Psychosocial problems, sexual violence victimization, and substance use among sexual and gender minorities in India: An intersectional multilevel analysis of individual heterogeneity and discriminatory accuracy (I-MAIHDA)

Sexual and gender minority (SGM) populations in India experience substantial psychosocial health burdens, which are likely patterned by intersecting axes of marginalized identities and positions within them. To characterize these intersectional inequalities and inform equitable intervention design, we investigated how psychosocial problems vary across combinations of SGM subgroup, sex work engagement, and socioeconomic status. We analyzed baseline data from the S3 Cohort Study of 799 men who have sex with men (MSM) and 630 transgender women in India. We employed intersectional Multilevel Analysis of Individual Heterogeneity and Discriminatory Accuracy (I-MAIHDA) using a Bayesian approach for six binary outcomes. Participants were nested within 16 social strata defined by SGM subgroup, sex work engagement, and socioeconomic status. We quantified between-stratum inequality (Variance Partition Coefficient, VPC), additive vs. interactive patterns (Proportional Change in Variance, PCV), and the discriminatory accuracy of intersectional strata in predicting outcomes (Area Under the Receiver Operating Characteristic Curve, AUC). All psychosocial problems were more prevalent among transgender women than MSM. Between-stratum inequality was strongest for depression (VPC = 9.0%; AUC = 0.66), followed by anxiety (VPC = 6.4%; AUC = 0.65), sexual violence victimization (VPC = 5.9%; AUC = 0.63), drug use (VPC = 5.5%; AUC = 0.62) and alcohol use (VPC = 4.7%; AUC = 0.62). For anxiety, problematic alcohol use, drug use, and sexual violence victimization, inequalities were driven predominantly by additive effects (PCV ∼ 62 to 96%). Depression showed primarily interactive effects (PCV = 43.5%; AUC = 0.66). Transgender women with indigenous identities who had low socioeconomic status and who engaged in sex work comprised the most marginalized stratum, with high adverse outcome prevalence. Health inequalities among SGM individuals are outcome-specific. Anxiety, problematic alcohol use, drug use and violence victimization outcomes follow cumulative disadvantage models with good discriminatory accuracy, while depression demonstrates interactive patterns, requiring targeted interventions.