Predicting HIV Incidence in the SEARCH Trial: A Mathematical Modeling Study

Area of Science:

• Epidemiology
• Mathematical Modeling
• Public Health Interventions

Background:

• The SEARCH study implemented community-based HIV testing and treatment in East Africa, reporting no significant difference in 3-year HIV incidence between arms.
• Mathematical modeling was employed to analyze the influence of viral suppression and community mixing on the SEARCH trial's outcomes.

Purpose of the Study:

• To estimate the impact of control arm viral suppression and community mixing on the results of the SEARCH trial.
• To model comparative HIV incidence under various scenarios, considering ART scale-up and community engagement.

Main Methods:

• Utilized the EMOD-HIV individual-based modeling software to create a model of SEARCH communities.
• Parameterized and calibrated the model with demographic, HIV prevalence, male circumcision, and viral suppression data.
• Estimated comparative HIV incidence by simulating different levels of ART scale-up, community mixing, and baseline testing effects.

Main Results:

• Pre-trial predictions indicated a 4%-40% reduction in HIV incidence, contingent on control arm ART linkage and community mixing.
• Modeled effect sizes were smaller than anticipated due to universal baseline testing and rapid ART uptake.
• Interim viral suppression data suggested potential reductions of up to 27% in the control arm and 43% in the intervention arm compared to a counterfactual scenario.

Conclusions:

• The active control arm in the SEARCH study likely reduced the trial's statistical power.
• Despite this, the SEARCH intervention demonstrated a significant reduction in HIV incidence (up to 43%) compared to a scenario without enhanced baseline testing and treatment linkage.