Reducing Caregiver Burden: A Randomized Psychoeducational Intervention for Caregivers of Persons With Dementia

Transcription

1 Copyright 1999 by The Cerontological Society of America The Cerontologist Vol. 39, No. 3, This 3-year randomized clinical trial tested the effectiveness of an interdisciplinary psychoeducational family group intervention in decreasing the caregivers' perceptions of the frequency and severity of behavioral problems in persons with dementia and their reactions to those problems, and in decreasing caregiver burden and depression. The intervention consisted of seven weekly, 2-hour multimedia training sessions that included education, family support, and skills training for 9 primary caregivers and their families. Repeated measures ANOVA was used to test for significant differences between the intervention and waiting list control groups over a 5-month period. The intervention was successful in reducing caregivers' negative reactions to disruptive behaviors and in reducing caregiver burden over time. Key Words: Family caregivers, Psychoeducational intervention, Randomized clinical trial Reducing Caregiver Burden: A Randomized Psychoeducational Intervention for Caregivers of Persons With Dementia Sharon K. Ostwald, PhD, RN, 1 Kenneth W. Hepburn, PhD, 2 Wayne Caron, PhD, 3 Theressa Burns, BS, OTR, and Robert antell, Ed, RN 5 Family caregivers occupy a pivotal place in the care of persons with dementia in tne United States. Families provide the bulk of home care (Rockefeller, 1991). They persist in their role for long periods of time, even when engaged in other activities (Scharlach & Boyd, 1989; Stone, Cafferata, & Sangl, 1987). ost want to retain the role, and their efforts, when appropriately supported, forestall institutionalization (ittleman, Ferris, Shulman, Steinberg, & Levin, 1996; Ostwald et al., 1993). Caregivers commit great quantities of time and money to the task (Ernst & Hay, 199). As the number of dementia patients increases and the costs of their care increase past the current $100 billion annual estimate, the importance of these caregivers will likely increase. Caregiving extracts a toll from family caregivers. The caregiver rote is linked with health and mental health problems, higher than usual psychotropic drug use, social isolation, family stress, burden, and depression (Clipp & George, 1990; Lieberman & Fisher, 1995; Shulz, O'Brien, Bookwala, & Fleissner, 1995; Skaff & This project was supported, in part, by a grant from the National Institute of Nursing Research (NR02971) to Sharon K. Ostwald, PhD, RN, Principal Investigator. The authors gratefully acknowledge the participation of the following research assistants: Amy Krasnoff, PhD, Karen Blount, RN, PH, Dawn Reckinger, PH, Wendy VanLoy, PhD, Laura Kirk, RN, S, Sue Newman, SW, Becky Weberg, A, and Christy Anderson- Jacob, A. In addition, the authors acknowledge the statistical assistance of Deborah Finstad, BA, and Kamiar Kouzekanani, PhD. Address correspondence to Dr. Sharon K. Ostwald, University of Texas- Houston, Center on Aging, 7000 Fannin St., Suite 720, Houston, TX sostwald@admin.hsc.uth.tmc.edu department of Family Practice and Community Health, University of innesota, inneapolis. 'Department of Family Social Science, University of innesota, St. Paul. "Veterans Affairs edical Center, inneapolis. 'Department of Psychology, University of innesota, inneapolis. Pearlin, 1992; Wright, 199; Zarit, Reever, & Bach- Peterson, 1980). A complex of causes is associated with the negative outcomes caregivers experience. Persistent behavioral problems, coupled with caregivers' perceived or actual inability to handle them, emerge as important sources of caregivers' stress and burden (Baumgarten etal., 1992; Gallagher-Thompson, Brooks, Bliwise, Leader, & Yesavage, 1992; ontgomery & Kosloski, 199; O'Donnell etal., 1992; Pruchno, ichaels, & Potashnik, 1990; Teri, Rabins, et al., 1992; Winogrond, Fisk, Kirsling, & Keyes, 1987; Winslow, 1997). At the same time, family disharmony and lack of support appear to be important contributors to overall caregiver burden (Gold, Cohen, Shulman, & Zucchero, 1995; Hannappel, Calsyn, & Allen, 1993). The literature on interventions to reduce the adverse impacts of caregiving suggests three general conclusions. Interventions that have taken multiple, rather than single, approaches (e.g., support and education vs support alone) have been more successful in affecting the targeted outcome, such as depression or caregiver burden (Greene & onahan, 1989; Haley, 1989; ohide et al., 1990; Toseland, Labrecque, Goebel, & Whitney, 1992). Intervention programs tailored to particular caregiver situations or to specific patient behaviors rather than a general framework or approach have been more successful in affecting the identified outcome (Chiverton & Caine, 1989; Lovett & Gallagher, 1988; ittelman et al., 1996; Zarit, Anthony, & Boutselis, 1987). Finally, previous research has shown that caregivers can acquire skills and knowledge that permit them to carry out sophisticated tasks (Gendron, Poitras, Dastoor, & Perodeau, 1996; Pinkston & Linsk, 198; Seltzer, Ivry, & Litchfield, 1987). Vol. 39, No. 3,

2 This article reports the results of a 3-year randomized trial of a psychoeducational intervention targeted to primary caregivers of community-dwelling persons with dementia. The intervention, the innesota Family Workshop (FW), aimed to improve outcomes for family caregivers and the persons with dementia through a multidimensional training approach provided to primary caregivers and other family members. ethods Theory Base and Intervention The FW used a general stress mediation model to conceptualize an approach to caregiver training (Aneshensel, Pearlin, ullan, Zarit, & Whitlatch, 1995; Lawton, oss, Kleban, Glicksman, & Rovine, 1991; Folkman, Lazarus, Gruen, & DeLongis, 1986; Pearlin, ullan, Semple, & Skaff, 1990). This model posits that multidimensional mediation may be needed to improve outcomes in a fundamentally stressful situation. The model suggests a number of factors that might mediate the impact of the caregiving situation. These factors include the caregiver's personal resources for caregiving (knowledge, skills, and belief in own abilities) and external resources for caregiving, especially the organization of the rest of the family to support the caregiver. Consistent with this theory base, we developed a 1-hour training workshop (provided in seven weekly, 2-hour training sessions) for primary caregivers and their families. BuiTding on the conclusions of previous intervention research, the workshop sought, through a multifaceted curriculum that emphasized the development of individualized strategies of care, to: 1. Provide caregivers with information about dementing diseases and how they affect persons with dementia, caregivers, and the family as a system. 