METHODS: To create a retrospective cohort of all adults with HIV released from jails and prisons in Connecticut, USA (2007-14), we linked administrative custody and pharmacy databases with mandatory HIV/AIDS surveillance monitoring and case management data. We examined time to LTC (defined as first viral load measurement after release) and viral suppression at LTC. We used generalised estimating equations to show predictors of LTC within 14 days and 30 days of release.

FINDINGS: Among 3302 incarceration periods for 1350 individuals between 2007 and 2014, 672 (21%) of 3181 periods had LTC within 14 days of release, 1042 (34%) of 3064 had LTC within 30 days of release, and 301 (29%) of 1042 had detectable viral loads at LTC. Factors positively associated with LTC within 14 days of release are intermediate (31-364 days) incarceration duration (adjusted odds ratio 1·52; 95% CI 1·19-1·95), and transitional case management (1·65; 1·36-1·99), receipt of antiretroviral therapy during incarceration (1·39; 1·11-1·74), and two or more medical comorbidities (1·86; 1·48-2·36). Reincarceration (0·70; 0·56-0·88) and conditional release (0·62; 0·50-0·78) were negatively associated with LTC within 14 days. Hispanic ethnicity, bonded release, and psychiatric comorbidity were also associated with LTC within 30 days but reincarceration was not.

INTERPRETATION: LTC after release is suboptimal but improves when inmates' medical, psychiatric, and case management needs are identified and addressed before release. People who are rapidly cycling through jail facilities are particularly vulnerable to missed linkage opportunities. The use of integrated programmes to align justice and health-care goals has great potential to improve long-term HIV treatment outcomes.