Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2022 Aug 1.
Published in final edited form as: J Subst Abuse Treat. 2021 Mar 5;127:108354. doi: 10.1016/j.jsat.2021.108354

Pre-incarceration polysubstance use involving opioids: A unique risk factor of postrelease return to substance use

Amanda M Bunting 1,*, Carrie B Oser 1,2, Michele Staton 2,3, Hannah K Knudsen 2,3
PMCID: PMC8217721  NIHMSID: NIHMS1680797  PMID: 34134861

Abstract

Objective:

Justice-involved populations are at increased risk of overdose following release from prison and jail. This risk is exacerbated by polysubstance use, including the use of opioids with other substances. This study explored pre-incarceration polysubstance use involving opioids as a unique risk factor for postrelease return to substance use.

Methods:

The study examined data from a cohort of 501 justice-involved persons who were enrolled in a therapeutic community treatment program while incarcerated. Latent profile validation identified profiles of polysubstance use involving opioids prior to incarceration. Multivariate logistic regression examined return to substance use, defined as self-reported relapse, and a time series model examined time in the community until a relapse event occurred.

Results:

A latent profile validation found six unique polysubstance opioid patterns prior to incarceration. Two of these profiles, primarily alcohol and primarily buprenorphine, were at increased and accelerated risk for relapse postrelease relative to a less polysubstance use profile. Both profiles at increased risk had pre-incarceration co-use of marijuana (≈45% of month) and nonmedical use of opioids (≈40% of month) but were unique in their respective near daily use of alcohol and nonmedical buprenorphine.

Conclusions:

Among persons who use opioids returning to the community, return to substance use occurs along a continuum of risk. Providers’ consideration of polysubstance use patterns during treatment may assist in mitigating adverse outcomes for patients postrelease.

Keywords: Opioids, Polysubstance use, Prison, Criminal justice, Reentry

1. Introduction

Substance use histories are endemic among individuals involved in the criminal justice system. More than 50% of individuals in state prisons and jails meet the criteria for a substance use disorder compared to only 5% of the general population (Bronson et al., 2017; James & Glaze, 2006). This disparity becomes more pressing in the current era of opioid use and the overdose epidemic. Individuals with justice involvement have long faced an increased risk of overdose and death following release from prison or jail (Binswanger, 2007), owing to returning use of substances after prolonged periods of substance abstinence and diminished tolerance. Research has found return to use of opioids post-release occurs in as many as 75% of formerly incarcerated persons (Fox et al., 2015a; Kinlock et al., 2008; Winter et al., 2015) as soon as 1 month postrelease (Binswanger et al., 2013; Lee et al., 2015). In Washington State the leading cause of death among formerly justice-involved individuals was overdose from opioids (Binswanger et al., 2013) and this trend likely extends across the United States. Notably from 1999 to 2009, research found more than 50% of the overdose deaths among formerly justice-involved persons to be due to use of opioids and other substances (Binswanger et al., 2013).

Polysubstance use (PSU) refers to these unique patterns of use of more than one drug. PSU can be simultaneous (i.e., two or more substances at the same time), sequential (i.e., one substance followed by another), or regular interval (i.e., two or more substances used in the same day/week/month). PSU involving opioids with a substance from another class is increasingly common and is a substantial contributor to overdose deaths due to the respiratory depression effects that are heightened when opioids are mixed with substances such as alcohol or benzodiazepines (Jones et al., 2018; Ruhm, 2017). In 2019, overdose due to the co-use of opioids and stimulants increased (O’Donnell, 2020) and the co-use of these substances is a concern in the current opioid epidemic (Cicero et al., 2019; Fogger, 2019).

