Mathematical Modeling to Inform Implementation of HIV Prevention Programs in the United States
Despite advances in the detection and treatment of HIV, the incidence of infection in the United States has increased in some subgroups over the past decade and remained stable in others. These trends point to a need for improved strategies for prevention that take into account the social, behavioral, and clinical context of different target populations. We conducted an internet-based survey to measure sexual behavior and use of pre-exposure prophylaxis (PrEP) among men who have sex with men (MSM) in Washington State (Chapter 1). Among 1,080 cisgender MSM respondents, 79% had heard of PrEP, 19% reported current use, and 36% of PrEP-naÃ¯ve men reported that they wanted to start taking it. Among high-risk men recommended to initiate PrEP, 31% were taking it. With the data from this survey, in combination with secondary data from local surveillance systems and other surveys, we developed a dynamic network-based mathematical model to evaluate the potential impact of PrEP on HIV incidence in Washington MSM (Chapter 2). In the context of the high levels of testing and treatment in Washington, our model estimated that HIV incidence at the end of the 10-year simulation would be 48-83% lower with continued or increasing use of PrEP relative to a counterfactual scenario with no PrEP use. In chapter 3, we constructed a static linear mathematical model to estimate the impact and optimal age for one-time routine HIV screening in terms of case detection, person-years of undiagnosed infection, and progression to symptomatic HIV/AIDS. When added to prenatal, risk-based, symptom-based, and partner notification testing, our model estimated that the impact of routine screening is likely to be modest. The percent of tests resulting in new diagnoses exceeded the recommended minimum of 0.1% only in a setting with high HIV incidence in groups that don’t engage in repeat, targeted testing. The results from these three projects provide important insights to inform local policies and HIV prevention strategies, demonstrating the value of applying mathematical modeling to inform public health practice. Our findings highlight the influence of epidemiologic context on the impact of interventions such as PrEP and HIV screening, underscoring the importance of using local data to define context-specific prevention strategies.