Evaluating potential mediators for the impact of a family-based economic intervention (Suubi+Adherence) on the mental health of adolescents living with HIV in Uganda

Abstract

Introduction: Many adolescents living with HIV in sub-Saharan Africa (SSA) experience poverty and have access to limited resources, which can impact HIV and mental health outcomes. Few studies have analyzed the impact of economic empowerment interventions on the psychosocial wellbeing of adolescents living with HIV in low resource communities, and this study aims to examine the mediating mechanism(s) that may explain the relationship between a family economic empowerment intervention (Suubi + Adherence) and mental health outcomes for adolescents (ages 10-16 at enrollment) living with HIV in Uganda.

Method: We utilized data from Suubi + Adherence, a large-scale six-year (2012-2018) longitudinal randomized controlled trial (N = 702). Generalized structural equation models (GSEMs) were conducted to examine 6 potential mediators (HIV viral suppression, food security, family assets, and employment, HIV stigma, HIV status disclosure comfort level, and family cohesion) to determine those that may have driven the effects of the Suubi + Adherence intervention on adolescents' mental health.

Results: Family assets and employment were the only statistically significant mediators during follow-up (β from -0.03 to -0.06), indicating that the intervention improved family assets and employment which, in turn, was associated with improved mental health. The proportion of the total effect mediated by family assets and employment was from 42.26% to 71.94%.

Conclusions: Given that mental health services provision is inadequate in SSA, effective interventions incorporating components related to family assets, employment, and financial stability are crucial to supporting the mental health needs of adolescents living with HIV in under-resourced countries like Uganda. Future research should work to develop the sustainability of such interventions to improve long-term mental health outcomes among this at-risk group.

Keywords: Adolescents; Depression; Economic intervention; HIV; Mental health; Structural equation model; Sub-saharan Africa; Uganda.

Fig. 1.

Structural equation model at 24-month follow-up. Note: 1. Entries marked bold are statistically significant standardized coefficients. 2. a = effect of the primary predictor on the mediator; b = effect of the mediator on the outcome; c’ = the direct effect of the primary predictor on the outcome; c = the total effect of the primary predictor on the outcome. 3. The model adjusts for potential covariates: age, gender, and primary caregiver 4. This figure only presents the unstandardized coefficients for a, b, c’, and c paths. All other unstandardized coefficients and standardized coefficients produced by the model are presented in Table 2.

Fig. 2.

Structural equation model at 36-month follow-up. Note: 1. Entries marked bold are statistically significant standardized coefficients. 2. a = effect of the primary predictor on the mediator; b = effect of the mediator on the outcome; c’ = the direct effect of the primary predictor on the outcome; c = the total effect of the primary predictor on the outcome. 3. The model adjusts for potential covariates: age, gender, and primary caregiver 4. This figure only presents the unstandardized coefficients for a, b, c’, and c paths. All other unstandardized coefficients and standardized coefficients produced by the model are presented in Table 2.

Fig. 3.

Structural equation model at 48-month follow-up. Note: 1. Entries marked bold are statistically significant standardized coefficients. 2. a = effect of the primary predictor on the mediator; b = effect of the mediator on the outcome; c’ = the direct effect of the primary predictor on the outcome; c = the total effect of the primary predictor on the outcome. 3. The model adjusts for potential covariates: age, gender, and primary caregiver 4. This figure only presents the unstandardized coefficients for a, b, c’, and c paths. All other unstandardized coefficients and standardized coefficients produced by the model are presented in Table 2.