Social
Intersectional mental health disparities among sexual and gender minority adults: an I-MAIHDA analysis of the All of Us Research Program Junjie Lu* Junjie Lu Lu Stanford University
Background: Sexual and gender minority (SGM) adults experience elevated mental health burden, yet traditional modeling assumes identity categories (sex assigned at birth, sexual orientation, gender identity) operate independently. Intersectional Multilevel Analysis of Individual Heterogeneity and Discriminatory Accuracy (I-MAIHDA), which fits mixed-effects models with intersectional strata as random effects, can reveal how multiple marginalized identities interplay to shape health outcomes beyond additive effects.
Methods: We analyzed 60,769 participants from the NIH All of Us Research Program who completed the Emotional Health Survey. Using Bayesian multilevel models, we examined depression (PHQ-9≥10), anxiety (GAD-7≥10), and probable PTSD (PCL-6≥14) across 16 intersectional strata defined by sex assigned at birth (male, female), sexual orientation (heterosexual, bisexual, gay/lesbian, other sexual minority), and gender identity (cisgender, gender minority). We compared traditional logistic regression and I-MAIHDA models.
Results: The highest prevalence was observed among Female-Other sexual minority-Gender minority individuals for depression (55.8%) and PTSD (60.6%), and among Female-Bisexual-Gender minority individuals for anxiety (49.3%). Male-Heterosexual-Cisgender individuals showed the lowest prevalence (13.6%, 9.5%, 13.3%, respectively). Variance Partition Coefficients indicated 4.9-6.8% of outcome variance was attributable to stratum membership. Notably, 62% (depression), 69% (anxiety), and 43% (PTSD) of between-stratum variance was attributable to interactions. Traditional regression showed significant sex and sexual orientation coefficients, but I-MAIHDA revealed these reflect intersectional interactions rather than independent contributions.
Conclusions: I-MAIHDA reveals SGM mental health disparities are driven by intersectional interactions, supporting precision mental health interventions.
