Estimating HIV incidence from case-report data: method and an application in Colombia

Area of Science:

• Epidemiology
• Biostatistics

Background:

• Accurate quantification of human immunodeficiency virus (HIV) incidence is crucial for epidemic tracking, especially in concentrated epidemics where data quality can be a challenge.
• Routinely collected case-reporting data presents a potential but underutilized resource for estimating HIV incidence.

Purpose of the Study:

• To evaluate a novel method for estimating HIV incidence using routinely collected case-reporting data.
• To assess the model's performance on synthetic data and its applicability to real-world data from Colombia.

Main Methods:

• A Bayesian framework was employed, constructing a flexible model of HIV infection, diagnosis, and survival.
• Time trends in infection hazard were modeled using penalized B-splines, with performance validated on synthetic datasets.
• The model was applied to Colombian case-reporting data and compared with existing estimation methods.

Main Results:

• The developed method demonstrated feasibility and successfully recovered various incidence trends in synthetic data experiments.
• Estimates for recent HIV incidence exhibited significant uncertainty.
• Application to Colombian data generated a credible incidence trajectory, suggesting substantially lower historical HIV incidence than previously estimated.

Conclusions:

• Estimating HIV incidence from case-report data is feasible in settings with robust data availability, though not without limitations.
• Further research is recommended to address data biases, explore alternative modeling functions, and incorporate additional data sources like mortality and antiretroviral therapy indicators.