The intraindividual dimensional structure and discriminant validity of subjective caregiver burden and depression

Caring for a family member with dementia places caregivers at risk for caregiver burden and increased depression (Schulz et al., 1995). Subjective caregiver burden (SCB) involves appraisals of entrapment, abandonment, resentment, and overload resulting from caregiving. Although this construct is well recognized and numerous measures exist, the dimensional structure of SCB remains unclear. Furthermore, the discriminant validity of SCB and depression has received only cursory attention using aggregate-level, cross-sectional methodologies (Fritz et al., 1997). Such studies ignore the implication of intra-individual variation over time in levels of caregiver's SCB and depression (Redinbaugh et al., 1995) for the validation of these constructs. Using a multivariate, replicated, single-subject, repeated measures design, this study tested five hypotheses regarding the intra-individual convergent and discriminant validity of SCB and depression.

Four women (mean age = 51 years) caring for a parent or (in-law) with dementia, and meeting diagnostic criteria for Major Depressive Disorder, completed daily ratings of depressive symptoms, SCB appraisals, automatic thoughts, and caregiving demands for an average of 88 days (range 73-103 days). Usable data was obtained from three caregivers. After statistically removing trend and cyclicity from items, data was analyzed using confirmatory dynamic factor analysis (CDFA; Molenaar, 1985; Mumma, 1999) and hierarchical dynamic time-series regression (DTSR; cf Hokanson et al., 1994).

Results indicated: (1) As hypothesized, the intra-individual dimensional structure of SCB was multi-factorial, consisting of from two to four oblique factors depending on caregiver. (2) The intra-individual dimensional structure of depression consisted of either two or three factors, with the two-factor models consistent with the Tripartite (Watson et al., 1995) or Bums and Eidelson's (1998) models. (3) Across caregivers, strong evidence of discriminant validity between the SCB and depression factors was found in combined CDFA models. (4) The prediction that intra-individual depression factors would be more temporally stable than SCB factors was partially supported. (5) As hypothesized, daily variability in SCB was generally predicted by caregiving context variables but not negative automatic thoughts while the reverse was generally true for daily variability in depression. The findings indicate that SCB appraisals and depressive symptoms represent distinct aspects of the caregiving experience.