| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||
RESEARCH ARTICLE |
Department of Educational Sciences and Psychology, Freie Universität Berlin, Germany.
FOR three reasons, we would like to thank Zarit, Stephens, Townsend, Greene, and Femia (2003)
for opening the discussion about the effectiveness of day care for caregivers in their response to our article "Evaluation of Geriatric Day Care Units: Effects on Patients and Caregivers" (Zank & Schacke, 2002). First, we discussed our results cautiously and did not conclude that day care is ineffective for caregivers. Yet, we do agree that there is a possibility of political misuse of our data. We certainly did not intend to support political decisions against day care. Second, the commentary gives us a chance to report our complex multimethodological approach and results that were beyond the scope of the original article (Zank & Schacke, 2002). Third, some of the issues raised could be due to differences between the American and the German health and care systems, which will be described later.
When we started our study in 1994, we were very much aware of the risk that standardized measures might not grasp certain effects that day care might have on caregivers. We, therefore, applied a multimethod approach. First, we chose the most prominent standardized measures at that time, namely, the Burden Interview (BI) and the Memory and Behavior Checklist (Zarit & Zarit, 1983/1990), and the Family-Caregiving Conflict and Job-Caregiving Conflict scales (Pearlin, Mullan, Semple, & Skaff, 1990). These quantitative data failed to show positive effects of day care on caregivers. We will give detailed reasons for our choice of instruments later (see Criterion 4).
Second, to examine the caregiver's stress, we developed a semistructured qualitative interview based on stress theory. This interview took approximately an additional 1–1.5 hr to administer. We conducted these qualitative interviews twice with caregivers of dementia patients only (N = 37, T1, and, after 9 months, T2 = 74 interviews). As the first step, we analyzed these qualitative interviews by using content-analytical techniques (Mayring, 1988). Then, to allow further statistical analyses, we developed response scales to measure the intensity of caregiving stress. Results from the analyses of variances of these data show that day care significantly reduced conflict between caregiving and job requirements, caregiving and family needs, and recreational constrictions. Initial results from these qualitative data were published in German (Schacke, 2001; 2002, Zank & Schacke, 2001, Zank, 2000), and an English article has been submitted (Schacke & Zank, 2002).
These results convinced us that day care is effective, not only for patients but also for caregivers; nevertheless, we have to accept that the results based on the standardized measures did not reflect this conclusion. Zarit and colleagues (2003) suspected several methodological flaws in our quantitative data that could explain our results. In the following, we will explain why we think that these methodologicial flaws did not exist. Finally, we will explain our interpretation of the results and consequences we drew.
Criterion 1: Use of Services on a Regular Basis for a Sustained Period of Time in the Treatment Group
Zarit and colleagues (2003) suggest that it was not clear whether day care was used on a regular basis for a sustained period of time. In our article, we stated that we only recruited caregivers and relatives who intended to use day care at least twice a week for at least 3 months. For the evaluation analyses, we included data only from those caregivers (relatives) who used services at least twice a week for 9 months. The mean number of visiting days was 3.4 days per week for at least 9 months, which, according to their commentary, Zarit and colleagues would accept as a sufficient therapeutic dosage.
In Germany, use of day care is subsidized by payments from care insurance. Accordingly, the number of days the care will be used is fixed in an agreement prior to the beginning of day care use and is obligatory, unless the contract is formally changed. Thus, a caregiver's or patient's use of day care is more carefully regulated than in the United States.
Criterion 2: Use of a Control Group Not Receiving Treatment
In their commentary, Zarit and colleagues (2003) questioned whether participants in our control group had had more alternative formal support or if the support they had received might have been sufficient to provide relief.
First, as we stated in our article, we accepted home-based community care in both groups because this type of support was also used in the treatment group by most of the caregivers in addition to day care. The home-based community care used in the entire sample was on average 3.8 hr per week. The mean level of use in the treatment group amounted to 3.2 hr; in the control group, it was 4.4 hr/week. This difference was not statistically significant, F(1, 41) = 0.49; p =.48).
It is important to note that this specific type of support is not comparable to day care because caregivers are not offered time away from the patient. As we described in our article, it covers assistance with clearly defined, basic self-care tasks such as washing, dressing, meals, or medication within a short time period.
