A BAYESIAN HIERARCHICAL MODEL FOR COMBINING MULTIPLE DATA SOURCES IN POPULATION SIZE ESTIMATION

Area of Science:

• Epidemiology
• Biostatistics
• Public Health

Background:

• Accurate population size estimates are vital for effective HIV/AIDS interventions targeting key populations.
• Key populations, including sex workers, people who inject drugs, and men who have sex with men, are difficult to enumerate directly.
• Existing indirect estimation methods often yield conflicting results, necessitating robust reconciliation approaches.

Purpose of the Study:

• To develop and present a Bayesian hierarchical model for combining and reconciling multiple indirect estimates of key population sizes.
• To address systematic errors inherent in different data sources used for population estimation.
• To apply the model for estimating the size of people who inject drugs in Ukraine and evaluate its performance.

Main Methods:

• Development of a Bayesian hierarchical model incorporating data from multiple sources and years.
• Explicit modeling of systematic error within each data source.
• Application of the model to estimate the population size of people who inject drugs in Ukraine.

Main Results:

• The Bayesian hierarchical model successfully integrated diverse data sources to produce a reconciled estimate for the target population.
• The model's ability to account for systematic error improved the reliability of the final estimates.
• Evaluation provided insights into the contribution of individual data sources to the overall estimate.

Conclusions:

• The proposed Bayesian hierarchical model offers a principled framework for combining and reconciling size estimates of key populations.
• This approach enhances the accuracy and reliability of estimates crucial for public health planning and HIV/AIDS control.
• The model's application in Ukraine demonstrates its utility for informing targeted interventions.