CAREGIVING AND SERUM BIOMARKERS: COMPARING EFFECTS ACROSS POPULATION-BASED AND CONVENIENCE SAMPLES

Several previous studies have compared biomarker indicators of stress, inflammation, and immune system functioning between family caregivers and noncaregivers. Most studies have used small convenience samples of specific caregiving subgroups (e.g., spouses of persons with dementia) and used poorly described noncaregiving comparison groups. Published reviews have noted these limitations and also documented clear inconsistencies in the findings, but definite conclusions have, nonetheless, been offered, and dysfunctional biomarker effects of caregiving are typically characterized as established facts. In our presentation, we will report meta-analyses that objectively summarize the previous findings and report new findings from the national, population-based, Reasons for Geographic and Racial Differences in Stroke (REGARDS) study. REGARDS includes a diverse sample of 2468 caregivers who had measures of C-reactive protein (CRP) and white blood cell count (WBC) obtained from serum samples. A matched sample of 2468 noncaregivers was identified using propensity-score matching techniques that balanced the two groups on 15 demographic, health history, and health behavior variables. Caregivers in the REGARDS study did not differ significantly from matched noncaregivers on the natural logarithm of CRP or on WBC (ps > .36), but the caregivers who reported high caregiving strain (17% of all caregivers) had significantly higher logCRP (M = 0.97) and WBC (M = 6.10) levels than less strained caregivers (Ms = 0.84 & 5.88, respectively; ps < .03). Differences between studies using convenience samples and those using population-based samples will be discussed, including the importance of using more rigorous methods to assemble suitable noncaregiving comparison groups.