Second, according to our data, caregivers in the control group spent 66.8 hr per week on caregiving tasks. Thus, one cannot assume that using 4 hr of paid help per week provided adequate support for the caregivers in the control group.
Criterion 3: The Two Groups Should Have No Significant Interventions Before Treatment
We followed the criterion of minimal use of comparable support before treatment by limiting our recruitment to those who had not received comparable support (e.g., day care, in-home respite) within the year previous to the treatment. Four patients in the treatment group and three in the control group had been in a hospital in the year prior to the first measurement, but all had been at home for at least 4 weeks before the intervention started.
Criterion 4: Use of a Theory-Driven Model for Evaluating Outcomes
To derive proximate indicators of perceived caregiving stress, we applied (a) the Burden Interview (Zarit & Zarit, 1983/1990), which measures subjective caregiving stress resulting from lack of privacy, restricted social activities, role conflicts, etc.; (b) the Memory & Behavior Problems Checklist (Zarit & Zarit, 1983/1990), which assesses depression and memory-related and disrupted behaviors in the care recipient; and (c) the subscales Family-Caregiving Conflict and Job-Caregiving Conflict from Pearlin and colleagues (1990). To assess whether the potential reduction of these dimensions of caregiving stress exerted a positive influence on caregivers' general well-being, we investigated caregivers' depression, self-esteem, and life satisfaction. The selection of outcome criteria is based on our sound knowledge of the intervention program under investigation and can be justified as follows: Adult day care is an intervention designed to provide time away from caregiving responsibilities on a regular basis. Thus, it can potentially enable the caregiver to spend a certain amount of time on personal needs such as regenerative and social activities or job and family obligations. As all institutions under investigation offered specialized services for dementia patients designed to stabilize and promote the patient's cognitive, social, and emotional resources, we also included the associated stress dimensions assessed by the Memory and Behavior Problems Checklist. Also, stress for the caregiver related to the patient's behavior problems might be alleviated by temporary separation from the patient. We selected depression, self-esteem, and life satisfaction as more distant outcome criteria because they are known to be affected by the caregiving situation (Alspaugh, Stephens, Townsend, Zarit, & Greene, 1999; Schulz, O' Brien, Bookwala, & Fleissner, 1995).
Although the measures we used did reflect dimensions that were likely to be responsive to use of day care, especially the BI also contained aspects that were less likely to be affected by the intervention (e.g., worries about the future), thus weakening the instruments' sensitivities to treatment effects. In our understanding, the main difficulties of our and other intervention studies are due to instruments that are too global and insensitive towards change. We, therefore, conducted the qualitative interviews described in the beginning as a supplementary analysis.
Three Other Differences
First, because of design constraints, we were not always able to conduct our baseline interviews prior to the time point at which the clients began using adult day care. We interviewed caregivers within the first 10 days, which does not differ much from the approach presented by Zarit, Stephens, Townsend, and Greene (1998) in which caregivers were interviewed within the first week.
Second, we wanted to evaluate geriatric day care and therefore had to include all patients. The day care units offer their services explicitly to both demented and nondemented patients. Nevertheless, 19 of the 20 day-care patients and 18 of the 22 control group persons suffered from dementia. This difference is not significant.
Third, we agree that a limitation of our study is the small sample size. This was because 50% of the day care users did not have a caregiver. The results from our qualitative data show, however, that even a small sample size can produce significant effects. The problem was not the sample size per se, but rather standardized instruments that were insensitive to change.
Conclusion
Given the limited effects found for the quantitative measures, we used qualitative data (1,175 pages of transcribed interviews) to develop a new, theory-driven questionnaire that includes different dimensions of strain and can describe different profiles of caregivers. In addition, this instrument also aims to be sensitive towards change and, therefore, provides a suitable measure for future intervention studies. This instrument was tested in a pilot study with 90 caregivers (Zank, Schacke, & Leipold, 2002). We have just finished our first measurement in a longitudinal study with 880 caregivers. This sample will allow us to test the instrument thoroughly and to collect prospective data on different interventions to further test the effects of day care for caregivers as we continue the cumulative research process.
 |
Acknowledgments |
|---|
Received for publication December 2, 2002. Accepted for publication December 17, 2002.
|
References |
|---|
 TOP References |
|---|
| ||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