2. Develop and strengthen caregivers' practical skills for dealing with caregiving tasks on a day-to-day basis. 3. Strengthen caregivers' feelings of confidence and belief that they are able (competent) to deal with issues, day in and day out.. Improve family communication and cooperation by increasing the congruence of caregivers' and families' perceptions about the nature and level of impairment in the person with dementia and by facilitating the family's ability to work collaboratively to find solutions to current management problems. The workshop curriculum focused particularly on linking the dementia-produced deterioration of cognitive functioning with changes in care receivers' behaviors and emotions and with day-to-day management. Cognitive functioning describes a person's ability to undertake and complete an activity,-taking into account the complexity of the activity (Allen, 1988). The concept includes information about distractibility, ability to engage meaningfully in goal-directed activity, the extent of an activity that can be undertaken, and the effectiveness of specific actions or use of objects. Knowledge about a person's level of cognitive functioning helps caregivers design caregiving strategies around the current abilities of the person with dementia. It helps them understand what activities are a good fit (e.g., in terms of scope, difficulty, and repetitiveness) for the person. It helps caregivers understand how much and what kind of direction to provide, what kinds of performance expectations to maintain, and how to prompt and communicate effectively. The workshop followed a written curriculum that identified teaching points and learning objectives for each session. The curriculum specified the faculty, exercises, readings, and homework for each session, and the workshop maintained a deliberate classroom format. One of the investigators monitored each session to ensure that the teaching points were covered in conformance with the curricular plan. At least two members of the interdisciplinary faculty (nurses, occupational therapists, family therapists, and educators) took part in each workshop. In the first four sessions, the workshop employed lectures, videotapes (of persons at different cognitive functioning levels), and exercises to provide information about dementia, its impact on the person and family, and cognitive functioning. In the fifth session, caregivers and other family members viewed videotapes of their own family members being assessed with the toast test portion of the Cognitive Performance Test (Burns, ortimer, & erchak, 199). At this time they were told of the result of the testing and given their family members' cognitive functioning levels. The last two workshop sessions sought to develop and enhance caregiving skills and mastery. In these sessions, participants reported on their in-home efforts to apply this new understanding of their care recipient's cognitive performance level to day-to-day management and behavior issues. Feedback from faculty and other participants concentrated on caregivers' reported strategies and their sense of accomplishment. The FW maintained a deliberate focus on family involvement in dementia care. In addition to the two sessions specifically addressing the impact of the disease on the family, the workshop used group process techniques throughout to attempt to involve all family participants as active learners and to serve as resources for the primary caregiver in his or her development of skills and mastery. In order to free caregivers to attend the workshop, the persons with dementia were invited to attend a concurrent group. This group was optional, except for the first two sessions, and provided day care-like activities adapted for participants' levels of functioning (e.g., cooking tasks and games). During the first two sessions of this group, tests were administered individually to determine care recipients' cognitive functional levels on the Allen scale. This scale delineates six ordinal levels of global function ranging from normal (level 6) to profoundly disabled (level 1). Allen, Earhart, and Blue (1992) also delineate five modes of performance within each functional level that further quantify behavior variations observed within each level. One author (T.B.), trained in this model of assessment, observed patients' performance in the group, administered performance tests, and used the ini- 300 The Gerontologist

3 ental State Exam (SE; Folstein, Folstein, & chugh, 1975) scores and caregiver information about daily activities to determine the level. The performance tests included those developed by Allen and colleagues (Kehrberg, Kuskowski, ortimer, & Shoberg, 1992) and the toast test from the Cognitive Performance Test (Burns et al., 199). The toast test asks the cognitively impaired person to make a piece of toast and put butter and jelly on it. The request, although reflecting a common, everyday activity, involves planning and following a sequence of steps (plug in and operate a toaster, remove bread from a wrapper, uncover butter, choose a knife, open a jar of jelly, etc.). Individuals were videotaped performing the toast test. These performance measures gave a rounded level score, which we felt to be adequate for teaching purposes. Hypotheses We hypothesized that, compared with those in a waiting list control group, caregivers of persons with dementia enrolled in the WF would report decreased frequency and severity of behavior problems in the care receiver. We also posited that primary caregivers enrolled in the FW would demonstrate decreased levels of burden and depression and decreased negative reactions to patient behaviors, compared with the caregivers in the waiting list control group. Design A random assignment experimental design was used to study the effects of the workshop. We recruited participants continuously throughout the project. We began new workshops at approximately 2-month intervals when at least four family groups were available to begin. As families agreed to take part in the study, we completed informed consent procedures and asked them to complete an initial series of questionnaires and interviews. Once these were done, each family was randomly assigned by a computer program to one of two groups (each family had an equal chance of assignment to either group). Families in the first group were invited to participate in the next 7- week workshop (generally within 3- weeks after the initial data gathering). Those in the second group were put on a waiting list for a workshop to begin 5-6 months later. Following randomization, the same interview and questionnaire data were collected from each family at points approximately 3 months and 5 months after the initial