JMIR Publications

ABSTRACT

Background:

Structural racism has long been identified as a driver of health disparities, including mental health. While multiple studies have assessed the effects of individual factors such as poverty and educational attainment, few have explored the cumulative impact of social determinants of mental health and how they predict racial disparities. Milwaukee County, which has a significant racial divide and socioeconomic variation, is an ideal locale for this investigation.

Objective:

To explore the association between structural racism and mental health disparities using geospatial analysis and deep learning

Methods:

A multi-step machine-learning pipeline was utilized, narrowing down a comprehensive list of health determinants to 12 crucial factors. These factors were then analyzed using deep learning techniques, culminating in the identification of three distinct mental health risk profiles in Milwaukee communities, revealing disparities in racial representation across these areas.

Results:

Smoking, poverty, insufficient sleep, lack of health insurance, sedentariness, and low educational attainment were the most defining factors. These determinants accounted for 95.11% of the variability in poor mental health in communities. The proportion of Black community members in high-risk areas was 2.23 times higher than expected without racial disparities.

Conclusions:

The findings suggest that structural racism shapes mental health disparities in Milwaukee County. Black community members are particularly affected, supporting the need for tailored interventions that address both individual and systemic determinants. This study underscores the value of employing geospatial and deep learning methods to better understand complex social factors influencing mental health, setting the stage for more targeted public health strategies.