Statistical Methods for Combination HIV Prevention Approaches

(PI: Chen)

This project is in response to the Funding Opportunity Announcement (FOA) RFA-MH-14-180, entitled “Methodologies and Formative Work for Combination HIV Prevention Approaches (R01)”. As recognized in the RFA, “recent advances in biomedical interventions with critical behavioral aspects (e.g., Pre-exposure Pro- phylaxis [PrEP], Treatment as Prevention [TasP]) have changed how HIV prevention and treatment are concep- tualized.” Moreover, “reductions in HIV incidence will only be achieved through implementation of combinations of interventions that include biomedical and behavioral interventions, as well as components that address social, economic, and other structural factors that influence HIV prevention and transmission. However, combined prevention intervention approaches rely on synergies of multiple elements that can be challenging to design, implement, and evaluate.”

In fact, for the past two decades, we have led or collaborated on design and analysis for many HIV/AIDS prevention trials conducted by the HIVNET, and the later HIV Prevention Trial Network (HPTN), including high- profile studies such as the EXPLORE (HIVNET 015) Study of behavioral intervention among men having sex with men (MSM), the HPTN 043/ACCEPT Study of community mobilization, mobile testing, same-day results and post- test support for HIV in sub-Saharan Africa and Thailand, the HPTN 052 Study of TasP among sero-discordant couples, and the ongoing HPTN 071 Study of a combination package on population-level HIV incidence in Zambia and South Africa (PopART). From our experience, we have developed a broad and deep understanding of the background and research objectives outlined in this RFA.

In response, we identify four imminent challenges faced by researchers, policy makers and practitioners, when designing, implementing and evaluating combined prevention intervention approaches for optimal public health impact. They include: 1) developing and applying powerful and easy-to-implement trial design(s) to assess the effectiveness of combined behavioral and biomedical interventions; 2) assessing the public health impact of combinations of effectiveness-proven prevention interventions; 3) utilizing established government surveillance systems to monitor resource use, improve program implementation and measure impact of combination preven- tion interventions; and 4) evaluating the predictions of complex epidemic models, particularly as they may be used in planning of combination intervention studies. To address these
particular methodological challenges, we specifically aim to:

Aim 1. Develop statistical methods for factorial designs that can be used to assess combination prevention intervention approaches.

Aim 2. Develop statistical methods for using time-varying attributable risk functions (ARF) to assess the overall public health impact of combination prevention intervention approaches. 

Aim 3. Develop statistical methods for using surveillance data to improve estimates and inferences of HIV prevention measures. 

Aim 4. Develop a statistical framework for evaluating the predictions of complex epidemic models and describe experiments that can be used to test this framework.