Inside and Beyond the Capital: Latent Profiles of Gay Men in Peru

Authors

DOI:

https://doi.org/10.34190/icgr.9.1.4610

Keywords:

machine learning, Gay men, Artificial intelligence (AI), Regional studies

Abstract

Research on sexual minority health in Peru has largely focused on the capital city, overlooking how geographic inequalities shape heterogeneity among gay men living outside Lima. This study addresses this gap by comparing latent sociodemographic and risk-related profiles of gay men residing in Lima and in Peru’s inland regions, where structural constraints and social exclusion may operate differently. Using data from the 2017 First Virtual Survey for LGBTIQ+ People in Peru, we analyze two subsamples: gay men living in inland regions (n=657) and those residing in Lima (n=1,528). The study applies a two-stage unsupervised analytical framework across both settings. Behavioral, social, and structural indicators are summarized through dimensionality reduction techniques, followed by the evaluation of multiple clustering algorithms. Optimal solutions are selected based on internal validity metrics, and feature relevance is examined using Random Forest–based importance measures to support interpretability while preserving cross-regional comparability. Results reveal marked differences in both the number and composition of latent profiles across geographic contexts. In inland regions, the optimal solution identifies two clusters primarily differentiated by age and employment status, with secondary differences in access to STI/HIV information and health services. In contrast, the Lima subsample yields three distinct clusters, indicating greater internal heterogeneity. The largest Lima cluster is dominated by young adults with lower employment levels and higher prevalence of mental health problems and suicidal ideation, while older clusters display greater socioeconomic stability and lower psychosocial vulnerability. From a gender research perspective, these findings demonstrate that sexual minority inequalities are shaped less by individual risk behaviors than by gendered structural and spatial contexts. Geographic location influences not only levels of vulnerability but also the internal configuration of risk and protection among gay men in Peru. The study underscores the need for territorially differentiated research and interventions that move beyond capital-centered and behavior-focused frameworks.

Author Biographies

Alejandro Aybar-Flores, Universidad del Pacífico

Alejandro Aybar Flores is a project analyst and research assistant specializing in machine learning, statistics, and data science applications in social and health research. He holds degrees in Information and Business Engineering and has published on financial inclusion, public health, and behavioral modeling using advanced computational and econometric methods.

Elizabeth Espinoza-Portilla, Universidad de San Martín de Porres

Medical Doctor with a Master’s in Public Policy Studies (Ortega y Gasset Institute, Complutense University of Madrid), a Master’s in Demography from Cayetano Heredia University, and a PhD in Government and Public Policy. RENACYT researcher, graduate professor at ESAN University, and CEO and Founder of Conecta Salud SAC.

Downloads

Published

2026-04-25