A Bayesian hierarchical model with novel prior specifications for estimating HIV testing rates

Area of Science:

• Epidemiology
• Biostatistics
• Public Health

Background:

• Human immunodeficiency virus (HIV) infection and acquired immunodeficiency syndrome (AIDS) represent a significant global health challenge.
• Accurate estimation of the HIV testing rate is critical for effective public health interventions and disease control strategies.

Purpose of the Study:

• To develop and validate a novel Bayesian hierarchical model for estimating the HIV testing rate.
• To improve the accuracy of HIV incidence and testing rate estimations by incorporating temporal dependencies.

Main Methods:

• A two-level Bayesian hierarchical model was employed, utilizing Poisson and multinomial distributions.
• Latent HIV infections were modeled at the first level, while diagnosed and undiagnosed cases were modeled at the second level.
• A new class of priors accounting for temporal dependence was introduced, and an adaptive rejection Metropolis sampling algorithm was used for posterior computation.

Main Results:

• The proposed model effectively estimates the HIV testing rate using annual HIV/AIDS surveillance data.
• Simulation studies and analysis of US national data demonstrated the model's accuracy and efficiency.
• The incorporation of temporal dependence in priors enhanced estimation accuracy.

Conclusions:

• The developed Bayesian model provides a robust framework for estimating HIV testing rates and incidence.
• This methodology can significantly aid in understanding HIV/AIDS epidemiology and informing public health policies.
• Accurate HIV testing rate estimation is vital for monitoring and controlling the HIV epidemic.