Prison Drug Treatment’s Wider Effects: From Reducing Recidivism to Rehabilitation

Prison-based drug treatment programs have long been judged by a single yardstick: whether they keep former inmates from returning to crime. A new quasi-experimental study from Israel, published in the American Journal of Criminal Justice, suggests that this narrow metric may be profoundly misleading. Tracking thousands of released prisoners over five years, researchers found that the country’s main incarceration-based drug treatment programs largely failed to reduce recidivism—yet they delivered dramatic, life-saving benefits in other domains, cutting post-release death rates by as much as 80 percent and substantially improving employment and earnings among program participants.

The research team, led by Badi Hasisi of the Hebrew University of Jerusalem together with David Weisburd of Hebrew University and George Mason University, Noam Haviv of John Jay College, and Efrat Shoham of Ashkelon Academic College, evaluated the four main incarceration-based drug and alcohol treatment programs operated by the Israel Prison Service. Their motivation was twofold. First, most existing evaluations examine only a single program in a single facility, making it difficult to compare intensive, well-resourced interventions against more basic ones. Second, a growing body of scholarship—including a 2022 consensus report from the National Academies of Sciences, Engineering, and Medicine—has challenged recidivism itself as the primary measure of correctional success, arguing that rearrest and reincarceration capture only a fraction of what rehabilitation actually means for people reentering society.

The stakes of the question are high. Substance abuse is endemic in correctional systems: in Israel, a prison service survey found that roughly 36 percent of inmates had been diagnosed with a substance use disorder, while in the United States more than half of the prison population meets criteria for such a disorder—a rate approximately five times that of the general population. Drug dependency is tightly linked to crime, poor physical and mental health, and premature death. The period immediately following release is particularly lethal: a meta-analysis cited by the authors found that drug-related deaths spike in the weeks after release, largely from unintentional opioid overdose, with some studies attributing 45 to 77 percent of post-release fatalities to accidental overdose. Against this backdrop, the design of prison treatment programs—ranging from sophisticated therapeutic communities to rudimentary detoxification units—has direct consequences for survival, not merely for public safety.

Israel’s prison system offered an unusually rich natural laboratory. Its flagship program at Hermon Prison, a purpose-built national rehabilitation center set in a rural, campus-like environment without high walls, operates an intensive therapeutic community model lasting nine to twelve months, with structured stages, group therapy, Twelve-Step meetings, educational classes, and graduated responsibilities. A second framework, the Drug Rehabilitation Centers embedded in medium- and high-security facilities, follows similar therapeutic community principles and duration but lacks Hermon’s distinctive social and physical climate. A third, the Lev Room program, offers far more limited detoxification-oriented treatment in drug-free wings of ordinary prisons, typically without dedicated staff or resources. Dropout rates were high across the board—65 percent for the Hermon drug program, 75 percent for its alcohol counterpart, 58 percent for the DRC framework, and a striking 85 percent for Lev Room.

To evaluate these programs, the researchers assembled a dataset of 61,826 individuals released from Israeli prisons between 2004 and 2012, with recidivism information current through July 2015. Because assignment to treatment was voluntary and therefore non-random, they employed propensity score matching, a statistical technique that pairs each treated individual with a statistically “twin” non-participant based on a comprehensive set of pre-treatment characteristics: age, schooling, marital and immigration status, criminal history, age at first incarceration, offense type, violence markers, alcohol abuse history, and year of release. After screening out prisoners without substance dependence indicators and those with insufficient remaining sentence time to enter treatment, the team constructed matched intervention and comparison groups for each program. Matching quality was rigorously verified: whereas significant differences existed between groups on many variables before matching, post-matching differences vanished entirely, and standardized bias values fell below the accepted threshold of 20 for every covariate.

