Binary regression analysis with network structure of respondent-driven sampling data

TL;DR

This paper introduces a binary regression model that accounts for network dependence in respondent-driven sampling data, improving prevalence estimation in hard-to-reach populations like men who have sex with men in Brazil.

Contribution

It proposes a novel binary regression approach incorporating RDS network structure via a latent random effect, enhancing analysis of complex sampling data.

Findings

Effective estimation of HIV and Syphilis prevalence in RDS data

Model accounts for network dependence in binary responses

Application to real-world data from Brazil

Abstract

Respondent-driven sampling (RDS) is a procedure to sample from hard-to-reach populations. It has been widely used in several countries, especially in the monitoring of HIV/AIDS and other sexually transmitted infections. Hard-to-reach populations have had a key role in the dynamics of such epidemics and must inform evidence-based initiatives aiming to curb their spread. In this paper, we present a simple test for network dependence for a binary response variable. We estimate the prevalence of the response variable. We also propose a binary regression model taking into account the RDS structure which is included in the model through a latent random effect with a correlation structure. The proposed model is illustrated in a RDS study for HIV and Syphilis in men who have sex with men implemented in Campinas (Brazil).