Estimating HIV Risk Among Individuals Released from Incarceration in Washington State: A Data-Driven Approach to Targeted Prevention.

People being released from prison face a heightened risk of HIV acquisition, yet there is need for additional research to quantify that risk and identify which groups should be prioritized for prevention efforts. We performed a population cohort study of people released from Washington State prisons to estimate the HIV acquisition risk post-release and applied machine learning methods to identify subpopulations at highest risk. We linked state prison release rosters (1/1/2010 and 12/31/2021) to HIV surveillance data to calculate HIV incidence post-release and used a decision tree to partition the population into risk strata using demographic variables, reason for incarceration, mental health, and substance use assessments. Among 58,026 individuals released from prison, 120 were diagnosed with HIV in the three years following release, corresponding to a rate of 32.5 per 100,000 person-years. Specific subgroups of people who use substances were found to have HIV diagnosis rates 24-38 times higher than the general population. People released from prison in Washington State are at a risk of HIV acquisition 6.1 times higher than the rate among individuals not released from prison. Routinely collected data can identify high-risk subgroups, enabling targeted prevention programs in prison and after release.