data gathering. Families in the first group would have completed a workshop before the second and third data gathering sessions, but those on the waiting list would not have entered a workshop before completing the third data gathering session. No data were collected from waiting list families following their participation in the workshop. At the time of enrollment, all families were given a packet of information about resources available in the community for Alzheimer's care (e.g., adult day centers and support groups). The study personnel had no control over the use of other services by either immediate intervention or waiting list group families. We conducted 16 workshops over the course of the study. Workshops varied considerably in size. A typical workshop included 5 primary caregivers and 8-10 other family members. Some workshops were smaller ( primary caregivers and 5 other family members), and some were larger (7 primary caregivers and 15 other family members). Given the design of the study, the first two workshops included only families who had been assigned to participate immediately. The last workshop included only families from the waiting list condition. In the other 13 workshops, participants included both families in the immediate assignment group and families who had completed the three data gathering sessions and who were now off the waiting list. The workshop faculty had no way of knowing the study status (immediate or waiting list) of workshop participants, except in the first two and the last workshops. In those workshops in which immediate and waiting list families were mixed together, no identification of study status was made and no distinctions were made in terms of instruction. Because recruitment criteria were broad, each workshop included a mix of families, regardless of study status, dealing with a wide range of dementia-related problems, so study status could not be inferred by faculty or participants based on the dementia person's condition. Sample Persons with dementia and their families were recruited from local memory loss clinics, the Geriatric Research, Education, and Clinical Center at the inneapolis Veterans Hospital, other senior clinics and health centers, the inneapolis/st. Paul Alzheimer's Association, and local community hospitals and health and social service agencies. To be enrolled in the study, families needed to be caring for a person with a diagnosis of nonreversible dementia who 1. Was living in the community. 2. Demonstrated signs of mild to severe dementia; only those who were nonambulatory and required total care (judged to be at or below a telephoneadministered Functional Assessment Staging Test [Reisberg, 1988] score of 7b) were excluded. 3. Displayed behavior problems (as perceived by the caregiver).. Was able to accompany the caregiver to at least the first two weekly intervention sessions. In addition, at least one additional family member, other than the primary caregiver, needed to be willing to accompany the caregiver anctpatient to seven weekly sessions. Instruments Basic demographic information on the caregiver and patient was provided by the primary caregiver. Age was collected as a continuous variable. Income, educational level, and gender were collected as categorical variables. Family caregivers were assessed for Vol. 39, No. 3,

4 depression, burden, and negative response to the care receivers' behaviors using standard paper and pencil surveys. Cronbach's coefficient alpha was employed to estimate the reliability/internal consistency of the instruments used to measure the major variables of interest. The 22-item Revised Caregiver Burden Scale (Zarit, Orr, & Zarit, 1987) measures level of burden or distress associated with caregiving for a dependent elder. The scale has a reliability level of.91 and a test-retest correlation of.71 (Gallagher, Wrabetz, Lovett, Del aestre, & Rose, 1989). Reliability coefficients for this variable were calculated on this sample; at pretest, posttest, and follow-up the reliability coefficients were.91,.92, and.92, respectively. Caregiver depression was measured using the 20-item self-report Center for Epidemiological Studies- Depression scale (CES-D). The scale has strong evidence of convergent and discriminant validity; its reliability level is reported as.85 and the test-retest correlation as.67 (Radloff, 1977; Roberts, 1980). The reliability coefficients for caregiver depression in this sample were.88 at pretest,.88 at posttest, and.90 at follow-up. Frequency of behavioral symptoms and the caregiver's reaction to these symptoms were measured using the Revised emory and Behavior Problem Checklist (RBPC). The alpha for frequency of disruptive behaviors reported by the caregivers was.83 (Teri, Truax, et al., 1992). Reliability coefficients for frequency of disruptive behaviors at pretest, posttest, and follow-up in this study were.11,.7, and.71, respectively. The alpha for caregiver's level of distress associated with the disruptive behaviors, reported by Teri and colleagues, was.8. Reliability coefficients for this variable at pretest, posttest, and follow-up were.77,.7, and.68, respectively, in this study. Persons with dementia were assessed for cognitive ability using the SE. Scores on the SE range from 0 to 30; scores of 2 and below are associated with dementia. Reviews of the SE report an internal consistency of >.80, good validity, sensitivity >80%, and specificity >70% (Bachman, et al., 1993; Tombaugh & clntyre, 1992). The reliability coefficients for the SE in this sample were.80 at pretest,.83 at posttest, and.83 at follow-up. The individual doing the cognitive assessments was masked to the group assignment and did not participate in the group intervention sessions. Data Analysis Repeated measures analysis of variance (ANOVA) was used to test for significant differences between the caregivers in the intervention and waiting list control groups, at three points in time, on the major variables of (a) caregiver reaction to disruptive behaviors, (b) caregiver burden, and (c) caregiver depression. In addition, repeated measures ANOVA was used to test for differences between the patients in the intervention and waiting list control groups at preintervention, postintervention, and follow-up on (a) frequency of patient disruptive behaviors (as reported by caregivers), and (b) cognitive status. Specifically, a series of one between-subjects variable (treatment vs control) and one within-subjects variable (pretest, posttest, followup) repeated measures ANOVA was performed. This statistical technique, which removes from the error term variability among the subjects due to individual differences, was used to test intervention, time, and intervention by time interaction effects (Stevens, 1996). Of particular interest was