diagnoses (e.g., stroke) also suffering from dementia. Two of the three Main Line Health ADS centers served as intervention sites, in which eligible study participants received the usual day services and those offered through the ADS Plus Program. A third center served as the control site, in which eligible study participants received only the usual ADS. We chose this center to serve as the control site because its census was comparable to that of the other two centers. Staff at each ADS center consisted of a center director, family service provider, activity specialist, nurse, and program assistants, with a direct staff elder ratio of 1:5. Each of the three centers offered the same programming for impaired elders. Between 2002 and 2004, we recruited from the three sites families who self-identified as the primary caregiver and who had enrolled their relative in ADS (but for whom service had not yet begun). We defined the primary caregiver as the family member who selfidentified as the primary provider of care at home and who participated in the ADS admission process. There were no additional criteria for study participation. A total of 140 families who had enrolled their impaired family member in one of the adult day centers met eligibility criteria and were approached for study participation. Of these 140 families, 129 (92%) agreed to participate and enrolled (n = 67in ADS Plus, n = 62 in usual care) in the study. We approached families for study enrollment consecutively at each of the sites as they entered ADS. Those family caregivers who chose not to participate in the Vol. 46, No. 5,

3 study simply received the usual care for their impaired family member (e.g., customary ADS). Of the 129 caregivers enrolled in the study at baseline, 106 were available at the 3-month followup assessment (18% attrition), 74 were available at the 6-month assessment (43% attrition), and 58 (55% attrition) were available at the final 12-month assessment. Attrition by 12 months was due largely to nursing home or assisted living placement (n = 25); illness (n = 8) or death (n = 8) of the impaired person; care recipient refusal to attend ADS (n = 11); or other reasons (n = 19), such as relocation. Procedures At each site, the site social worker approached eligible family caregivers and described the study procedures specific to the site (e.g., interview and intervention at the intervention sites; interview only at the control site), obtained signed informed consent by using a form approved by the Main Line Health Care System, and conducted a baseline interview. Trained interviewers who remained masked to study hypotheses and group assignment subsequently interviewed study participants face-to-face on site or by telephone at 3, 6, and 12 months. The interview at each testing occasion consisted of questions pertaining to the health and well-being of both the caregiver and relative enrolled in ADS and took approximately 45 minutes. Study participants at the control group site received usual care. This involved use of ADS by their impaired family member and access to staff to obtain information as needed as is customary in usual care. Study participants at the two intervention sites received usual care (ADS) and the ADS Plus Program, which was delivered by the family service director (a licensed social worker) at each intervention site. ADS Plus Program ADS Plus, as described elsewhere (Reever, Mathieu, Dennis, & Gitlin, 2004), was based on Pearlin, Mullan, Semple, and Skaff s (1990) stress process model. The program targets the primary stressors of caregiving: behavior problems of the care recipient; and the physical, mental, and social health of family caregivers. Building on previous intervention studies that showed the multiple benefits of care management on caregiver well-being (Mittelman, Ferris, Shulman, Steinberg, & Levin, 1996; Mittelman, Ferris, & Steinberg, 1995), the approach was used to support positive intrapsychic factors (such as the perception of competence and rewards of caregiving) as well as introduce positive coping strategies (such as problem solving) in order to mediate objective stressors of caregiving. Families assigned to intervention received usual care consisting of ADS use for their impaired relative and ADS Plus. ADS Plus consisted of care management services that targeted the family caregiver s own concerns and needs. Family caregivers initially met face to face with the site service director in order to (a) identify areas of concern and needs; (b) develop a care plan to minimize identified areas of difficulty; and (c) implement an agreed-upon care plan that involved four components: counseling, education, referral, and periodic supportive contact with the site family service director. Following the initial meeting, the family service director met with family caregivers either face to face when their relative was brought to the center, or by telephone. Additionally, the family