Caring for a Family Member with Dementia: Evidence from a Cross Sectional Comparative Study on Caregiver Burden and Psychological Well-being

3. Results

Care recipients’ characteristics are shown in Table 1. It can be seen that the dementia and non dementia sub-samples were similar with respect to most of the socio-demographics taken into consideration by the study: gender, age, educational level, marital status, personal and household net monthly income. Statistically significant differences between the two sub-samples emerged only in terms of functional disability and number of reported chronic conditions: while dementia care recipients showed on average a more severe functional disability (see Table 2 for details at the item level), non dementia care recipients reported a higher number of chronic conditions (see Table 3 for details).

Table 1: Care recipient socio-demographic variables by dementia status

Table 2: Care recipient ADLs/IADLs ability by dementia status

Table 3: Care recipient chronic conditions by dementia status

Caregivers’ characteristics are reported in Table 4. Here, dementia and non dementia caregivers were comparable on all the variables related to the areas of the socio-demographics, caregiving circumstances and health status. This indicated that we obtained two highly homogeneous samples of caregivers with respect to these variables.
After this preliminary description of the dementia and non dementia dyads, five multiple regression analyses were run using the measures of caregiver burden, stress and psychological well being as criteria, and the variables on which dementia and non dementia dyads were found to differ (control variables) as well as care recipient dementia status as predictors. In Step 1, for all regression models we entered the control variables (care recipient functional disability, i.e. EASY care total score, and number of reported chronic conditions). In Step 2 of the analyses we entered the dementia status variable (1=dementia; 0=non dementia). The results for the five regression models are shown in Table 5.

Table 4: Caregiver-related variables (socio-demographics, caregiving circumstances and health status) by patient’s dementia status

Table 5: Multiple regression analyses of caregiver’s outcome measures regressed on care recipient dementia status (controlling for care recipient functional status and care recipient number of reported chronic conditions)^

Together, the control variables explained a significant portion of variance only in the case of the caregiver objective burden (13.2%). Parameters’ estimates indicated that the care recipient functional status was the only factor to affect this outcome measure. The standardised regression coefficient (β) was positive, suggesting that the higher the patient’s functional disability, the higher the caregiver’s objective burden. Step 1 regression model was not significant for any of the other outcome measures. It should be noted that we didn’t expect the control variables to explain a high portion of variance on the caregivers outcome measures. This is because most of the variables (caregiver gender, age, education, income etc.) which are known to influence these outcome measures were not included in the regression models, since they were found to be equally distributed in the dementia and nondementia subsamples. When the dementia status variable was included (Table 3), it added a modest but significant portion of explained variance in the case of caregiver stress (3.5%) and depression (4,0%). This indicated that, even after controlling for potentially confounding variables, the care recipient dementia status significantly predicted caregiver stress and depression. Since the β values were in both cases positive, the results indicated that dementia caregivers reported higher levels of stress and depression than non dementia caregivers. After this, to seek evidence for our second hypothesis (i.e., that spouse caregivers would report worse consequences from caregiving than non spouse caregivers), we carried out on the overall sample (N=413) a multivariate analysis of variance (MANOVA; dependent variables: caregiver subjective burden, stress, anxiety and depression; grouping variable: caregiver relationship to patient), followed by Tukey HSD post hoc tests. The mean values for the different groups of caregivers are shown in Table 6.

Table 6: Comparison of Spouses vs. Son/Daughters vs. Other relatives on measures of burden and emotional well-being

The overall MANOVA was statistically significant – Hotelling’s T: F(10.736)=4.08; p < .001 — and so were all the follow up tests: objective burden: F(2.373)=3.04, p<.05; subjective burden: F(2.373)=4.8, p<.01; stress: F(2.373)=11.42, p<.001; anxiety: F(2.373)=5.61, p<.01; depression: F(2.373)=15.3, p<.001. The post hoc analysis revealed homogeneous subsets of means, which showed that the three groups (Spouses: SP; Sons/Daughters: SD; Other relatives: Other) did not differ significantly on the objective burden measure. In terms of subjective burden, stress and anxiety spouses were not found to differ significantly from sons/daughters, the differences being mainly between spouses and sons/daughters as a group vs. other relatives (Subjective burden: SP = S/D, SP > Other; S/D > Other; Stress: SP = S/D, SP > Other, S/D > Other; Anxiety: SP=S/D, SP > Other, S/D = Other). However, results did show that spouses reported significantly higher levels of depression than sons/daughters, who in turn were significantly more depressed than other relatives (SP > S/D > Other). This finding replicated previous research in this area.

Pages: 1 2 3

Tags: ,