the effect of the intervention over time. All variables were tested for normality; the stem-and-leaf diagram showed one outlier on caregivers' perception of the frequency of patients' disruptive behaviors data which was not included in the analysis of that variable. The Box test showed that the assumption of homogeneity of covariance matrices had been met for all analyses. The number of subjects was not the same for all analyses due to missing data. The level of significance was set at.05. Results Recruitment, Randomization, and Attrition A total of 117 families met study criteria and were randomly assigned to either the immediate intervention group or a 5-month waiting list control group. The method of randomization resulted, by chance, in a very uneven distribution of subjects to the two groups. As a result of randomization, 72 families were assigned to immediate workshop participation and 5 families were placed on the 5-month waiting list. The average age of the caregivers was 65.6 years and the average age of care receivers was 77.1 years. Sixty-five percent of the caregivers and.% of the care receivers were female. Almost all of the sample was of Northern European descent, which is representative of the inneapolis-st. Paul area. In general, the participants were well educated, with 7.7% of the caregivers and 57.3% of the care receivers having at least some education beyond high school. The majority of the care receivers (8.9%) lived with the family caregivers. Ninety-four families completed the program. Families were considered to have completed the program if (a) the caregiver participated in preintervention, postintervention, and follow-up assessments, and (b) in the immediate intervention group, the caregiver and at least one other family member attended six of the seven workshop sessions. A total of 23 families did not complete the program for an attrition rate of 19.7%. Twelve families discontinued participation in the immediate intervention group (16%) and 11 families dropped out of the waiting list control group (26%). The dropouts and completers (both intervention and waiting list control groups combined) were compared on the basis of their baseline scores on the study's major outcome variables (i.e., caregiver's perception of disruptive behaviors, caregiver's reaction to disruptive behaviors, caregiver burden, and caregiver depression). The two groups were also compared on the basis of demographic variables of caregiver and patient age, caregiver and patient education, caregiver and patient gender, and caregiver and patient income. With the exception of patient age, none of the differences was 302 The Gerontologist

5 statistically significant at the.05 level. Results are summarized in Table 1. Reasons for discontinuing the program are known for 13 of the 23 families. Seven patients were admitted to long-term care facilities, and therefore were dropped from the analysis because they no longer met the study criteria of living in the community. Three families discontinued because the condition of the person with dementia had deteriorated, making it difficult to leave the house, and two families did not have transportation. In addition, one family discontinued the program due to family problems. The remaining 10 families did not give reasons for discontinuance. Comparison of Sample at Baseline The intervention and control groups were compared on the basis of demographic variables and patient mental status: (a) SE; (b) patient and caregiver age; (c) patient and caregiver gender; (d) patient and caregiver education (coded as either <high school or >high school); (e) patient and caregiver income (coded as either <$29,999 or >$29,999); (f) whether patient lived with the caregiver; and (g) caregiver's relation to patient. Significance, for each analysis, was set at the.05 level, based on available baseline data. We employed t test for independent samples to compare the two groups on the basis of SE and age. Fisher's two-sided Exact Probability test was used to compare the two groups on the basis of dichotomous variables of sex, education, income, and whether the patient lived with the caregiver. None of the differences were statistically significant at baseline, except caregiver education, which was only significant in the analysis that compared the cognitive status of the patients in the intervention and control groups and so was not relevant to the analysis. There was no difference in patient education between the two groups. Because of several cells with expected frequency of less than 5, statistical comparisons of the two groups on the basis of Caregiver's Relation to Patient could not be performed. The results are summarized in Table 2. There were five major analyses in this study: (a) caregivers' burden, (b) caregivers' depression, (c) caregivers' response to descriptive behaviors, (d) care receivers' descriptive behaviors, and (e) care receivers' Table 1. Comparison of Completers and Dropouts at Completers Dropouts Variable t P Disruptive Behaviors Response to Negative Behaviors Burden CES Depression Caregiver Age Patient Age t ( t Caregiver Education < high school >high school Patient Education < high school >high school Caregiver Gender ale Female Patient Gender ale Female Caregiver Income < $29,999 >$29,999 Patient Income < $29,999 >$29, X2(1 ; N = 11) = 2.1/ P z =.1 X 2 (1, N = 111) =.1, P =.71 X 2 d, N - 117) =.9, P =.8 X 2 d, N = T\7) = 1.1' P z =.29 X 2 d, N = 107) = 1.21 > P z =.27 X 2 (1, N = 101) =.20, P =.65 Note: Due to unequal sample sizes, Welch Approximate t test was used to compare the dropouts and completers on continuous variables. Yates' Correction for Continuity was applied in the calculation of chi-square values. Vol. 39, No. 3,

6 Table 2. Demographic Characteristics of Intervention and Groups at Baseline Table 2. Demographic Characteristics of Intervention and Groups at Baseline (Continued) Intervention Intervention Variable Analysis () or n () or n p Variable Analysis () or n () or n p SE Patient Age Caregiver Age Patient Gender Caregiver Gender (7.10) (7.20) (7.16) 1*7.62 (7.3) (7.12) (7.62) (7.55) (7.76) (7.2) 75.2 (7.70) (12.6) (12.52) (12.96) (12.58) 65.1 (12.59) 33, 19F 33, 20F 31, 19F 33, 18F 28, 17F 18, 3F 18, 35F 18, 32F 17, 3F 15, 30F (7.33) (7.33) (6.8) 19.2 (7.5) (6.9) (10.35) (10.35) 77.3 (9.60) (10.9) (10.39) (15.87) (15.87) 6.77 (1.92) 6.13 (15.9) (15.92) 16, 15F 16, 15F 16, 1F 16, 1F 16, 13F 8, 23F 8, 23F 8, 22F 8, 22F 8, 21 F Caregiver's 1 Relation to Patient Sp, 12 Ch 1 Sib, 3 In, 1 Fr 35 Sp, 12 Ch 1 Sib, 3 In, 1 Fr 33 Sp, 11 Ch 1 Sib, 3 In, 1 Fr 36 Sp, 11 Ch 1 Sib, 1 In, 1 Fr 31 Sp, 12 Ch 1 Sib, 1 In, 1 Fr 19 Sp, 12 Ch 19 Sp, 12 Ch 19 Sp, 11 Ch 19 Sp, 11 Ch 18 Sp, 11 Ch Note: Analyses: 1 = Disruptive behaviors, intervention group, n = 52, control group, n = 