service director provided targeted education materials through mailings. The purpose of each follow-up contact was to provide emotional support, counseling, education, and referral targeted to the specific identified concerns. Table 1 presents the specific treatment objectives for targeted areas of caregiver concern and the associated strategies provided. Family caregivers received an average of 1 hour per month of service contact. Of the four treatment components (counseling, education, referral, and follow-up), 63% of individuals received counseling, 60% received education, 39% were referred for additional professional services, and all caregivers (100%) received follow-up support. Outcome Measures We used five primary outcome measures in order to assess short- and long-term effectiveness: caregiver depression, burden, self-efficacy in managing troublesome behaviors, overall perceived well-being, and health behaviors. We also considered number of days in ADS and nursing home placement as potential long-term outcomes. Caregiver Depression. We used the short-form, 10-item version of the Center for Epidemiologic Studies Depression scale (Andresen, Malmgren, Carter, & Patrick, 1994) in order to measure caregiver depression. The CES-D has been used widely in intervention studies with family caregivers. It provides cutoff scores (e.g., 8 or greater for the short form) that aid in identifying individuals at risk for clinical depression. Each item assesses the frequency of occurrence of a depressive symptom within the past week, with response options ranging from 0 (rarely or none of the time) to 3 (most or all of the time). Scores range from 0 to 30 and represent caregivers summed responses across the 10 items, with higher scores indicating the presence of more depressive symptoms (Cronbach s alpha =.80 for this sample). Caregiver Burden. In order to assess the extent to which caregivers felt stressed in their caregiving role, we used 12 items from the short-form of the Zarit Burden Interview (Be dard et al., 2001). Response options range from 1 (never) to 5 (nearly 632 The Gerontologist

4 Table 1. Overview of Caregiver (CG) Concerns Addressed in ADS Plus and Specific Intervention Strategies Concern Goals Objectives Examples of Strategies Provided Physical health Increase perceived health Decrease obstacles to good health Increase time for engagement in healthy behaviors Decrease CG sick time Mental health Increase CG confidence managing behaviors Decrease CG depression Increase CG sense of control Increase CG gains Learn about CR illness and use as a basis for caregiving strategies Manage stress Address isolation Identify and engage in Pleasant events Refer to exercise and nutrition programs Refer to workshops on home care skills and health Provide written/verbal information on health topics Teach time management for health behaviors Educate about CR illness Teach CG stress management skills Provide regular CG support via regular phone or meeting Provide counseling or refer CG for counseling Social health Decrease CG strain Know need for support Provide support or refer to CG Have the confidence and skills to get support support group Provide workshops/information to Expand CG support network plan for future Engage CG in networking opportunities Facilitate family meetings always). Scores range from 1 to 75 and represent the caregivers summed responses across the 15 items (Cronbach s alpha =.88 for this sample). Memory and Problem Behaviors and Self-Efficacy. Using the Zarit and Zarit (1990) 24-item Memory and Problem Behaviors scale, we asked caregivers to indicate the frequency of occurrence of each behavior as well as their level of confidence in handling memory loss, depressive and disruptive behaviors. For each of the 24 behaviors, caregivers responded on a 5-point scale (0 = not at all confident to 4= very confident). Scores ranged from 0 to 96 and represented the caregivers average confidence level managing behaviors. It was not feasible to calculate Cronbach s alpha for the self-efficacy scale because the number of rated behavior items varied among caregivers. Perceived Change in Well-Being. In order to assess perceived change in well-being, we used the Perceived Change Index (Gitlin et al., 2003). This brief, easily administered, and valid self-report measure assesses self-perceptions of temporal change in well-being or the extent to which caregivers appraise whether a range of affective and caregiving concerns have stayed the same, improved, or deteriorated over the past month (Gitlin, Winter, Dennis, & Hauck, in press). The items reflect factors that may place caregivers at increased risk for negative affective consequences and cumulative health risks. We asked caregivers to rate 13 items (e.g., ability to manage day to day) along a 5-point scale (1 = became much worse to 5 = improved a lot over the past month). This approach to evaluating areas of improvement or decline is