Effect of family caregiving on depression outcome among older European adults

Most caregiver health studies rely on observational data and traditional regression methods, which fail to account for time-varying confounding, limiting causal inference. This analysis uses inverse probability weighting (IPW) to appropriately account for time-varying confounding in caregiving-depression relationship.

Using seven waves (2004-2019) of the Survey of Health, Ageing and Retirement in Europe, we estimated the effect of caregiving on probable depression (having 4+ symptoms) comparing daily and some caregiving to no caregiving. We accounted for censoring and potential measured confounding by time-fixed covariates (gender, number of children, country, and education) and time-varying covariates (age, employment, marital status, income, physical limitations, psychiatric medication, receiving help, previous caregiving, and previous depression) using IPCW and IPTW. The product of the two weights was applied to a marginal structural model to obtain the causal estimand on the prevalence difference scale. Confidence intervals were derived from bootstrapping.

Among 36 346 participants and 67 800 person-waves, compared to no caregiving in the last year, daily caregiving was associated with a 6.7 percentage point (95 % CI: 4.8 %, 8.6 %) increase in the prevalence of probable depression, after accounting for time-fixed and time-varying covariates. Some caregiving was not associated with probable depression (PD = 0.5, 95 % CI = -0.8 %, 1.8 %).

Our results support existing findings that high levels of caregiving may increase the prevalence of probable depression, while lower levels of caregiving do not. Accurate documentation of the relationship between caregiving and health outcomes is foundational in creating evidence-based policies to support healthy aging.