The recidivism results were sobering. Across a five-year follow-up period, none of the four programs produced a statistically significant reduction in reincarceration beyond the first year. The sole exception was the Hermon drug program, which lowered the risk of returning to prison by a relative 13.9 percent during the first year after release—an effect that faded entirely thereafter. The Hermon alcohol program, the DRC programs, and the Lev Room program showed no significant recidivism effects at any point in the follow-up. The authors interpret this cautiously: even the flagship program’s modest first-year impact may signal that gains erode without sustained aftercare and community-based continuity of care.

Yet when the researchers broadened the outcome measures—linking prison records to national mortality, employment, and income data from Israel’s Central Bureau of Statistics—a strikingly different picture emerged. Participants in the Hermon drug program died at dramatically lower rates than their matched counterparts. One year after release, mortality among program participants was approximately 80 percent lower (0.52 percent versus 2.04 percent), meaning untreated individuals faced a fourfold higher risk of death. The gap persisted through the third year, when the comparison group remained about 60 percent more likely to have died. Even in years four and five, when the difference no longer reached statistical significance, the protective trend continued. Given that overdose is the leading cause of death among formerly incarcerated people with drug dependence, this finding suggests that intensive treatment may function as a genuine life-saving intervention, even where it fails to alter criminal trajectories.

Employment and income outcomes reinforced the pattern. Hermon drug program participants worked significantly more months per year than their matched peers across the entire five-year follow-up—41 percent more in year one, roughly 30 percent more by year five—and earned significantly higher annual incomes in years one, three, and five, with earnings about 40 percent higher by the end of the fifth year. The DRC program showed similar though somewhat smaller benefits: employment months were 37 percent higher in the first year and 78 percent higher by year five, while income advantages reached 72 percent at the five-year mark. The authors emphasize that for a population burdened by ill health, absent work histories, and the stigma of a criminal record, even modest absolute gains in employment represent meaningful achievements—and known correlates of reduced reoffending over the long term.

The study’s implications reach well beyond Israel. The authors argue that evaluations anchored solely in recidivism risk concluding that prison-based drug treatment “doesn’t work,” when in fact such programs demonstrably improve the health, economic stability, and survival of some of society’s most vulnerable individuals. A public health perspective, they contend, treats reduced mortality and improved employment as valuable outcomes in their own right, regardless of short-term effects on crime. This reframing has cost-benefit consequences as well: weighed against the substantial social costs of preventable overdose deaths, chronic unemployment, and untreated addiction, drug treatment programs may deliver a favorable return on investment even without strong effects on reoffending. The researchers also call for criminal justice agencies to integrate reentry data across health, labor, and corrections domains so that program success can be assessed in its full dimension.

The authors are careful to acknowledge limitations. As with all non-experimental research, unobserved variables—such as prior employment or health histories, which were not systematically recorded—could influence the propensity models. The voluntary nature of treatment raises the possibility that participants possess higher baseline motivation than the general prison population, though the matching design was chosen precisely to mitigate such selection bias. High dropout rates, particularly in the weaker programs, complicate interpretation; notably, a supplementary completers-only analysis revealed promising recidivism reductions in the Hermon drug program across the full five-year follow-up, hinting that treatment dosage matters. The binary measure of reincarceration may also understate progress along the “desistance” trajectory, where reductions in the frequency or severity of offending, even without full cessation, constitute meaningful movement toward reintegration.

Taken together, the findings deliver a pointed message to criminologists and policymakers alike: the question “does prison drug treatment reduce recidivism?” is too narrow to capture what these programs actually do. The Hermon program’s most dramatic and durable effects were measured not in arrest statistics but in lives saved and livelihoods rebuilt. As the National Academies report quoted in the study urges, the field may need to reverse the polarity of recidivism—from a measure of failure to a component of success—recognizing that successful reintegration encompasses health, housing, employment, and social connection far beyond what any rearrest rate can register. In an era of renewed interest in rehabilitation, this study suggests that the true value of treating addiction behind bars may be measured first in who survives, and only later in who stays free.