commonly used in pharmacological research and has only recently begun to be used in caregiver research (REACH II Investigators, 2006). We reverse coded negatively valenced items (e.g., feeling overwhelmed) and averaged ratings across the 13 items. Higher mean scores indicated more positive change in perceptions of well-being (Cronbach s alpha =.91 for this sample). Time for Healthy Behaviors Checklist. We used six items from the National Institutes of Health multisite Resources for Enhancing Alzheimer s Caregivers Health (REACH II) initiative. Items were In the past 6 months, did you miss one doctor s appointment? forget to take medication? slow down and rest when sick? have enough time to see a doctor when necessary? have enough time to exercise? overall have time to rest? Responses to all six items were dichotomous (0 = no, 1 = yes). We reverse coded responses and then summed them to form a 6-point scale, where 0 indicated engagement in highrisk health behaviors and 6 indicated time for engaging in healthier behaviors (a =.68). Use of ADS and Nursing Home Placement. We calculated a single-item index representing the total number of days that a care recipient had attended ADS from the date of the caregiver s baseline interview. For those study participants whose care recipient stopped attending ADS, each site director ascertained whether it was due to nursing home placement or another reason. Either the family had informed the center of the placement or the director conducted a follow-up telephone call to determine care recipient status. Statistical Analyses We used chi-square and Wilcoxon rank-sum tests, as appropriate, in order to determine if the usual Vol. 46, No. 5,

5 care and intervention caregivers and their relatives differed significantly at baseline in demographic characteristics; relative s level of physical and cognitive functioning and memory and problem behaviors; caregiver depression, burden, self-efficacy and healthy behaviors; as well as attrition and ADS attendance. Next, we examined the group means, variances, and actual range of values for all outcomes and participant attendance. We also examined the distribution of the residuals from the regressions for each outcome measure in order to identify violations of the normality assumption; we found no transformations to be necessary. Based on the extant literature, we identified several covariates. We conducted stepwise (stepdown) regressions for each outcome in order to identify covariates consistent across outcomes as well as potential demographic factor interactions that might also covary with the outcome measures. This resulted in one interaction that was common to all of the stepwise regressions: care recipient gender by caregiver education. We subsequently included in the regression analyses other covariates (caregiver race, gender, and education; and care recipient race and gender) that were consistent across outcomes, as well as care recipient illness characteristics (number of activities of daily living [ADLs] and instrumental ADLs [IADLs] with which the care recipient needed assistance, plus number of memory and problem behaviors). Adjusting for baseline values, we performed analysis of covariance procedures on the short-term, 3-month outcomes (the primary study endpoint) in order to examine main treatment effects. Because this was a quasi-experimental design and we found control and intervention site differences in caregiver and care recipient characteristics, we evaluated whether there were interaction effects of treatment by demographic factors for each of the outcome measures. Secondarily, in order to examine long-term treatment effects, we included in the analyses all participants with a baseline and 6- or 12-month interview (n = 69). Data across all four assessment time points were available for 19 usual care and 36 intervention caregivers. Five usual care and 8 intervention caregivers had both baseline and 6-month data, whereas only 1 caregiver with 12-month data did not have data for the 6-month assessment. For each of the five outcomes, we performed linear mixed-model analyses (adjusting for baseline values and the above-mentioned covariates) using restricted maximum likelihood parameter estimation and an unstructured within-subject covariance matrix. The p values we report were not corrected for multiple endpoints in these analyses. We calculated effect size for each treatment outcome by dividing the adjusted mean difference by the pooled standard deviation to derive Cohen s d (Cohen, 1987). Finally, we examined differences between the intervention and control groups in days to discontinuation from ADS by using a Kaplan Meier survival analysis procedure and compared curves by using a Cox proportional hazards model adjusting for the same set of demographic and care recipient illness characteristics. We performed all analyses using SPSS, version 13.0 (SPSS, Chicago, IL). Results Caregivers across all three