Metrics of Mobility by Sex are Associated with HIV Incidence in Rural Kenya and Uganda

The complex and dynamic nature of human mobility requires use of multiple measures and sex-stratified analyses to fully understand its influence on HIV acquisition risk in specific populations and settings. Longitudinal population-based studies designed to measure the influence of mobility (which includes both short and long-term, temporary and permanent changes of residence across defined geopolitical boundaries) on HIV acquisition risk are rare, but needed to understand the ongoing challenges that mobility poses to progress in ending HIV. Incidence of HIV acquisition over 3 years was measured in mobile and non-mobile adults in 32 rural communities in three regions of Uganda and Kenya participating in the Sustainable East Africa Research in Community Health (SEARCH) trial (NCT01864603) from 2013-2017. Poisson regression models were used to estimate incidence rate ratios (IRRs) for HIV acquisition among categories of mobile relative to non-mobile adults, with sex-stratified multivariable models adjusted for community clustering using robust estimators. At baseline, 11,337 adult residents (9.7%) reported living at least 1 month outside their community in the past 12 months (8% women, 11.8% men, p < 0.001). By year three, 8502 HIV incidence cohort members (7.3%) had out-migrated (7.7% women, 7.1% men, p = 0.046), and 5747 adults remaining in the cohort (4.9%) reported living at least 1 month outside of their community in past 12 months (4.1% women, 5.9% men, p < 0.001). Over 3 years, the risk of HIV acquisition was 1.9 times higher in those living >= 1 month outside their community in the past 12 months compared to those who had not (Adj IRR = 1.88, 95% CI 1.48-2.38). HIV acquisition risk was 42% higher among men who spent some nights away in the past month at baseline (Adj IRR = 1.42, 95% CI 1.11-1.81), but not for women (Adj IRR = 1.06, 95% CI 0.85-1.32). Mobility was significantly associated with a risk of HIV acquisition among adults in rural communities in Kenya and Uganda in 2013-2017. The effect of mobility on HIV incidence varied by both sex and pattern (frequency, duration) of mobility (SEARCH ClinicalTrials.gov number, NCT01864603).