AI-driven Multimodal Representation Learning for Latent Mediation Structure Discovery of Socioeconomic Disadvantage, Psychosocial Factors, and Cardiometabolic Multimorbidity

The research examined how social disadvantage is associated with multiple diseases using artificial intelligence. The study included 20,804 participants from the All of Us Research Program with complete data. Researchers analyzed data from social, psychosocial, clinical, laboratory, behavioral, and genetic perspectives. The strongest indirect association was found between a dimension of socioeconomic disadvantage, psychosocial vulnerability, and cardiometabolic multimorbidity. Psychosocial vulnerability was characterized by poorer mental health, greater loneliness, lower social well-being, and lower health literacy. The resulting health outcome was associated with high blood pressure, diabetes, high cholesterol, obesity, chronic kidney disease, and heart disease. The research suggests that psychosocial vulnerability may contribute to the connection between social disadvantage and cardiometabolic multimorbidity.