centers (N = 129) were on average 62.6 years of age, were Caucasian (79.8%), were women (75.9%), had a high school or better education (95.1%), and had been caregiving for approximately 5.6 years. Most were daughters (40.3%) or wives (29.5%) of the impaired elders and rated their own physical health as good or very good (M = 3.6, SD = 1.02). Similarly, care recipients were on average 80.4 years of age, were Caucasian (81.3%), were men (59.7%), had a high school or better education (79.2%), needed assistance with 4 basic ADLs and 7 IADLs, and had low cognitive status as measured by the Mini-Mental State Examination (Folstein, Folstein, & McHugh, 1975). As shown in Table 2, at the 3- and 12-month follow-up, control and intervention group caregivers did not differ significantly in age, gender, education, relationship to the care recipient, or number of years of caregiving. However, we observed a statistically significant racial difference: More Whites were in usual care than in the intervention, and more African Americans were in the intervention than in usual care. Similarly, we observed a statistically significant racial difference between care recipients in usual care and in ADS Plus: A greater proportion of relatives of control group families were White. Also, relatives of family caregivers in usual care and in intervention differed significantly in the number of IADLs with which they needed help (clients in usual care required more assistance than those in intervention). However, we observed no statistically significant differences between relatives of families in intervention and control groups with regard to age, gender, primary diagnoses, help with ADLs, the number of reported memory and problem behaviors, or cognitive status. Table 3 presents means and standard deviations by treatment group. Upon entry into the study, caregivers tended to be moderately burdened and at risk for depression. On average, they engaged in nearly four healthful self-care behaviors. At 3-month follow-up, they had not perceived much change in their feelings of well-being over the past month and had made use of ADS approximately 26 days (SD=14.1) of the 90-day intervention period (not shown in Table 3). Short-Term (3-Month) Effects of ADS Plus (n = 106) As hypothesized, there was a main effect for caregiver depression such that ADS Plus caregivers (M = 7.2, SD = 5.0) experienced a statistically 634 The Gerontologist

6 Table 2. Background Characteristics of the 3-Month and Long-Term Samples 3-Month Sample (N ¼ 106) Long-Term Sample (N ¼ 69) Characteristic ADS Usual Care (n ¼ 49) ADS Plus (n ¼ 57) v 2 Z p ADS Usual Care (n ¼ 25) ADS Plus (n ¼ 44) v 2 Z p Caregiver Age, M (SD) 66.0 (13.2) 62.1 (13.8) (14.7) 60.2 (12.9) Race (%) White African American Other Gender (%) Male Female Education (%) , high school high school high school Relation to care recipient (%) Spouse Child Other Years caregiving, M (SD) 6.7 (12.0) 5.3 (5.6) (3.7) 5.1 (5.5) Care recipient Age, M (SD) 80.7 (7.5) 80.9 (8.6) (8.2) 81.2 (9.5) Race (%) White African American Other Gender (%) Male Female Primary diagnoses (%) Dementia Heart disease Cerebrovascular accident Arthritis Diabetes Other ADL requiring help, M (SD) 3.1 (2.7) 3.6 (2.5) (2.5) 3.7 (2.5) IADL requiring help, M (SD) 8.2 (1.5) 6.4 (3.0) (0.2) 0.8 (0.3) Memory and Behavior Problems 9.9 (5.2) 11.0 (5.1) (4.7) 11.4 (4.7) Checklist, M (SD) MMSE, M (SD) 15.8 (10.2) 13.5 (10.3) (10.7) 13.4 (10.5) Notes: ADS = adult day services; ADL = activity of daily living; IADL = instrumental ADL; MMSE = Mini-Mental State Examination; SD = standard deviation. Bold p-values indicate statistical significance. significant decrease in depression compared with usual care participants (M = 8.6, SD = 6.6; d =.43, p =.016). We also found a statistically significant difference between usual care and intervention caregivers at 3-month follow-up in memory and problem behavior self-efficacy (p =.013, d =.56). That is, caregivers enrolled in the ADS Plus Program reported a gain in confidence (M = 3.14, SD = 0.83) from baseline, whereas those in usual care reported a decrease in confidence (M = 2.88, SD = 0.98). Similarly, caregivers in ADS Plus (M = 3.29, SD = 0.62) differed significantly from caregivers in usual care (M = 2.90, SD = 0.46) in how they perceived changes in their overall well-being (p =.001, d =.80). Caregivers assigned to ADS Plus tended to report a change for the better, whereas those assigned to usual care tended to report feeling worse. At 3 months, there were no statistically significant differences in healthy behaviors between usual care and ADS Plus participants. There was a marginal but nonsignificant treatment effect for caregiver burden at the 3-month follow-up. Although caregivers in the Vol. 46, No. 5,

