A Bayesian approach to estimate changes in condom use from limited human immunodeficiency virus prevalence data

Area of Science:

• Epidemiology
• Biostatistics
• Mathematical Modeling

Background:

• Evaluating large-scale human immunodeficiency virus (HIV) interventions requires accurate behavioral data, often limited or biased.
• Estimating changes in condom use over time is crucial for assessing intervention effectiveness.

Purpose of the Study:

• To develop and apply a novel Bayesian inference methodology to estimate condom use time trends from HIV prevalence data.
• To assess the feasibility of evaluating HIV interventions using limited prevalence and behavioral data.

Main Methods:

• Employed a Bayesian inference methodology integrated with an HIV transmission dynamics model.
• Utilized particle Markov chain Monte Carlo (pMCMC) methods for estimation.
• Explored novel formulations for time-varying condom use parameters, including diffusion-driven trajectories and sigmoid curves.

Main Results:

• Numerical simulations demonstrated the method's ability to provide informative results on the amplitude of condom use increases during interventions.
• The approach showed good sensitivity and specificity in detecting changes in condom use.
• The method successfully evaluated a real-world HIV intervention using a small number of prevalence estimates.

Conclusions:

• This Bayesian approach offers a robust framework for estimating condom use trends and evaluating HIV interventions, particularly when data is scarce.
• The methodology is adaptable for similar applications in diverse public health contexts.
• The study highlights the potential of advanced statistical modeling in public health program evaluation.