Estimating the Effects of Prison Visitation and Length of Incarceration on Post-Release Mortality and Reincarceration

People released from incarceration have an increased risk of mortality and an increased risk of subsequent incarceration compared to the general population. However, scant research has focused on which aspects of incarceration are harmful. We used data from the state of Rhode Island to investigate two aspects of incarceration, receiving visitors and incarceration length, and their effect on post-release mortality and reincarceration. In both aims, we followed a cohort of people released from the sentenced population of the Rhode Island Department of Corrections (RIDOC) for up to five years. We identified deaths from the Rhode Island Department of Health vital records and reincarcerations in the RIDOC system. In the first aim, we estimated the effect of hypothetical interventions that increased the number of visits received. We found visitation was common, with 98% of people receiving at least one visit during their incarceration; however, the average number of visits received per month decreased over the course of someone’s incarceration. We found that none of the interventions on visitation had an impact on post-release mortality and that all the interventions to increase visitation, including the most minimal interventions, prevented reincarceration. For example, under a hypothetical intervention to increase each person’s mean monthly visits by two, we estimated a five-year risk difference of -0.04 (95% CI: -0.05, -0.04).In the second aim, we estimated the effect of two hypothetical interventions to reduce incarceration length on post-release risk of mortality and reincarceration. The median length of incarceration was 3.3 months (interquartile range: 1.0, 10.1). We found that neither intervention had an effect on mortality. We found that reducing incarceration lengths to those observed in the 1990s in the United States led to a slight increase in risk of reincarceration (five-year risk difference: 0.0074 (95% CI, 0.0012, 0.0129)). Intervening to modeling incarceration length after those observed in Norway led to a slight reduction in reincarceration until 2 years of follow-up (6-month risk difference: 0.0078 (95% CI: -0.0143, -0.0035)).