Climate change drives increase in modeled HIV prevalence

Long-term changes in temperature can negatively affect human livelihoods with resulting implications for health outcomes. While increasing attention has been paid to the direct health consequences of climate change, there has been less focus on indirect health effects, partly due to the relative complexity of establishing and modeling these outcomes. Here, I leverage a dataset of over 400,000 individuals across 25 countries in Sub-Saharan Africa, coupled with high-resolution climate data for the region, to test the effect of long-run temperature changes on HIV prevalence. I find that warmer periods are linked with an increase in HIV prevalence, particularly for younger generations. Both economic and behavioral changes likely drive this linkage: I find evidence that male migration and sex-market use increase with higher temperatures. I generate projections for future HIV prevalence using two methods. First, I project directly using the empirical estimate, an approach frequently used in econometric studies of climate change impacts. Second, I develop a mechanistic model of HIV spread, incorporating the estimated temperature effect. I find that agreement between the models is strong (77%). Under the RCP8.5 scenario for warming, I find that climate change will lead to between 11.6 and 16.0 million additional cases of HIV by 2050, an increase in prevalence of 1.4–2.1 percentage points.