Mortality & the Criminal Legal System: Inside the Process of a Wilson Center Research Project
The Project
In 2024, more than 20,000 people were released from North Carolina prisons, and over 76,000 were living under community supervision through the Department of Adult Corrections. "These people have served their time," says Michele Easter, Associate Professor in Psychiatry and Behavioral Sciences, "and it is in everybody's interest that they integrate back into our communities." What that integration looks like, and how often it fails in the most irreversible way, is at the center of a new research project at the Wilson Center for Science and Justice.
“There are a lot of challenges facing people as they reenter," Easter says, "and one very blunt way to measure that is to look at mortality rates.” Several studies have measured mortality rates of formerly incarcerated people over the weeks, months, or years after release, in North Carolina and other states, consistently finding that they die at higher rates than the general population, even when controlling for factors such as age. In North Carolina, for example, people released from prison were 40 times more likely to die from an opioid overdose in the two weeks following release than the general population during the 2000 to 2018 era.
Most existing research focuses almost exclusively on people leaving incarceration. This project goes further, comparing three groups: people released from prison without probation, people who served prison time and were then placed on probation, and people on probation whose conviction would have been eligible for prison time but who were never incarcerated. The project will also compare mortality rates across different county contexts and account for individual demographic factors such as race. Led by Easter and data scientist Rita Grunberg, with contributions from former postdoctoral fellow Madeline Stenger, the team aims to build North Carolina-specific mortality data across all three groups and examine how county-level conditions shape outcomes.
Building the Data Infrastructure
The project draws on the statewide North Carolina death registry spanning 2013 to 2023 and data from the North Carolina Department of Adult Corrections. The central technical challenge is linking those two datasets. Names are misspelled, birthdates transposed, and identities changed over time.
To handle this, the team used probabilistic matching software called Link King, which ranks potential matches across six confidence levels based on how well identifiers like name, birthdate, and race align. After quality checking, 23,966 out of 411,449 Department of Corrections records were successfully matched to death records, covering both people released from prison and people on probation.
That process also surfaced a structural limitation. Women are less likely to be linked across datasets due to name changes after marriage, which may mean their deaths are systematically underrepresented in this study. Correcting for that gap is an ongoing part of the work.
One advantage of working with death registry data, Easter notes, is its durability. "Unlike arrest or court data, these data are likely to be available indefinitely — so we can accumulate many years for potential study." That stability adds the possibility of tracking trends and policy changes across time as the dataset grows.
The Challenge of Classifying by County
The next major challenge is developing a county-level measure of what the team calls "reentry friendliness." The premise is that where someone lives after release or during supervision shapes their mortality risk. Access to treatment, community resources, and local attitudes all vary considerably by county, and no systematic measure of this currently exists.
To build this metric of “reentry friendliness” for counties across the state, the team will piece together existing healthcare access data and public records alongside perspectives from OurJourney, a reentry organization whose staff bring lived experience and help people navigate resources across North Carolina. Integrating that on-the-ground knowledge with quantitative county-level data is meant to produce a fuller picture than either source could offer alone. Once the reentry friendliness classification is finalized, the analysis itself can begin.
Long-term Goals of the Research
The intended audience is state-level and local decision makers with authority over how resources move through the criminal legal system and into the communities people return to. The researchers aim to highlight mortality among people leaving incarceration for policymakers: documenting how many die, identifying where reentry is most challenging, and offering concrete strategies for reducing those deaths.