31; 2 = Response to negative behaviors, intervention group, n = 53, control group, n = 31; 3 = Zarit Burden Inventory, intervention group, n = 50, control group, n = 30; = Depression, intervention group, n - 51, control group, n - 30; 5 = SE, intervention group, n = 5, control group, n = 29. Education: A = <high school, B = >high school; Sex: = male, F = female;lncome: C = <$29,999, D = >$29,999. Relation: Sp = spouse; Ch = child; Sib = sibling; In = in-law; Fr - friend. cognitive status. Not all caregivers completed all items in the instruments. In accordance with human subject protection protocols, subjects were free not to respond to questions. Thus, the number of subjects reported in these analyses varies. Patient 1 Education Caregiver 1 Education Patient Income Caregiver Income 20A, 30B 21A, 30B 21A, 27B 21A, 29B 19A, 25B 15A, 36B 16A, 36B 16A, 33B 17A, 33B 16A, 27B 27C, 16D 28C, 16D 27C, 16D 28C, 15D 26C, 10D 23C, 23D 2C, 23D 22C, 23D 25C, 21 D 2C, 16D 13A, 17B 13A, 17B 13A, 16B 12A, 17B 10A, 17B A, 27B A, 27B A, 26B A, 26B 3A, 25B 19C, 10D 19Q 10D 19C, 9D 18C, 10D 17C, 9D 10C, 20D 10C, 20D 10C, 20D 10C, 19D 10C, 18D Caregivers' Burden A comparison of the means and standard deviation scores for the intervention and waiting list control groups for caregivers' burden are shown in Table 3. With respect to caregiver burden, the effect of the intervention over the 5-month period was of particular interest. The intervention by time interaction (showing a preintervention to follow-up measurement effect) was significant, F(2,156) = 5.53, p =.005. Analysis of simple effects showed a marginal increase from pretest to posttest, (9) =.50, p =.62, and a moderate decrease from posttest to follow-up, (9) = 2.23, Table 3.,, and ean Scores () and Standard Deviations () for the Intervention and Waiting List Groups on the Zarit Burden Inventory (N = 80) Intervention Group (n = 50) Group (n = 30) Patient 1 Living w/ 2 Caregiver Yes, 8 No 39 Yes, 8 No 37 Yes, 7 No 39 Yes, 6 No 36 Yes, 5 No Yes, Yes, Yes, Yes, Yes, No No No No No a 5.13 a Statistically significant decrease at.05 level. Statistically significant increase at.05 level b b The Gerontologist

7 Table.,, and ean Scores () and Standard Deviations () for the Intervention and Waiting List Groups on the Center for Epidemiological Studies-Depression Scale (N = 81) Intervention Group (n = 51) (n = Group 30) p =.03, in the intervention group, suggesting a delayed effect. In the waiting list control group, on the other hand, a marginal decrease from pretest to posttest was observed, (29) =.77, p =.5, followed by a substantial increase from posttest to follow-up, (29) = 3.38, p =.002. The intervention, F(1,78) =.29, p =.60, and time, F(2,156) =.31, p =.73, effects were not statistically significant. Caregivers' Depression ean scores on the CES-D for the intervention and waiting list control groups did not fall above the accepted cutoff for depressive symptomatology (16 points) on the initial pretest measurement. There were no significant intervention, F(1,79) = 2.10, p -.15, or intervention by time interaction effects, F(2,158) = 1.92, p =.15. However, there was a significant time effect, F(2,158) = 1.29, p =.001. The mean scores, shown in Table, illustrate that the scores rose for both groups during the intervention and fell after the intervention. The fall in the intervention group was sharper but did not reach the level of significance. Trie pattern, however, was quite different for the 26 participants whose initial CES-D scores fell above the cutoff score of 16. The intervention and waiting list control groups had almost identical baseline mean scores. By the follow-up assessment 5 months later, the mean scores for the waiting list group had risen whereas the mean scores for the intervention group had fallen (see Table 5). The intervention by time interaction effect, F(2,8) = 2.50, p =.09, although not statistically significant (perhaps clue to small sample Table 5.,, and ean Scores () and Standard Deviations () for the Intervention and Waiting List Groups Who Scored Above the Cutoff for Depressive Symptoms (>16) on the Center for Epidemiological Studies-Depression Scale (N = 26) Intervention Group (n = 16) (n = Group 10) size), suggests that caregivers who entered the program with depressive symptomatology might have achieved some benefits from the intervention over time. The intervention, F(1,2) = 2.15, p =.16, and the time, F(2,8) = 1.5, p =.2, effects were not significant. Caregivers' Response to Disruptive Behaviors A comparison of the means and standard deviation scores for caregivers' negative responses to the care receivers' disruptive behaviors is shown in Table 6. The time effect, F(2,16) = 3.15, p =.0, and the intervention by time interaction effect, F(2,16) =.60, p =.01, were statistically significant. The intervention effect alone was not statistically significant, F(1,82) =.03, p =.87. As can be seen in Table 6, in the intervention group, the mean scores decreased from pretest to posttest and from posttest to followup. Analysis of simple effects showed that the change from pretest to posttest was statistically significant, (52) = 2.3, p =.02, whereas the change from posttest to follow-up was not statistically significant, (52) = 1.72, p =.09. In the waiting list control group, however, the mean scores decreased insignificantly from pretest to posttest, (30) = 1.07, p =.29, but increased from posttest to follow-up, (30) = 1.66, p =.11. Caregivers in the intervention group were able to achieve a reduction in their negative reactions to the disruptive behaviors displayed by their family member in the short term, and to maintain that change even after the completion of the program. Care Receivers' Disruptive Behaviors Disruptive behaviors are a major source of stress and relationship strains between caregivers and care receivers. The FW was successful in reducing caregivers' negative reactions to disruptive behaviors and reducing caregiver burden over time, but was not successful in significantly reducing the behaviors reported by caregivers. As can be seen in Table 7, the mean scores for caregiver's perception of frequency of disruptive behaviors decreased for both groups from pretest to posttest. These scores increased slightly from posttest to follow-up in the intervention group but remained below the pretest level. They increased sub- Table 6.,, and ean Scores () and Standard Deviations () for the Intervention and Waiting List Groups on the Caregiver Response to Disruptive Behavior Subscale of the Revised emory and Behavior Problems ChecklisH/V = 8) Intervention Group (n = 53) 6.76 a 5.00 a (n = Statistically significant decrease at.05 level. Group 31) Vol. 39, No. 3,