7 Table 3. Comparison of Usual Care and ADS Plus Caregivers on Outcome (N = 106) Baseline 3-Month Follow-Up Usual ADS Care (n ¼ 49) ADS Plus (n ¼ 57) Usual ADS Care (n ¼ 49) ADS Plus (n ¼ 57) Difference of Outcome Measure M (SD) M (SD) Adjusted Means a 95% CI p CES-D 9.33 (6.6) 9.89 (4.7) 8.61 (6.6) 7.24 (5.0) , Burden (9.1) (10.3) (10.5) (9.8) , Memory and behavior problem self-efficacy b 3.42 (0.7) 2.87 (0.9) 2.88 (1.0) 3.14 (0.8) , Time for healthy behaviors 4.23 (1.8) 3.60 (1.6) 4.20 (1.7) 3.60 (1.8) , Perceived change in well-being 2.90 (0.5) 3.29 (0.6) , Notes: ADS = adult day services; CI = confidence interval; CES-D = Center for Epidemiologic Studies Depression scale; SD = standard deviation. Bold p-values indicate statistical significance. a Adjusted for baseline, care recipient gender, number of activities of daily living and instrumental activity of daily living for which care recipient needed assistance, care recipient memory and behavior problems, caregiver race and education, and care recipient gender by caregiver education interactions. b N = 103. intervention experienced a decrease in self-reported burden from the baseline assessment, caregivers in usual care also reported a small decline in burden (Table 3). We did not have specific hypotheses regarding interaction effects, but rather conducted exploratory analyses of covariance in order to evaluate whether treatment effects were in part due to site differences in caregiver and care recipient characteristics. However, these analyses did not reveal any statistically significant interactions between treatment condition and caregiver or care recipient demographic factors. Long-Term Effects of ADS Plus A total of 69 family caregivers had sufficient data for inclusion in the long-term effects analysis (n =25 controls, n = 44 intervention). There were no large or statistically significant differences between the 69 caregivers included in the analyses and the 60 caregivers who were not included in the long-term analyses with regard to age, education, relationship to the care recipient, years of caregiving, care recipient s age, primary diagnosis, amount of assistance needed with IADLs and ADLs, or care recipient s cognitive status. There were, however, differences by group assignment (v 2 = 8.31, p =.004), caregiver race (v 2 = 8.57, p =.014), care recipient race (v 2 = 7.61, p =.022), and care recipient gender (v 2 = 3.48, p =.062). More specifically, 61.7% of caregivers in the control group compared with 38.3% of caregivers in the intervention group lacked sufficient data to be included in the long-term effects analysis. Likewise, 90.0% of White caregivers and 10.0% of non-white caregivers also lacked sufficient data for inclusion in the long-term effects analysis. Table 4 presents the estimates of treatment group differences for each separate mixed-model analysis. These estimates represent the impact of the predictors on the outcome of interest, averaged over the 6- and 12-month assessments after the effects of the other predictors in the model have been controlled. In the depression analysis, for example, the estimate of 3.52 for treatment condition indicates that, across both assessments, intervention caregivers had an average depression score 3.52 points lower than the average score for the control group participants; as expected, this difference was statistically significant (p =.005). Baseline depression (p =.001) and memory and problem behavior (p =.040) also proved to be statistically significant in the model. Similarly, for memory and problem behavior selfefficacy, the estimate of a.51 difference for the effects of the intervention indicates that caregivers in ADS Plus had an average self-efficacy score that was.51 points higher than the average score for the control caregivers across the 6- and 12-month assessments; consistent with our prediction, this difference was statistically significant (p =.007). The only other predictors in this model that were statistically significant were baseline self-efficacy (p =.002) and number of memory and problem behaviors (p =.003). The estimates of the long-term effects of the ADS Plus intervention on caregiver burden, perceived change in well-being, and time for healthy behaviors, although in the expected direction, were not statistically significant when averaged across the 6- and 12-month time points. This may have been due to the sample size. Attendance and Discontinuation Our analyses also revealed a statistically significant difference in attrition between the usual care and ADS Plus groups (v 2 = 20.6, p =.038), such that of the 129 caregivers enrolled in the study, 43 (69%) in usual care were lost to follow-up compared with 636 The Gerontologist

