Evaluation of health impacts of a disability-inclusive graduation programme among ultra-poor people with disabilities in Uganda: secondary analysis of a cluster randomized trial.

Background: People with disabilities experience significant health inequalities and barriers to healthcare access globally. While poverty alleviation interventions show promise for improving health outcomes, evidence specifically for people with disabilities remains limited. This study evaluated the effectiveness of a disability-inclusive graduation (DIG) programme on health outcomes among ultra-poor people with disabilities in Uganda. Methods: We conducted a two-arm, parallel cluster-randomized controlled trial in four districts of Northern Uganda. Clusters were randomly assigned to either the DIG intervention (96 clusters) or control group (89 clusters). This analysis focused on households with people with disabilities, as identified by the Washington Group Short Set questions and verified by BRAC programme managers. Households in treatment clusters received up to 18 months DIG intervention between December 2020 and June 2022, combining asset transfers, cash support, skills training, financial inclusion activities, and disability-specific services including rehabilitation and assistive devices. The primary outcome was experience of illness/injury in the past 3 months, assessed at both first follow-up (immediately post-intervention) and second follow-up (about 16 months post-intervention), with secondary outcomes including unmet health needs, mental health status, unmet assistive product needs, and healthcare expenditure. Effects were estimated using linear mixed-effects regression or generalized estimating equations, reporting minimally-adjusted and fully-adjusted mean differences (FAMD) or odds ratios (FAOR) with 95% CIs. The trial was registered with RIDIE (RIDIE-STUDY-ID-626008898983a) and ISRCTN (ISRCTN-78592382). Findings: At baseline, 691 participants (370 intervention, 321 control) were included. The DIG intervention did not significantly impact overall illness/injury prevalence at either first follow-up (41.18% vs 45.86%, FAOR 0.84, 95% CI 0.58-1.22) or second follow-up (55.65% vs 53.98%, FAOR 1.07, 95% CI 0.74-1.56). However, the intervention demonstrated a progressively strengthening effect on reducing unmet health needs, from marginal improvement immediately post-intervention (FAOR 0.56, 95% CI 0.31-1.02, p = 0.06) to significant reduction at 16 months post-intervention (FAOR 0.4, 95% CI 0.22-0.71, p = 0.002). Notably, the intervention produced temporal potential shifts in disease patterns, with malaria showing contrasting trends between follow-up periods. Sex-differentiated effects emerged by second follow-up, with females in the intervention group experiencing fewer injuries (FAOR for interaction 0.17, 95% CI 0.04-0.74, p = 0.02) but more pain-related conditions compared to males (FAOR for interaction 2.43, 95% CI 1.05-5.59, p = 0.04), though these subgroup findings require replication in future studies. No significant differences were observed in mental health outcomes or health expenditure. Interpretation: This first randomized evaluation of a disability-inclusive graduation programme demonstrates that while economic empowerment alone may not reduce overall illness prevalence among people with disabilities, it can progressively improve healthcare access over time. The temporal evolution of effects and emerging sex-differentiated impacts highlight the need for sustained support and gender-sensitive approaches in future disability-inclusive poverty reduction programmes, with additional health-specific components to achieve broader improvements in health outcomes. Funding: PENDA, funded by the UK Foreign, Commonwealth and Development Office.