Illicit substance use after release from prison among formerly incarcerated primary care patients: a cross-sectional study

The criminal justice system has an exceptionally broad reach in the United States. At any one time, over 2 million people are incarcerated with 1.3 million in state prisons, 630,000 in local jails, nearly 200,000 in federal prison, and 40,000 in immigration detention centers [1]. Problematic substance use is common among this population with more than 80% of people in jail or prison reporting having used illicit substances in their lifetimes [1, 2]. The majority of people in jails (53%), state prisons (56%), or federal prisons (50%) met DSM-IV criteria for a substance use disorder (SUD) at the most recent national surveys [3‐5]. Additionally, according to a report by the Substance Abuse and Mental Health Services Administration (SAMHSA), 19% of males on probation (i.e. sentenced and serving time in the community) aged 18–49 had a drug use disorder (DUD) and over a quarter had an alcohol use disorder (AUD) in 2012 [6]. However, few incarcerated individuals receive evidence-based SUD treatment, and substance use often continues during and after incarceration [7, 8].

During incarceration, less than 20% of individuals with SUDs receive formal treatment. Pharmacotherapies, such as methadone maintenance treatment for opioid use disorder, are rarely offered in correctional settings; even when available, only a small fraction of eligible individuals access treatment [9]. Addressing substance use and SUDs in the criminal justice population will require additional attention and new approaches.

Substance use disorders are chronic relapsing conditions. Even those who stop substance use during incarceration may resume use post-release, which introduces several risks [9]. Numerous studies have documented greatly elevated risk of death when people are released from jail or prison with the leading cause of death being drug overdose [10, 11]. People in jail or prison who resume injecting drugs post-release are also at risk for transmitting viral infections, such as HIV or Hepatitis C Virus [12]. Because of the co-occurrence of SUDs and mental health conditions, post-release substance use may also worsen mental health status and prevent engagement in needed medical care [13]. Additionally, post-release substance use can lead to re-incarceration. Formerly incarcerated individuals with SUDs or substance-related criminal charges are more likely to be re-incarcerated than those without substance involvement [14, 15]. Qualitative research suggests that substance use post-release may be due to poor mental health, environmental exposures (e.g., substance-using peer groups), or life stressors related to community re-entry, such as challenges finding work and stable housing [16‐18]. Additional research is needed to better understand substance use among people who have been released from jail or prison.

Post-release substance use can carry risk even when individuals do not have SUDs. General conditions of parole prohibit the use or possession of a controlled substance [19]. Drug testing procedures for individuals on parole vary depending on state regulations, but post-release substance use could lead to parole violations [20]. In New York State, for instance, any possession of drug paraphernalia or use or possession of a controlled substance without medical authorization may result in parole revocation [21]. Therefore, research on substance use post-release should include individuals both with and without SUD diagnoses.

While substance use and SUDs are common prior to incarceration, there is a dearth of data on the rate of substance use post-release. One systematic review highlighted that most studies of SUD care post-incarceration have reported criminal justice outcomes (e.g., re-incarceration) instead of substance use outcomes. [22] Additionally, existing research has focused on cohorts of individuals in SUD treatment who likely carry the greatest risk of substance use resumption [23‐26]. For example, a study assessing the efficacy of a therapeutic community treatment program for formerly incarcerated individuals with SUDs found that 79% of participants in the treatment group resumed illicit drug use within 5 years post-release [24]. A clinical trial assessing opioid use disorder treatments at prison release found that more than three-quarters of the study arm that received pre-release counseling resumed heroin use at 3 months post-release [25]. A cohort study tracking individuals released from jail with varying levels of substance use found that 43% had at least one substance dependence symptom 1 year post-release; over a quarter (28%) reported cocaine use and a third (33%) reported cannabis use [26]. These data highlight the chronicity of severe SUDs; however, it also possible that individuals with less severe substance use do cut down or stop substance use post-release, which has implications for monitoring and service provision.

In this study, we investigated substance use in a diverse group of individuals who were recently released from incarceration. These data could improve generalizability of knowledge regarding substance use resumption post-incarceration, which to date has been mostly derived from individuals receiving SUD treatment. We also explored factors associated with substance use following release from incarceration. Understanding the trajectory of substance use following release from incarceration for people with and without SUDs can guide the development of targeted interventions for those at greatest risk of poor substance use-related outcomes.

