In 2017, the Madison Police Department launched the Madison Addiction Recovery Initiative (MARI), a collaborative effort with a host of local agencies aimed at creating a more effective response to the opioid epidemic.
Gabriel Zayas-Caban
Rather than arresting individuals who had committed a nonviolent, drug-use-related crime, officers could offer eligible adults a personalized treatment program for substance-use disorder. The diversion program, which has since evolved into the Madison Area Addiction Recovery Initiative, and similar efforts in communities across the country offer an alternative to the use-arrest-use cycle (and possible incarceration) that hinders those individuals and strains the criminal justice system.
In order to accomplish that mission, however, the programs need to do more than merely exist.
They need to be effective.
Working in close partnership with the MARI team, operations researchers from the University of Wisconsin-Madison Department of Industrial and Systems Engineering have analyzed the initiative’s performance. By their estimate, MARI reduced the odds of recidivism within six months.
The team published its findings—the first in an ongoing study—in the journal Drug and Alcohol Dependence.
“It’s promising; that’s the biggest thing that we’re trying to answer. At least in six months, is it doing what we expect it to? Is it reducing crime?” says Veronica White, a PhD student who was first author on the study, which is part of her dissertation work on the opioid crisis in Dane County and the effects and costs of various potential community initiatives and policies.
Veronica White
The UW-Madison team, which also included Assistant Professor Gabriel Zayas-Caban, PhD student Sebastian Alvarez Avendano and Professor Laura Albert, used three types of analyses—all methods that fall under the bucket of causal inference from observational data—to parse the MARI data. (To account for the lack of a randomized assignment that is typical of a controlled trial, since there was no control group of MARI-eligible individuals who didn’t participate in the program, Alvarez employed a newer approach to estimating the complier average causal effects analysis in order to compare program participants to a historical comparison group.)
White says the research group’s partnership with the MARI team has been vital in providing deeper context for the data and enabling insightful analysis. She’s helped with data collection and met regularly with Aleksandra Zgierska (now a professor at the Penn State College of Medicine and a former UW-Madison faculty member and principal investigator for MARI), and Joe Balles (a retired Madison Police captain who’s coordinated MARI through the nonprofit Safe Communities Madison-Dane County).
“I hope it can be a model for future initiatives, in terms of how academic researchers should be involved from the beginning,” says White, who is co-advised by Zayas-Caban and Albert. “Not only were we able to use different tools that they had probably never heard of, like causal inference, but we were also able to help make sure they were capturing the data that could use those tools.”
The UW-Madison operations researchers will next look at MARI’s effects on recidivism at 12 months, and White also sees opportunities to probe the program’s influence on the health outcomes of participants. Zayas-Caban notes that while diversion programs have grown, thorough analyses like this one have been slower to follow.
“These kind of community policing initiatives are popping up everywhere,” says Zayas-Caban, whose expertise is in applying causal inference methods to healthcare and policy analysis. “This is, I think, pretty close to being one of the first studies to really systematically collect this data and analyze it in an academic environment.”
The study was funded by grants from the National Science Foundation and the U.S. Department of Justice Bureau of Justice Assistance.
Zayas-Caban is the Jane R. and Jack G. Mandula Assistant Professor in the Department of Industrial and Systems Engineering. Albert is the David. H. Gustafson Chair of the Department of Industrial and Systems Engineering.