8 Table 4. Linear Mixed Model of Treatment Effects for Five Separate Caregiver Outcomes (N = 69) Outcome Estimate a p Confidence Interval Depression , 1.11 Burden , 1.41 Memory and behavior , 0.87 problem self-efficacy Perceived change in well-being , 0.32 Time for healthy behaviors , 0.99 Note: Bold p-values indicate statistical significance. a Adjusted for baseline, care recipient gender, number of activities of daily living and instrumental activities of daily living for which care recipient needed assistance, care recipient memory and behavior problems, caregiver race and education, and care recipient gender by caregiver education interactions. 28 (42%) in ADS Plus. The difference between the two treatment groups was most apparent with respect to nursing home placement: 17 (27%) control care recipients had been placed in a long-term care nursing facility, whereas only 8 (12%) ADS Plus care recipients had been placed in a facility. There was also a statistically significant difference between treatment groups in the total number of days of attendance (Z = 3.35, p =.001), with intervention caregivers utilizing ADS more days than did usual care caregivers (Ms = 92.3 days vs 55.4 days, respectively). As shown in Figure 1, usual care participants discontinued earlier and at a higher rate (70% at one year) than ADS Plus participants (41% discontinuation at one year; log rank = 12.6, p, 0.001). A Cox proportional hazard analysis showed that the main effect for treatment condition remained statistically significant after controlling for caregiver race and education, patient gender, illness characteristics, and a patient gender by caregiver education interaction (Wald chi square = 8.61, p =.004). Discussion This evaluation of the ADS Plus Program provides evidence that augmenting ADS with a care management service that systematically targets family caregivers and their concerns results in immediate and long-term clinically significant quality-of-life improvements for both the caregiver and impaired elder. Immediate benefits for caregivers included a reduction in depressive symptoms, increased confidence in managing troublesome behaviors, and enhanced overall perceived well-being. Thus, the care management service had a direct effect for caregivers within a relatively short time frame (3 months). This is in contrast to other caregiver intervention studies that show that improvement occurs gradually and over an extended period of time (Mittelman et al., 1995). Reducing depressive symptoms may help caregivers provide more effective and longer-term care to their impaired relative and may also minimize the caregivers own mortality risk (Schulz & Beach, 1999). Troublesome behaviors such as physical and Figure 1. Kaplan-Meier Analysis of Days to Discontinuation by Treatment Group. Vol. 46, No. 5,

9 verbal aggressiveness, agitation, or incontinence are often triggers for nursing home placement. Thus, enhancing caregivers confidence in managing these complex and difficult behaviors may support their ability to keep their relative at home. Over time, family caregivers derived additional benefits from ADS Plus, including continued reductions in depressive symptoms. There was a trend towards decreased burden, but this was not statistically significant. The finding of continued benefit in reducing depressive symptoms is in contrast to the findings of other studies, which have shown that ADS use alone may reduce caregiver burden but not depression (Cox, 1997; Gottlieb et al., 1995). In this study, families enrolled in ADS Plus scored on average lower than the clinical cut-off score of 8 for depressive symptoms, whereas families in usual care scored above the cut-off at each testing occasion. We also found that, over time, families enrolled in ADS Plus utilized more days of ADS than families who received the usual care. This may have two important benefits: It affords older people who are impaired increased exposure to an enriched care setting and improved quality of life, and it provides caregivers with needed respite and other potential forms of help. Given the consistent finding that family caregivers tend to underutilize helpful services especially those family members caring for persons with dementia (Brodaty, Thomson, Thompson, & Fine, 2005; Lawton, Brody, & Saperstine, 1989) increased day care use is an important treatment outcome. Furthermore, there were fewer nursing home placements for families in ADS Plus than for those receiving ADS only, with a little more than double the number of placements among controls than among intervention group participants. All of these represent important public health outcomes with benefits extending to families, their relatives, and to society at large (in terms of reduced costs associated with nursing home placement). ADS Plus is a relatively low-cost intervention that requires a minimal expenditure of time from center staff for its delivery. It is easily integrated into the usual care provided at adult day centers by extending the role of existing staff, usually a staff social worker. Also, it presents a minimal burden of participation for caregivers because treatment contact occurs when the impaired relative is brought to the center or via telephone and mail. In this study, all families received the minimum intervention exposure, which included initial problem