Loneliness, depressive symptoms, and life satisfaction among very old Mexican American adults: a repeated-measures GEE mediation

OBJECTIVE: To examine the concurrent association between loneliness and life satisfaction across survey waves and assess whether depressive symptoms statistically account for this association among very old Mexican American adults. METHODS: Data were drawn from three waves of the Hispanic Established Populations for the Epidemiologic Study of the Elderly (H-EPESE). The analytic sample included 944 person-wave observations from adults aged 82 years and older, of which 892 from 596 individuals were included in multivariable analyses after excluding missing data. Analyses were restricted to self-respondents. Ordinal logistic generalized estimating equation (GEE) models with a cumulative logit link accounted for repeated measures and adjusted for demographic, cognitive, health, and financial factors. Indirect effects evaluated the role of depressive symptoms in the association between loneliness and life satisfaction. RESULTS: < 0.001). The indirect effect through depressive symptoms was statistically significant (a × b = -0.48, 95% Monte Carlo CI [-0.70, -0.30]). Better self-rated health, absence of pain during standing or walking, lower financial strain, and higher cognitive function were also associated with higher life satisfaction. CONCLUSIONS: Loneliness is strongly associated with lower life satisfaction, with depressive symptoms partially and modestly accounting for this association. Interventions addressing both may improve well-being in very old adults.