8 Table 7.,, and ean Scores () and Standard Deviations (5/7) for the Intervention and Waiting List Groups on the Disruptive Behavior Subscale of the Revised emory and Behavior Problems Checklist (N = 83) Intervention Group (n = 52) (n = Croup 31) stantially in the waiting list control group during the same time period, suggesting a quadratic trend. However, the intervention, F(1,81) =.6, p =.3, the time, F(2,162) = 2.71, p =.07, and the intervention by time interaction, F(2,162) = 2.62, p =.08, effects were not statistically significant. Care Receivers' Cognitive Status Persons with dementia in both the intervention and control groups showed a decrease in their cognitive abilities. The mean scores and standard deviations for the pretest, posttest and follow-up are shown in Table 8. We did not hypothesize an effect of the intervention on cognitive function, but monitored the cognitive status as a measure of comparability between groups. The mean scores for patients in the intervention and control decreased during the 5-month period. There was no intervention effect, F(1,72) =.92, p =.32, and no intervention by time effect, F(2,1) =.81, p =.5. However, there was a significant time effect, F(2,1) = 12.80, p = The analysis of simple effects showed that the change from posttest to follow-up was statistically significant, (73) = 3.9, p =.001. The pretest to posttest was not statistically significant at the.05 level, t(73) = 1.23, p =.22. Discussion This interdisciplinary psychoeducational study demonstrated that involving primary caregivers and other family members in a short-term (7-week) workshop with a clearly defined curriculum that combines infor- Table 8.,, and ean Scores () and Standard Deviations () for the Intervention and Waiting List Groups on the ini-entaf State Exam on Patients (N = 7) Intervention Croup (n = 5) (n = Group 29) mation and behavior management training with attention to family support can have significant positive effects for caregivers. Significant changes occurred in caregivers' ratings of their reactions to care receivers' disruptive behaviors. In addition, a significant difference in caregiver burden emerged between intervention and waiting list caregivers 5 months after the initial assessment. A trend toward improvement in depressive symptoms in the small number of intervention group participants who entered the program with depressive symptomology was also observed. The hypothesized effect of the workshop on the caregivers' burden scores were not apparent until the 5-month follow-up measurement, suggesting delayed treatment effects. The caregivers in the intervention group had moderate but significant reductions in burden posttest to follow-up, but even more strikingly, the caregivers in the waiting list control group had substantial increases in their burden scores during the same period of time. This suggests that the workshop not only resulted in a moderate reduction in burden after the intervention, but also prevented a significant increase, as was seen in the waiting list control group. These effects cannot be explained by improvements in the care receivers in the intervention group as compared to the control group. There were no significant differences in the frequency of disruptive behaviors between the groups, and both groups of care receivers demonstrated cognitive decline on the SE. Nor, could these 5-month follow-up findings be explained by a change in living arrangements. We were concerned that changes in living arrangements, such as admission to a nursing home, might selectively affect caregivers' burden and depression in the follow-up period. Therefore, to be included in the analysis, caregivers needed to be providing assistance to patients with dementia in the community at all three points in time. Thirty percent of the attrition in this study was related to eliminating from the analysis families who had admitted a relative to a nursing facility prior to the follow-up assessment. These results are somewhat puzzling and unanticipated. We had thought the positive effects might be seen more immediately and had considered that the effects might either remain in place or improve with time. The stress mediation theory on which the curriculum is based offers one possible explanation of the results. According to the model on which we based the workshop curriculum, successful mediation of stress in the caregiving situation depends on the interaction of a number of factors. These include the acquisition of knowledge, the acceptance, in many cases, of the care receiver's true condition, and changes in patterns of family interaction and organization for caregiving. They also include the development of caregivers' confidence in the effectiveness of the management techniques taught in the workshop and in their ability to apply these effectively in day-to-day caregiving. Tnus, the workshop included a great deal of information that had to be put to use and practiced in order to be understood and integrated. It may be that the development and integration of these mediators occurred over time and continued after the workshop, thereby 306 The Gerontologist

9 explaining its delayed effect. However, a cross-over design would have strengthened confidence in the findings and this interpretation. Competing explanations should, however, be considered. Although randomization of subjects should ensure that there was no systematic bias in the access of either study group to external services, differential use could have occurred and could, in part, account for study results. Because we had no control over the kind and amount of services participants were receiving outside the workshop, it is possible that participants randomized to the immediate intervention learned about services from the group and acquired, from each other, skills or confidence to seek and ask for additional care. In this case, study effects could have been produced by added services. A second issue is raised by mixing, in virtually all of the workshops, families randomized to the immediate intervention with those coming off the waiting list. By the time of their entry to a workshop, the typical waiting list caregiver was significantly more burdened by caregiving and more bothered by the care receiver's disruptive behavior than was the typical caregiver randomized to the immediate intervention. This could have resulted in the caregivers randomized to the immediate intervention being exposed to levels of upset that might have affected response to their own situation. In this case, the delayed effects observed in the 5-month follow-up measurement may also reflect the impact of a kind of contamination produced by mixing the two study groups. There is a great deal about the process of the workshop that needs further study. The workshop sought to provide caregivers and their families with opportunities to look at their