identification and care planning and follow-up contact with the service director. Service directors spent an average of one hour per month per caregiver in ADS Plus, and most of this contact occurred via telephone (Reever et al., 2004). We found that families in ADS Plus used adult day services for a total of 36.9 more days (or more than one month more) than did families in usual ADS, resulting in increased revenue for the two intervention centers. Thus, revenue generated from the additional days of attendance that result from this program may help defray the costs associated with providing care management. Moreover, as was noted, the most frequent method of contact with families in ADS Plus was by telephone, which is a cost-efficient approach to delivering this intervention (Reever et al., 2004). Thus, the modest investment for implementing the care management intervention may be offset by the additional revenue brought in as a result of increased attendance. Several limitations to the present study temper these findings. The primary limitation concerns the design of the study. We did not randomly assign study participants to either usual care or intervention, and there were site differences with regard to caregiver and care recipient characteristics. Also, due to the lack of participant randomization, we were unable to control for potential confounders such as the quality and number of staff family interactions that naturally occurred and may have differed across sites. Nevertheless, we did adjust for group differences in all analyses, and an evaluation of treatment interaction by demographic factors revealed no differential effects of the intervention, thus strengthening support for our findings. Furthermore, families were not able to choose treatment group (intervention or control), but rather were enrolled consecutively into the study at a particular site that had been a priori designated as either intervention or control. This does provide some assurance of the strength of the main treatment effects. Moreover, the use of a comparison control group represents a design improvement over previous efforts to evaluate ADS treatment effects (Zarit, 1990) while retaining the ecological validity of the study. Another limitation is that, given the homogeneity of our sample, it is unclear whether our findings can be generalized to other centers with more racial or ethnic diversity and in different geographic areas. Finally, missing data appears to be nonrandom, in that there is a substantial and differential dropout rate between groups for the long-term analyses. However, in order to address this we compared individuals who stayed in the study to those who dropped out at their 3-month endpoints and found no large or statistically significant differences. Thus, this suggests that (at least through 3 months, the primary endpoint of this study) there was random missing data. Despite these limitations, the findings from this study are promising. The effect sizes for treatment outcomes in this study ranged from small to moderate and appear more favorable than those reported by other caregiver intervention studies (Sorensen et al., 2002). The study clearly demonstrates the potential benefits of targeting both the older person using ADS as well as the family caregiver in assessment and treatment, in order to maximize benefits for each in service programs that traditionally target only the older person. More 638 The Gerontologist

10 research is required to (a) substantiate these findings in other ADS centers, (b) examine the various components of the intervention (e.g., education, referral) in order to determine the contribution of each element to overall improvement, and (c) conduct formal cost analyses. In sum, this evaluation of ADS Plus demonstrates that systematically combining caregiver care management with ADS is not only feasible, but results in clinically significant public health benefits to family caregivers and their impaired relatives. ADS Plus has social validity in that families overwhelmingly evaluated the program as supportive and invaluable, offering, as one respondent stated, a critical safety net. This augmentation approach may also serve as a model for integrating caregiver assessment and treatment into other long-term-care service programs that traditionally target only the impaired older adult. Finally, although an important objective of ADS is to help families delay nursing home placement, this study suggests that without systematically targeting the concerns of family caregivers, this objective may not be fully achieved. References Andresen, E. M., Malmgren, J. A., Carter, W. B., & Patrick, D. L. (1994). Screening for depression in well older adults: Evaluation of a short form of the CES-D. American Journal of Preventive Medicine, 10(2), Be dard, M., Molloy, D. W., Squire, L., Dubois, S., Lever, J., & O Donnell, M. (2001). The Zarit burden interview: A new short version