Depression and psychiatric help-seeking as joint outcomes: a Bayesian multilevel analysis of household-level dependence.

Abstract

BACKGROUND: Depression and the utilization of psychiatric services (PSU) are complex mental health outcomes shaped by a wide range of individual, familial, and socioeconomic factors. However, within the context of Türkiye, limited attention has been devoted to the simultaneous modeling of these outcomes and to the systematic examination of their shared household-level determinants. METHODS: This study conducts a concurrent analysis of two binary outcomes—self-reported depression and receipt of professional psychiatric care within the past 12 months—using data from a nationally representative household survey. The analytical sample comprises 22,742 individuals nested within 10,028 households across Türkiye. A Bayesian multilevel bivariate logistic regression model was employed to account explicitly for the hierarchical structure of the data, with individuals nested within households. Household-level random intercepts were specified for both outcomes, and the correlation between the latent processes underlying depression and PSU was directly estimated. Weakly informative priors were adopted, and model estimation was performed using the No-U-Turn Sampler (NUTS), an efficient variant of the Hamiltonian Monte Carlo algorithm. RESULTS: The results reveal pronounced clustering of both depression and PSU at the household level. A strong positive correlation between household-level random effects indicates that individuals experiencing depression are substantially more likely to utilize psychiatric services within the same family context. At the individual level, men are significantly less likely than women to report depressive symptoms or to seek psychiatric care. Tobacco and alcohol use are positively associated with depression risk, whereas adequate rest and healthier lifestyle behaviors exhibit protective effects. Increases in body mass index (BMI) are associated with a higher likelihood of depression. In contrast, active employment and higher household income are linked to reduced depression risk. Additionally, a range of sociodemographic and economic characteristics exert significant effects on both depression and PSU. CONCLUSIONS: These findings demonstrate that depression and psychiatric service utilization (PSU) are shaped not only by individual-level characteristics but also by shared familial and structural factors. The results underscore the need for mental health policies that move beyond narrowly individual-centered frameworks and instead adopt holistic strategies that incorporate household dynamics and broader socioeconomic conditions.