Advancing caregiver research: weighing efficacy and feasibility of interventions.

D periodic questioning of the merits and utility of the vast number of publications focused on family caregivers by key scholars in the field (e.g., George, 1990; Zarit, 1989), the number of studies of caregiving and family caregivers has not diminished. A review of the contents of The Journals of Gerontology and The Gerontologist in just the past three years reveals a total of 61 articles concerned with caregiving issues. When attention is given to the study designs and data analysis techniques reported in these articles, there is a clear indication that the work in this area has advanced beyond a natal stage to a more sophisticated adolescence. When, however, attention is given to the research questions most frequently addressed, there is reason to question the advancement of this field of study. A large majority of research on caregiving continues to use cross-sectional data to describe the prevalence of various family members as caregivers, differences among family members and cultural groups in the patterns of care provided, and the consequences (or lack of consequences) of such care for various family members. A disproportionate emphasis has been placed on psychological distress and caregiver burden, while studies of the ameliorative effects of existing or potential support services have received minimal attention. The article by Bass, Noelker, and Rechlin (1996) in this issue is a welcome contribution to a very small body of literature that has lacked conceptual and methodological sophistication. Given the pragmatic, long-term care policy issues that have served as the impetus for research in this arena, the dearth of intervention and service-impact studies might be viewed as an enigma. I would argue, however, that progress in this area could not be made without the cumulative contributions of the descriptive studies of caregiver prevalence, care patterns, and care consequences conducted over the past decade. This body of literature has honed our understanding of the diversity of caregiving contexts by extending studies to more heterogenous populations and utilizing more sophisticated multivariate analyses. Consequently, we have come to understand that the apparent discrepancies in findings regarding the prevalence and consequences of caregiving reflect the complexity and diversity of caregiving contexts (e.g., Aneshensel et al., 1995) and the critical need to carefully define the phenomenon of interest (Stone, 1991). Clearly, there is a convergence of work concluding that there is no single, generic caregiver role; rather, caregiving is a role that emerges from prior role relationships and is influenced by values, beliefs, and life circumstances of an individual caregiver (Aneshensel et al., 1995; Pearlin, 1992). Equally important, there is a growing recognition that caregiving is a dynamic process that unfolds over time and can be likened to a career that can vary in quality and length (Montgomery and Kosloski, 1994). Unfortunately, the documentation of great diversity in caregiving contexts tends to mask the consistency of patterns that are emerging in the literature. To date, only a few efforts have been made to synthesize the literature (e.g. Knight, Lutzky, and Macofsky-Urban, 1993; Schulz et al., 1995) and thereby provide a sense of cohesiveness to a seemingly discrepant body of literature. Future research on caregiving issues could benefit from meta-analyses and literature syntheses that would more clearly reveal the consistency of patterns in caregiving contexts and careers. The well-researched complexity of caregiving contexts has two implications for studies concerned with the development and assessment of caregiver support services and interventions. First, an understanding of the complexity of caregiving should sensitize scholars to the conceptual and methodological limitations inherent in previous studies of support services and raise questions about the accuracy of disappointing negative findings regarding the impact of interventions (e.g., Lawton, Brody, and Saperstein, 1989; Montgomery and Kosloski, 1995). Second, the diversity of caregiving situations should prompt scholars to give adequate attention to efficacy and feasibility when undertaking new investigations of the impact of support services or interventions intended to help caregivers. Study designs that do not acknowledge the complexity and diversity of caregiving contexts are likely to suffer from Type II errors. Without targeting services to appropriate segments of the caregiver population, and without attention to dosage, there is little chance that support services will have an impact sufficiently large and dispersed to be detected by measures now available to social scientists. Clearly, the support needs of caregivers are not uniform, but are influenced by gender, generation, culture, family histories, and duration of caregiving. Nor are these needs for support stable over time. Consequently, there is minimal likelihood