Date of Award

January 2026

Document Type

Thesis

Degree Name

Medical Doctor (MD)

Department

Medicine

First Advisor

Benjamin A. Howell

Abstract

Individuals released from incarceration face a heightened risk of mortality, with overdose as the leading cause. Most post-release mortality research has focused on prison settings; less is known about risk of death following jail release. Jail exposures are heterogeneous, ranging from one short stay to multiple longer stays. Understanding jail heterogeneity and its association with mortality can help inform targeted interventions.

We conducted a retrospective cohort study using linked jail, prison, and death records of individuals in Minnesota with at least one jail stay from January 1, 2017, through August 31, 2024. We excluded anyone under 18 years old, those who died during incarceration, and those who had a prison exposure during the study period.

We first conducted latent class analysis to group individuals based on four jail exposure characteristics: total number of jail days, average length of stay, number of jail stays, and the presence/absence of jail transfer. Using the resultant classes, we then conducted survival analysis with a piecewise Cox proportional hazards model to examine the association between latent class and all-cause, overdose, and opioid-involved overdose mortality, splitting the follow-up time before and after one-month post-release. We adjusted all analyses by age at last release, race, and gender.

Among the 270,054 individuals released from jail in Minnesota between January 1, 2017, and August 31, 2024, who experienced a total of 533,280 incarceration episodes, 8,401 died during the study period. Overdose was the leading cause of death (n=2,001), with 72% of overdoses involving opioids. Most individuals in this sample were male (70%) and white (72%). Individuals spent a median of 3 total days in jail (IQR 1-7), with a median of 1 jail stay (IQR: 1-2), a median average length of stay of 2 days (IQR: 1-4), and 14% experienced a jail transfer.

Latent class analysis yielded a 3-class model: 1) Class 1 (Limited Jail Exposure) comprised individuals who spent the least amount of time in jail, experienced the shortest and fewest stays, and had the lowest proportion of people experiencing a transfer, 2) Class 2 (Extended Jail Stays) was characterized by individuals who experienced a few prolonged jail stays with a moderate proportion of people experiencing a jail transfer, and 3) Class 3 (High Jail Churn) comprised those who had short, frequent jail stays with the greatest proportion of people experiencing a transfer.

During the first month post-release, compared to those with limited jail exposure (Class 1), the risk of all-cause mortality was significantly higher for those with extended jail stays (Class 2; HR: 2.60, 95% CI: 1.87-3.60) and those with high jail churn (Class 3; HR: 3.78, 95% CI: 3.06-4.66). After the first month post-release, compared to those with limited exposure (Class 1), the risk of long-term all-cause mortality was significantly greater for those with extended jail stays (Class 2; HR:1.58, 95% CI: 1.43-1.73) and those with high jail churn (Class 3; HR: 2.58, 95% CI: 2.39-2.77).

Regarding overdose mortality, within the first month post-release, compared to individuals with limited jail exposure (Class 1), the risk of overdose mortality was significantly higher for those with extended jail stays (Class 2; HR: 5.16, 95% CI: 3.20-8.32) and those with high jail churn (Class 3; HR: 6.88, 95% CI: 4.94-9.61). After the first month post-release, the risk of overdose mortality was significantly greater for those with extended jail stays (Class 2; HR: 2.09, 95% CI: 1.75-2.50) and those with high jail churn (Class 3; HR: 4.35, 95% CI: 3.85-4.91), compared to those with limited jail exposure (Class 1).

Lastly, during the first month post-release, individuals with extended jail stays (Class 2; HR: 4.11, 95% CI: 2.21-7.63) and high jail churn (Class 3; HR: 7.75, 95% CI: 5.30-11.34) experienced significantly higher risk of opioid-involved overdose mortality than those with limited jail exposure (Class 1). After the first month post-release, compared to people with limited jail exposure (Class 1), those with extended jail stays (Class 2; HR: 1.84, 95% CI: 1.47-2.31) and with high jail churn (Class 3; HR: 4.74, 95% CI: 4.13-5.45) faced significantly higher risk of opioid-involved overdose mortality.

In conclusion, differences in jail exposure patterns are important markers of mortality risk following release from jail. Individuals with extended jail stays and high jail churn experience significantly higher risk of all-cause, overdose, and opioid-involved overdose mortality, compared to those with limited jail exposure. Those who experience high jail churn, characterized by many short jail stays, face the greatest mortality risk. Interventions to reduce jail churn, length of stay, and transfers may mitigate mortality in this vulnerable population.

Comments

This thesis is restricted to Yale network users only. This thesis is permanently embargoed from public release.

Download

Share

COinS