situations more objectively. The use of various teaching strategies (minilectures, group and individual support, and videotapes) and homework assignments that asked caregivers to identify and solve specific problem behaviors confronting them in their caregiving role were intended to take into account the variety of learning styles among adults and to involve them in experiential learning. Videotapes of the care receivers completing a task (making a piece of toast) were intended to break through the denial of disease so often present in families. We can point to anecdotal evidence of the effectiveness of the workshop's educational strategies. A number of family members commented on how seeing their relative's videotape helped them understand and acknowledge the devastating effects of the dementia. A number of caregivers reported the workshop enabled them to redefine the situation and its meaning, to try out new approaches to troublesome behaviors, and to make appropriate adaptations to the environment. However, we recognize that much more systematic formative study should be undertaken to evaluate the effectiveness of various components of this psychoeducational intervention. The family involvement required in this intervention was greater than usually reported in caregiver intervention programs. In addition to the primary caregiver, at least one other family member was required to attend the sessions, but attendance was not limited to just one other family member. In one family, eight others came to almost every session. Involvement may have provided sustained support to caregivers; that could partly explain the delayed effects. Family involvement may also have hindered some caregivers. As group members proceeded through the curriculum together, the interaction between family members and primary caregivers varied greatly. Some families provided emotional support and reinforced caregivers' motivation to work on the program material. In other cases, divisions remained or widened and may have added, rather than reduced, stress. Although it was our impression that family participation was a positive factor in caregivers' outcomes, a much more detailed examination is needed to assess the effect of family involvement in producing long-term effects on caregivers. The workshop format had a number of serious limitations. Because it was a research-based intervention, the workshop had no other context in the lives of its participants. Unlike other interventions that have shown positive impact on caregivers (e.g., ittleman et al., 1996), the FW was not a part of some other structure like a medical practice, an HO, or even an Alzheimer's Association chapter. In consequence, the workshop could offer neither continuity nor followup. We did note, by the end of each workshop, that a number of participants voiced a wish for follow-up. Similarly, the workshop could exercise no control over the other services participants used either while taking part in the workshop or waiting to participate in it. Although a training program like this appears valuable as a stand-alone intervention, it would be important to study whether it might produce even better outcomes if it were a component of a larger, more comprehensive system of care for persons with dementia and caregivers, a system with the ability to identify and meet very specific service needs of caregivers. An additional limitation to this study was the homogeneity of the participating caregivers. Although minority participants were representative of the population as a whole in the inneapolis-st. Paul area, they represented fewer than 10 of the 117 families who were randomized to participate in the FW. Therefore, no conclusions can be drawn about the appropriateness of the FW for specific ethnic or racial groups. In addition, the caregivers were well educated; 71.1% had at least some education above the high school level. Well-educated individuals may seek group educational experiences, such as this workshop. Their previous success in educational settings may also have oeen a major factor in the enthusiasm and motivation that they displayed in learning the complex content. A workshop such as this might be intimidating to caregivers who have had lejs experience or success in educational settings. Other formats might be more successful with caregivers with different demographic characteristics; significant adaptations might need to be made to the educational materials for groups with different ethnic or racial or educational profiles. One other design issue stands as a limitation. The unexpected results of randomization produced study groups of very unequal size. We had hoped to recruit larger numbers of families to the study and to have a Vol. 39, No. 3,

10 more typical distribution through randomization. Although the groups were basically equivalent on the major variables under study at baseline, the unequal groups and low number of participants both affect power and both need to be addressed in future studies. Nevertheless, the findings of this 3-year randomized study are consistent with the findings of other researchers: interventions with multiple approaches targeted to specific behavior problems can be successful in affecting outcomes, such as caregiver burden. In addition, this study demonstrated that caregivers could experience less distress from care receivers' disruptive behaviors even when dementia is progressing. The decrease in burden in the intervention group between the posttest and the follow-up, coupled with the dramatic increase during the same time period in the waiting list control group, supports the need for further research to replicate these findings and better explain the underlying mechanisms. References Allen, C. K. (1988). Occupational therapy: Functional assessment of the severity of mental disorders. Hospital and Community Psychiatry, 39, Allen, D. K., Earhart, C. A., & Blue, T. (1992). Occupational therapy treatment goals for the physically and cognitively disabled. Rockviile, D: American Occupational Therapy Association. Aneshensel, C. S., Pearlin, L. I., ullan, J. T., Zarit, S. H., & Whitlatch, C. J. (1995). Profiles in caregiving: The unexpected career. San Diego, Academic Press, Inc. Bachman, D. L, Wolf, P. A., Linn, R. T., Knoefel, J. E., Cobb, J. L, Belanger, A. J., White, L. R., & D'Agostino, R. B. (1993). Incidence of dementia and probable Alzheimer's disease in a general population: The Framingham study. Neurology, 3, Baumgarten,., Battista, R.N., Infante-Rivard, C, Hanley, J. A., Becker, R., & Cauthier, S. (1992). The psychological and physical health of family members caring for an elderly person with dementia. Journal of Clinical Epidemiology, 5, Burns, T., ortimer, J. A., & erchak, P. (199). Cognitive performance test: A new approach to functional