Structural Patterns of Intersectional Stigma Among Sexual and Gender Minorities: A Psychometric Network Analysis of the All of Us Cohort

People with HIV (PWH) have high rates of co-occurring mental health disorders (MHDs), yet patterns of comorbidity remain unclear. Mapping these networks may identify central conditions whose treatment could disrupt broader comorbidity clusters and improve mental health outcomes in this vulnerable population. To map comorbidity networks of mental health disorders among people with HIV and identify diagnoses with the highest co-occurrence with other disorders. Cross-sectional psychometric network analysis of electronic health records and survey data from the National Institutes of Health All of Us Research Program (Release 8; participants enrolled May 6, 2018–October 1, 2023). This study used electronic health records and survey data from the All of Us Research Program, a large, nationwide, population-based cohort with intentional oversampling of underrepresented groups across more than 700 US health care institutions Participants in this study include those with a validated computational phenotype for identified HIV. Of 6,664 participants with confirmed HIV, 5,868 had electronic health data; 3,078 with at least 2 encounters bearing mental health disorder ICD-10 codes comprised the analytic cohort. Participants were not exposed to any medications or treatments in this study. As part of the All of Us Research Program, participants are regularly included in the data collection and dissemination process and have access to all publicly available data. No identifying information is included in this study. We estimated comorbidity networks that include mental health disorder co-occurrence based on ICD-10 codes, with connections between disorders representing conditional log-odds associations. Centrality was measured using strength and expected influence, and clusters or communities of mental health disorders were detected using the Louvain algorithm. Sensitivity analyses relaxed the mental health disorder criterion from at least 2 to at least 1 code, using nonparametric bootstrapping. Participants are predominantly aged at 50 years or above (69.2%). This sample includes 48% Black/African American, 78% non-Hispanic white, and 64% male participants, with 91% with health insurance. Common mental health disorders included tobacco-related disorders (49%), major depressive disorder-single episode (45%), other anxiety disorders (45%), recurrent major depressive disorder (28%), adjustment/stressor-related disorders (25%), and substance-related disorders (alcohol 21%, cocaine 20%, other psychoactive substances 19%). The comorbidity network revealed 2 principal clusters: (1) affective/anxiety-related disorders, and (2) substance use/psychotic-related conditions. Five diagnoses exhibited the highest centrality via strength and expected influence: other psychoactive substance, cannabis-related, personality, cocaine-related, and recurrent depressive disorders. Among people with HIV, comorbidity networks are organized around tightly connected substance-related and affective/anxiety disorder clusters. A small set of highly central disorders may serve as strategic targets for integrated screening and treatment. Addressing these hubs could disrupt broader comorbidity patterns and improve mental health care for people with HIV. • Examines intersectional discrimination using network analysis • Discrimination patterns differ by gender identity status • Race, ancestry, and religion are structurally central • Edge differences persist after multiple-testing correction • Findings generate hypotheses for future intervention research