| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||
RESEARCH ARTICLE |
1 The Pennsylvania State University, University Park.
2 Kent State University, Ohio.
3 Case Western Reserve University, Cleveland, Ohio.
4 Administration on Aging, Washington, District of Columbia.
THE Zank and Schacke (2002)
study makes a valuable and original contribution by demonstrating that adult day care can improve functioning among its clients. At the same time, the authors draw conclusions about a lack of benefit to family caregivers of those clients that do not appear warranted because of limitations in the study. This is an important consideration both for scientific reasons and because policy makers may seize on null findings as a way of justifying low levels of support for community-based programs.
The evaluation of adult day care and other community-based services is a difficult enterprise with many potential threats to internal validity. There is a long history of reports showing little or no effects of respite services, but which on closer examination turn out to have substantial methodological problems that seriously compromised the test of their effectiveness (for reviews see Zarit, 1990; Zarit & Leitsch, 2001). Our research group has described these problems in past work and incorporated research strategies in our own study of adult day care that minimized the problems and demonstrated significant amounts of stress reduction for family caregivers who use adult day care on a regular and sustained basis (Zarit, Stephens, Townsend, & Greene, 1998). In framing their study, Zank and Schacke cite our work and state that their study used the same criteria we had to assure an adequate test of adult day care. Their study, however, may have varied considerably on four of the six criteria they say they have adopted from our work, and these differences may account for the null findings. The specific criteria where their study may have differed from ours are examined.
Criterion 1: Use of Services on a Regular Basis for a Sustained Period of Time
We felt that past work (e.g., Lawton, Brody, & Saperstein, 1989) had failed to show treatment effects because caregivers did not receive adequate amounts of service. There was often no consideration of what was a therapeutic dosage, and whether it was maintained over time. Consequently, we set a minimum criterion for service use and eliminated anyone who did not use this amount (Zarit et al., 1998). By contrast, the only information presented by Zank and Schacke (2002) about service use is that day care clients or their caregivers stated at baseline that they intended to use day care at least twice a week for 3 months, but not if they actually used services at that level throughout the evaluation period.
Criterion 2: Use of a Control Group Not Receiving Treatment
The effects of adult day care on caregivers are likely to come about through providing relief from ongoing care demands. Programs such as in-home respite may provide relief in a similar manner to adult day care (Berry, Zarit, & Rabatin, 1991). As a result, a control group that uses in-home respite may not differ from caregivers using adult day care for their relative. For that reason, our study limited participation in the control group to caregivers who used none or only small amounts of comparable services (Zarit et al., 1998). In Zank and Schacke (2002), the control group was receiving regular in-home service in the morning and evening. Zank and Schacke argued that this was an ideal control group because the day care sample was receiving similar help. No information was provided on how much help the two groups were actually receiving. If controls received more hours of in-home help than the day care group, then that might be functionally equivalent to the relief that caregivers receive through adult day care. Furthermore, even if the amount of home help received by both groups was the same, that amount might be sufficient to provide relief to family caregivers in the home-help-only condition. In other words, comparing two groups who both receive some amount of treatments that have similar effects does not yield a clear answer about possible benefits of either program.
Criterion 3: The Two Groups Had No Significant Intervention Before Treatment
We thought it was critical that adult day care produce a significant change in the daily routines of caregivers for it to lead to reductions in care-related strains. If caregivers were receiving regular in-home respite and simply replaced that with adult day care, it seemed unlikely that there would be much difference in their experience. For that reason, our study did not include anyone who was receiving more than a minimal amount of comparable services prior to enrolling in adult day care (Zarit et al., 1998). Zank and Schacke (2002) imply that they followed this procedure, but they do not state explicitly if that actually was the case.
Criterion 4: Use of a Theory-Driven Model for Evaluating Outcome
Both Zank and Schacke (2002) and our group used the same framework (Pearlin Mullan, Semple, and Skaff's [1990] stress process model), to guide our evaluations. The measures used by Zank and Schacke, however, do not appear to reflect closely those dimensions of the stress process that are likely to be responsive to this type of intervention. This is a common problem in respite studies, where dependent measures are selected without careful consideration of the mechanisms by which an intervention is likely to lead to change (Zarit, 1990; Zarit & Leitsch, 2001). Constructs used by Zank and Schacke, such as "life satisfaction," "self-esteem," "burden," and "family conflict," have many potential determinants beyond the stressors likely to be affected by adult day care use. By contrast, our study used Pearlin and colleagues' model to identify those specific aspects of the stress process that adult day care might alter. We identified that the major change that occurs with adult day care is giving caregivers time away from their relative, which reduces exposure to primary objective stressors (activity of daily living disability, behavior problems). We hypothesized that this reduction in exposure would have its greatest effect on caregivers' immediate subjective experience of the time, effort, and energy involved in giving care, which represents primary subjective stressors in Pearlin and colleagues' framework. These expectations were borne out in our study (Zarit et al., 1998).
Other Differences
Five other differences between the Zank and Schacke (2002) study and our work may be relevant. First, Zank and Schacke allowed for a longer interval (up to 2 weeks) than we did between starting day care and when baseline measures were administered. As they noted, many caregivers reported that they had experienced changes in their subjective strain beginning on the first day of use, that is, prior to the completion of the baseline interview.
Second, the studies differed in the composition of their samples. Our study focused only on caregivers of day care clients with dementia, whereas Zank and Schacke (2002) included clients both with and without dementia. Dementia produces greater strain on caregivers than typically found with noncognitive disorders (Ory, Hoffman, Yee, Tennstedt, & Schulz, 1999). It may be that some of the caregivers of nondemented clients in the Zank and Schacke study were not experiencing much strain and, therefore, could not show a reduction in strain as a result of the treatment.
The third difference is that we limited our sample to caregivers who had regular involvement with the day care client, and who would thus be likely to experience the kind of strains likely to be relieved by adult day care. Zank and Schacke (2002) do not indicate if they used a similar inclusion criterion.
Fourth, the studies were conducted in different countries. We cannot rule out the possibility that even an optimally designed study in Germany might have reached similar conclusions, perhaps because of variations in caregiving context and values, compared with the United States.
Fifth, perhaps the most serious limitation in the Zank and Schacke (2002) article is the small sample size. Although the sample size for testing the effects of adult day care on clients is adequate (
in the treatment group; 
in the control group), it drops to 20 and 22 for the analysis of caregivers. By contrast, our study had a total sample of 324 for a comparison after 3 months in day care and 193 for a 1-year comparison. In addition, the method of handling missing data used by Zank and Schacke (mean substitution) would be likely to shrink the variance and increase the probability of Type II error, especially in a small sample.
In conclusion, we believe that Zank and Schacke (2002) did not adequately test their hypotheses about the effects of adult day care on family caregivers. As past experience has shown, null findings can be taken as an indication that services do not work (e.g., Callahan, 1989) and can become part of a self-fulfilling expectation among service providers and policy makers toward helping older people. The issues raised by the Zank and Schacke study underscore the importance of developing field trials that are funded and designed at a sufficient level to provide an adequate test of the benefits of interventions.
 |
Acknowledgments |
|---|
Received for publication July 16, 2002. Accepted for publication October 28, 2002.
|
REFERENCES |
|---|
 TOP REFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  E. E. Femia, S. H. Zarit, M. A. P. Stephens, and R. Greene Impact of Adult Day Services on Behavioral and Psychological Symptoms of Dementia Gerontologist, December 1, 2007; 47(6): 775 - 788. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