This cross-sectional study utilized baseline data from the Transitions Clinic Network (TCN, www.​transitionsclini​c.​org), a multi-site prospective longitudinal cohort study of post-incarceration medical care.

Of the 751 participants who completed the TCN baseline survey, the median age was 47, participants were mostly male (85%), non-white (47% black, 30% Hispanic), and had not graduated from high school (59%). Participants were most commonly on parole (80%), lived in institutional housing (39%), and unemployed (92%). The median time incarcerated during participants’ most recent prison term was 4 years (Interquartile range: 2–8 years). The median time from prison release to engagement at a TCN site was 5 weeks (IQR: 2–9 weeks). Among clinical factors, slightly less than half of all participants reported a prior diagnosis with depression (46%) or a drug use disorder (45%) (see Table 1).

The proportion of participants reporting any illicit substance use and any alcohol use soon after release from incarceration was 18% and 23%, respectively. The 134 participants who reported any illicit substance use post-release differed in demographic, social and clinical characteristics from those without illicit substance use. In bivariate analysis there were significant associations between post-release illicit substance use and younger age, male gender, not being on parole, housing status, psychiatric diagnoses, incarceration history and time to engagement at a TCN site. Illicit substance use was positively associated with unstable housing and doubling up and negatively associated with institutional housing. Spending fewer years incarcerated during the most recent prison term was associated with post-release illicit substance use. Reporting a prior diagnosis of depression, bipolar disorder or DUD was also associated with post-incarceration illicit substance use (Table 1).

In regards to post-release illicit substance use, cannabis use was most common with 12% of participants reporting post-release cannabis use. Fewer participants reported post-release cocaine or opioids use: 4% for each substance (see Table 2). Of participants with a DUD diagnosis, 67% reported receiving treatment during incarceration. The most common form of treatment reported was narcotics anonymous (61% of those who received treatment). Formal programs (20%), one-on-one counseling (20%), and pharmacotherapy (4%) were less commonly reported. One participant reported receiving art therapy.

In the multivariable regression model several variables remained significantly associated with post-release illicit substance use, including male gender (aOR = 3.91, 95% CI: 1.73–8.81), housing with friends or family (aOR = 3.33, 95% CI: 1.20–9.28), time incarcerated during most recent prison term (aOR = 0.93, 95% CI: 0.89–0.98), weeks elapsed before engagement with TCN (aOR = 1.07, 95% CI: 1.03–1.10), being on parole (aOR = 0.58, 95% CI: 0.34–0.99), and having an DUD diagnosis (aOR = 2.27, 95% CI: 1.40–3.67) (see Table 3). In the first sensitivity analysis among those only with DUD, housing with friends or family was no longer significantly associated with illicit substance use, but the point estimate of the odds ratio remained similar to that of the full sample (aOR = 2.74, 95% CI: 0.65–11.56). Other variables maintained statistical significance. In the second sensitivity analysis, restriction of the sample to only those on parole was not found to affect which covariates in the multivariable model maintained significance.

In our cohort of individuals recently released from prison who initiated medical care at a transitions clinic, 18% reported illicit substance use between their prison release and first primary care appointment. In multivariable analysis, we found that post-release substance use was associated with expected risk factors such as drug use disorders, male gender, parole status, and time elapsed between release and the first medical encounter. Interestingly, housing status—specifically, living “doubled up” with friends or family members—had among the strongest association with post-release substance use, and this did not change when we excluded participants who were not monitored by parole (data not shown). Also, greater amount of time incarcerated at the latest prison term (in years) was associated with lower odds of post-release illicit substance use. These findings suggest areas, such as post-incarceration aftercare for drug use disorders or structured housing environments, where interventions could reduce the risk of post-release substance use and perhaps consequences of substance use.

