Stochastic models to demonstrate the effect of motivated testing on HIV incidence estimates using the serological...

Area of Science:

• Epidemiology
• Biostatistics
• Public Health

Background:

• The Serological Testing Algorithm for Recent HIV Seroconversion (STARHS) is used to estimate HIV incidence.
• STARHS assumes independence between HIV infection and testing, which may not hold true in clinical settings.
• Previous studies show STARHS estimates often exceed cohort-based incidence estimates.

Purpose of the Study:

• To investigate whether testing motivated by suspected HIV exposure biases STARHS estimates.
• To compare the robustness of STARHS and cohort-based incidence estimation methods.

Main Methods:

• Monte Carlo simulations of 10,000 men who have sex with men were conducted.
• Simulations modeled both independent infection/testing and testing motivated by suspected exposure.
• Both STARHS and cohort-based incidence estimation methods were applied to simulated and empirical data.

Main Results:

• Under simulated independence, STARHS and cohort estimates aligned with clinical data.
• Motivated testing inflated STARHS estimates but did not affect cohort-based estimates.
• The degree of testing motivation significantly impacted STARHS estimates.

Conclusions:

• Cohort-based HIV incidence estimates are reliable even when testing is influenced by infection risk.
• STARHS HIV incidence estimates are susceptible to bias from motivated testing.