Machine learning-based assessment of the built environment on prevalence and severity risks of acne

Abstract

Understanding the determinants of acne prevalence and severity is crucial for effective prevention and management of this dermatological condition. While urban interventions have long-lasting, far-reaching, and costly implications for health promotion, the associations between built environments (BEs) and acne need further investigation. To address this gap, our study utilizes a nationwide cross-sectional sample of 23,488 undergraduates from 90 campuses in China to conduct a comprehensive analysis. We examined the combined and specific contributions of BEs in relation to other domains of acne-related factors in acne development. By employing the optimal random forest model, our findings reveal that BEs collectively ranked as the second-largest contributors to both the overall prevalence of acne among all participants and the severity of acne in the affected individuals. Moreover, our analysis identifies curvilinear associations between acne and most BEs, underscoring the importance of incorporating BE considerations into the prevention, diagnosis, and management of acne.