Our findings add to the literature on substance use after prison release by focusing on a general population instead of only participants enrolled in SUD treatment. Our post-release illicit substance use incidence was at the low end of the range of previously published studies (18% vs. 22–88% at 3–6 months [22, 25, 32] and 70–95% at 1–3 years post-release [33‐35]. Our study enrolled primary care patients who were released from prison, while prior studies mostly enrolled individuals enrolled in SUD treatment who are likely at highest risk of relapse. In our study, a history of drug use disorders was common (45%) and associated with post-release substance use, but even among those with a drug use disorder history, only 24% reported illicit substance use after the time of release. One study that is commonly cited in the scientific literature (338 times as per Google Scholar, searched on August 5, 2018) and policy reports, estimates that 95% of substance-involved people in prison will relapse to substance use post-release; however, the results should be interpreted with acknowledgement of the sampling frame, which selected for individuals with severe substance use disorders [33]. Our study also has limitations regarding generalizability (see below), but there is likely high variability in risk of substance use after release from incarceration.

Important factors that likely influenced our lower incidence of substance use were that we engaged participants soon after their release, most were monitored by parole, and our sampling strategy selected for a cohort that was older than prior studies. The median time elapsed between release and the first medical encounter among the TCN group was only 5 weeks. Many prior studies reported substance use over longer periods of time post-release. This is important for two reasons. First, it is plausible that substance use incidence is low in the first months after release, but then increases proportionally with time. This is consistent with our data, which shows that each additional week between release and the first medical encounter was associated with a 7% increased odds of substance use. This is also consistent with a 2004 prospective study following a general cohort of formerly incarcerated individuals that found illicit substance use rates of 22% at 4–6 months post-release [32]. Qualitative data highlight how recently incarcerated people with SUDs may express confidence and motivation to avoid substance use soon after release, but challenges during community reentry and accompanying emotional distress may lead to substance use [36]. Second, the majority of our participants were monitored by parole, which may have affected decisions around substance use. If participants’ parole monitoring included urine drug testing, this may have effectively discouraged illicit substance use. Data are conflicting about the types of monitoring practices that are most effective, but close supervision of substance use and certain and immediate consequences are considered best practices. [37] Third, the median age within our cohort was 47 years, and most national surveys in the United States suggest that incidence and prevalence of alcohol and drug use disorders decrease with age [38]. The prior studies cited above that reported post-release substance use had enrolled participants that were approximately 7–17 years younger than our cohort. Older age was not significantly associated with illicit substance use in our cohort, but selection of an older sample may have affected our low reported estimates of substance use. Engaging formerly incarcerated individuals in the early post-release period and capitalizing on parole’s influence on substance use could support abstinence from substance use.

Another interesting finding from our study is that housing status may also be associated with post-release substance use. Individuals who were “doubled up,” meaning living with friends or family members, were at higher risk of post-release substance use than those housed in other settings. Studies of housing in the post-release period generally measure the effects of institutional or supportive housing on substance use outcomes, which has demonstrated significant reduction in substance use associated with residence in supportive housing for sufficient time in the post-release period. [39‐41] Individuals who are “doubled-up” with friends or family members may be at particularly high risk for illicit substance use due to lack of institutional support or exposure to acquaintances also using substances. Another important consideration is that participants living outside of institutional settings may have been under less drug testing surveillance, both from the program, but also from the state.

There were several limitations to our study. Our data comes from a cross sectional survey so we cannot make any statements about causality. Refusal to participate in the study was not systematically collected, which could affect generalizability of substance use estimates. The substance use outcomes are based on self-report, and a summary of data was shared with clinicians, so participants may have under-reported substance use. However, some studies with this population have demonstrated higher rates of substance use upon self-report in comparison to urine drug testing. [25] This study’s substance use outcomes also include cannabis, which may not be appropriate in states where it is currently legal, but is still important nationally as positive drug tests are a common reason for re-incarceration [42]. This was a secondary analysis, and our multivariable regression model was exploratory, so associations should be confirmed in studies specifically designed to test these hypotheses. Finally, the participants were older than most cohorts of formerly incarcerated individuals. Also, we only enrolled participants who engaged in primary care. Therefore, younger individuals and those who do not engage in medical care may have higher rates of substance use.

Due to the high volume of prison releases annually, high prevalence of SUD diagnoses in this population, and high recidivism rate, more studies are needed to understand substance use following release from incarceration. Our data suggests that overall substance use may be lower than expected post-release, but highlights some areas—such as less supervised housing—where substance use may be more common. Preventing negative consequences of substance use post-release should be a high priority for clinicians and policy-makers. Substance use education and treatment services should be available post-release and targeted to those with greatest treatment needs.