and screening version. The Gerontologist, 41, Brodaty, H., Thomson, C., Thompson, C., & Fine, M. (2005). Why caregivers of people with dementia and memory loss don t use services. International Journal of Geriatric Psychiatry, 20, Cohen, J. (1987). Statistical power analyses for the behavioral sciences (Rev. ed.). Hillsdale, NJ: Erlbaum. Cox, C. (1997). Findings from a statewide program of respite care: A comparison of service users, stoppers, and nonusers. The Gerontologist, 37, Droes, M., Meiland, F., Schmitz, M., & van Tilburg, W. (2004). Effect of combined support for people with dementia and carers versus regular day care on behaviour and mood of persons with dementia: Results from a multi-centre implementation study. International Journal of Geriatric Psychiatry, 19, Family Caregiver Alliance. (2001). Fact sheet: Selected caregiver statistics. Retrieved August 21, 2005, from content_node.jsp?nodeid=439 Folstein, M. F., Folstein, S. E., & McHugh, P. R. (1975). Mini-mental state : A practical method for grading the cognitive state of patients for the clinician. Journal of Psychiatric Research, 12, Gitlin, L. N., Winter, L., Corcoran, M., Dennis, M., Schinfeld, S., & Hauck, W. (2003). Effects of the home environmental skill-building program on the caregiver care recipient dyad: Six-month outcomes from the Philadelphia REACH initiative. The Gerontologist, 43, Gitlin, L. N., Winter, L., Dennis, M. P., & Hauck, W. (in press). Assessing perceived change in well-being of family caregivers : Psychometric properties of the Perceived Change Index (PCI) and response patterns Manuscript submitted for publication. Gottlieb, B. H., & Johnson, J. (1995). Impact of day care programs on family caregivers of persons with dementia. Guelph, Ontario, Canada: Gerontology Research Center, University of Guelph. Lawton, M. P., Brody, E. M., & Saperstein, A. R. (1989). A controlled study of respite services for caregivers of Alzheimer s patients. The Gerontologist, 29, Mittelman, M. S., Ferris, S. H., Shulman, E., Steinberg, G., & Levin, B. (1996). A family intervention to delay nursing home placements of patients with Alzheimer s disease. A randomized controlled trial. Journal of the American Medical Association, 276, Mittelman, M. S., Ferris, S. H., & Steinberg, G. (1995). A comprehensive support program: Effect on depression in spouse-caregivers of AD patients. The Gerontologist, 35, Ory, M. G., Hoffman, R. R., Yee, J. L., Tennstedt, S., & Schulz, R. (1999). Prevalence and impact of caregiving: A detailed comparison between dementia and nondementia caregivers. The Gerontologist, 39, Pearlin, L. I., Mullan, J. T., Semple, S. J., & Skaff, M. M. (1990). Caregiving and the stress process: An overview of concepts and their measures. The Gerontologist, 30, REACH II Investigators. (2006). Enhancing the quality of life of Hispanic/ Latino, Black/African American, and White/Caucasian dementia caregivers: The REACH II randomized controlled trial. Manuscript under review. Reever, K. E., Mathieu, E., Dennis, M. P., & Gitlin, L. N. (2004). Adult day services plus: Augmenting adult day centers with systematic care management for family caregivers. Alzheimer s Care Quarterly, 5, Schulz, R., & Beach, S. (1999). Caregiving as a risk factor for mortality: The Caregiver Health Effects study. Journal of the American Medical Association, 282, Schulz, R., O Brien, A., Czaja, S., Ory, M., Norris, R., Martire, L. M., et al. (2002). Dementia caregiver intervention research: In search of clinical significance. The Gerontologist, 42, Schulz, R., O Brien, A. T., Bookwala, J., & Fleissner, K. (1995). Psychiatric and physical morbidity effects of dementia caregiving: Prevalence, correlates, and causes. The Gerontologist, 35, Sorensen, S., Pinquart, M., & Duberstein, P. (2002). How effective are interventions with caregivers? An updated meta-analysis. The Gerontologist, 42, U.S. General Accounting Office. (1994). Long-term care: Diverse, growing population includes millions of Americans of all ages (GAO/HEHS Report No ). Washington, DC: Author. Zarit, S. H. (1990). Interventions with frail elders and their families: Are they effective and why? In M. A. P. Stephens, J. H. Crowther, S. E. Hobfoll, & D. L. Tennenbaum (Eds.), Stress and coping in late-life families (pp ). Washington, DC: Hemisphere. Zarit, S. H., & Leitsch, S. A. (2001). Developing and evaluating community based intervention programs for Alzheimer s patients and their caregivers. Aging and Mental Health, 5(Suppl.), S84 S98. Zarit, S. H., Stephens, M. A. P., Townsend, A., & Greene, R. (1998). Stress reduction for family caregivers: Effects of adult day care use. Journal of Gerontology: Social Sciences, 53B, S267 S277. Zarit, S. H., & Zarit, J. M. (1990). The Memory and Behavior Problems Checklist and the Burden Interview. University Park: The Pennsylvania State University. Received April 14, 2006 Accepted April 24, 2006 Decision Editor: Linda S. Noelker, PhD Vol. 46, No. 5,