assessment in Alzheimer's disease. Journal of Geriatric Psychiatry and Neurology, 7, 6-5. Chiverton, P., & Caine, E. D. (1989). Education to assist spouses in coping with Alzheimer's disease: A controlled trial. Journal of the American Geriatrics Society, 37, Clipp, E. C, & George, L. K. (1990). Psychotropic drug use among caregivers of patients with dementia. Journal of the American Geriatrics Society, 38, Ernst, R. L, & Hay, J. W. (199). The U.S. economic and social costs of Alzheimer's disease revisited. American Journal of Public Health, 8, Folkman, S., Lazarus, R. S., Cruen, R. J., & DeLongis, A. (1986). Appraisal, coping, health status, and psychological symptoms. Journal of Personality and Social Psychology, 50, Folstein,. F., Folstein, S. E., & chugh, P. R. (1975). "ini-ental State": A practical method for grading the cognitive state of patients for the clinician. Journal of Psychiatric Research, 12, Gallagher, D., Wrabetz, A., Lovett, S., Del aestre, S., & Rose, J. (1989). Depression and other negative effects in family caregivers. In E. Light & B. D. Lebowitz (Eds.), Alzheimer's disease treatment and family stress: Directions for research (pp ). Washington, DC: U. S. Government Printing Office. Gallagher-Thompson, D., Brooks, J. O., Bliwise, D., Leader, J., & Yesavage, J. A. (1992). The relations among caregiver stress, "sundowning" symptoms, and cognitive decline in Alzheimer's disease. Journal of the American Geriatrics Society, 0, Cendron, C, Poitras, L., Dastoor, D. P., & Perodeau, G. (1996). Cognitive-behavioral group intervention for spousal caregivers: Findings and clinical considerations. Clinical Gerontologist, 17, Gold, D. P., Cohen, C, Shulman, K., & Zucchero, C. (1995). Caregiving and dementia: Predicting negative and positive outcomes for caregivers. International Journal of Aging and Human Development, 1, Greene, V. L., & onahan, D. J. (1989). The effect of a support and education program on stress and burden among family caregivers to frail elderly persons. The Gerontologist, 29, Haley, W. E. (1989). Group intervention for dementia family caregivers: A longitudinal perspective. The Gerontologist, 29, Hannappel,., Calsyn, R. J., & Allen, G. (1993). Does social support alleviate the depression of caregivers of dementia patients? Journal of Gerontological Social Work, 20, Kehrberg, K., Kuskowski,., ortimer, J., & Shoberg, T. (1992). Validating the use of an enlarged, easier-to-see Allen cognitive level test in geriatrics. Physical and Occupational Therapy in Geriatrics, 70(3), 1-1. Lawton,. P., oss,., Kleban,., Glicksman, A., & Rovine,. (1991). A two-factor model of caregiving appraisal and psychological wellbeing. Journal of Gerontology: Psychological Sciences, 6, P181-P189. Lieberman,. A., & Fisher, L. (1995). The impact of chronic illness on the health and well-being of family members. The Gerontologist, 35, Lovett, S., & Gallagher, D. (1988). Psychoeducational interventions for family caregivers: Preliminary efficacy data. Behavior Therapy, 19, ittelman,. S., Ferris, S. H., Shulman, E., Steinberg, G., & Levin, B. (1996). A family intervention to delay nursing home placement of patients with Alzheimer's disease. Journal of the American edical Association, 276, ohide, E. A., Pringle, D.., Streiner, D. L., Gilbert, J. R., uir, G., & Tew,. (1990). A randomized trial of family caregiver support in the home management of dementia. Journal of the American Geriatrics Society, 38, 6-5. ontgomery, R. J. V., & Kosloski, K. (199). A longitudinal analysis of nursing home placement for dependent elders cared for by spouses vs adult children. Journal of Gerontology: Social Sciences, 9, S62- S7. O'Donnell, B. F., Drachman, D. A., Barnes, H. J., Peterson, K. E., Swearer, J.., & Lew, R. A. (1992). Incontinence and troublesome behaviors predict institutionalization in dementia. Journal of Geriatric Psychiatry and Neurology, 5, Ostwald, S. K., Leonard, B., Choi, T., Keenan, J., Hepburn, K., & Aroskar,. A. (1993). Caregivers of frail elderly and medically fragile children: Perceptions of ability to continue to provide home health care. Home Health Care Service Quarterly, 1, Pearlin, L. I., ullan, J. T., Semple, S. J., & Skaff,.. (1990). Caregiving and the stress process: An overview of concepts and their measures. The Gerontologist, 30, Pinkston, E.., & Linsk. N. L. (198). Behavioral family intervention with impaired elderly. The Gerontologist, 2, Pruchno, R. A., ichaels, J. E., & Potashnik, S. L. (1990). Predictors of institutionalization among Alzheimer disease victims with caregiving spouses. Journal of Gerontology: Social Sciences, 5, S259-S266. Radloff, L. S. (1977). The CES-D scale: A self-report depression scale for research in the general population. Applied Psychological easurement, 1, Reisberg, B. (1988). Functional assessment staging (FAST). Psychopharmacology Bulletin, 2, Roberts, R. E. (1980). Reliability of the CES-D scale in different ethnic contexts. Psychiatry Research, 2, Rockefeller, J. D., IV. (1991). A call for action: The Pepper Commission's blueprint for health care reform. Journal of the American edical Association, 265, Scharlach, A. E., & Boyd, S. (1989). Caregiving and employment: Results of an employee survey. The Gerontologist, 29, 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, Seltzer,.., Ivry, J., & Litchfield, L. C. (1987). Family members as case managers: Partnership between the formal and informal support networks. The Gerontologist, 27, Skaff,.., & Pearlin, L. (1992). Caregiving: Role engulfment and the loss of self. The Gerontologist, 32, Stevens, J. (1996). Applied multivariate statistics for the social sciences. ohawk, NJ: Lawrence Erlbaum Associates. Stone, R., Cafferata, G. L., & Sangl, J. (1987). Caregivers of the frail elderly: A national profile. The Gerontologist, 27, Teri, L., Rabins, P., Whitehouse, P., Berg, L., Reisberg, B., Sunderland, T., Eichelman, B., & Phelps, C. (1992). anagement of behavior disturbance in Alzheimer disease: Current knowledge and future directions. Alzheimer Disease and Associated Disorders, 6(2), Teri, L., Truax, P., Logsdon, R., Uomoto, J., Zarit, S., & Vitaliano, P. P. (1992). Assessment of behavioral problems in dementia: the revised memory and behavior problems checklist. Psychology and Aging, 7, Tombaugh, T. N., & clntyre, N. S. (1992). The ini-ental State Examination: A comprehensive review. Journal of the American Geriatrics Society, 0, Toseland, R. W., Labrecque,. S., Goebel, S. T., & Whitney,. H. (1992). An evaluation of a group program for spouses of frail elderly veterans. The Gerontologist, 32, Winogrond, I. R., Fisk, A. A., Kirsling, R. A., & Keyes, B. (1987). The relationship of caregiver burden and morale to Alzheimer's disease patient function in a therapeutic setting. The Gerontologist, 27, The Gerontologist