In addition to risk of overdose, PSU is a unique risk factor of postrelease substance use. Individuals who engage in PSU tend to be younger with more severe substance use histories and justice histories (Betts et al., 2016; Darke & Hall, 1995; Green et al., 2011). More severe substance use (e.g., multiple substance use disorders, longer duration of use, injection use) and a greater number of substances used are associated with postrelease return to substance use (Chalana et al., 2016; Håkansson & Berglund, 2012; Kopak et al., 2016), indicating that a return to substance use postrelease may be more likely for individuals who engage in more prolific PSU prior to incarceration. Further, the likelihood to return to substance use following treatment may differ based on primary substance use (Nordfjærn, 2011), indicating that PSU patterns will affect return to use in ways not yet explored. Other risk factors for return to use include comorbid mental health diagnoses, lack of social supports, and socioeconomic factors (e.g., unemployment, homelessness; Hser et al., 2007; Kopak et al., 2016; Manuel et al., 2017)—all factors that are more common among justice-involved populations (James & Glaze, 2006; Western, 2006). Compounded with unique challenges to reentry such as barriers to housing, employment, and health care, individuals returning to the community are significantly burdened to meet their basic needs and at-risk following a period of prolonged abstinence (Mallik-Kane & Visher, 2008; Sung et al., 2011; Western, 2006). Given the prevalence of substance use among justice-involved persons and the other challenges that they face to reentry, understanding patterns of PSU and other factors that may be associated with postrelease return to substance use is important.

1.1. Postrelease substance use: Defining relapse

When considering postrelease return to substance use, sometimes referred to as relapse, we should first discuss the relevant theoretical and operational definitions of the term. Historically, research has viewed substance use outcomes as dichotomous such that an individual is either abstinent or relapsed because of any substance use (Miller, 1996), without consideration of whether the individual views their return to use as actually constituting an ongoing and problematic return to use. Dichotomous abstinent/relapse definitions that researchers have imposed based on any use imbue moral implications and fail to consider the gradual process of recovery (Miller, 1996; White & Ali, 2010.; White, 2007). Moreover, the use/no-use dichotomy lacks consideration of the chronic disease model of addiction, with remission defined as sustained abstinence, no longer meeting DSM criteria, and/or no problems related to substance use (Bradizza et al., 2006; Calabria et al., 2010). In the late 1980s, some researchers began to focus on recovery, considering it as "drug abuse and related behavior [that] are no longer problematic in the individual's life" (Leukefeld & Tims, 1986, p.185)—a conceptual frame that does not require complete abstinence as a measure of success. In 2010, the Substance Abuse and Mental Health Services Administration defined recovery as, "a process of change through which individuals improve their health and wellness, live a self-directed life, and strive to reach their full potential" while concurrently stating that abstinence is "the goal for those with addictions" (SAMHSA, 2012). While "recovery" may not require abstinence, a "relapse" to use does pose an inherent risk to one’s ability to sustain the success as outlined in these definitions of recovery.

Given an emphasis on person-first language and autonomy of individuals as active agents in their own recovery (White & Ali, 2010), the current research allows individuals to determine if their postrelease substance use was a relapse event. Despite a move to eliminate stigmatizing language, the term relapse remains in use and well understood by individuals in treatment programs (e.g., Binswanger et al., 2012; Fox et al., 2015b). Thereby when an individual reports that they have relapsed, they may be reporting that they have returned to problematic use as perceived in their own life. Further, the timing of this self-reported relapse event (i.e., how long individuals remain in the community before a relapse occurs) has important implications for reentry and is considered as an additional outcome in this study.

1.2. Current research

The current study sought to understand the risk factors associated with postrelease return to substance use, defined as self-reported relapse, among a population of individuals who engage in opioid PSU. While prior research has indicated high rates of postrelease relapse for individuals who use opioids (Binswanger et al., 2013; Fox et al., 2015a; Kinlock et al., 2008; Lee et al., 2015), how pre-incarceration PSU patterns, in conjunction with relevant risk factors for justice populations, influence postrelease return to substance use remains unclear. The goal of this study was to identify significant pre-incarceration PSU opioid patterns as independent predictors in a multivariate model of self-defined relapse among a justice-involved sample enrolled in a substance use treatment program while incarcerated.

2. Methods

2.1. Sample

This study includes 501 participants from the Criminal Justice Kentucky Treatment Outcome Study (CJKTOS). CJKTOS monitors outcomes of substance use programs (SAP) operated by the Kentucky Department of Corrections (DOC), through a partnership with the University of Kentucky Center on Drug and Alcohol Research (UK CDAR). The SAP is a six-month long program following the therapeutic community model (DeLeon, 2000). Individuals entering DOC prisons, jails, and community custody are eligible for entry to a SAP if they have 12–24 months remaining to serve on their sentences, a reported substance use history, and no recent disciplinary violations. Consent to baseline assessment is included in DOC consent to treatment and written consent at baseline is obtained for individuals who wish to be contacted for a follow-up survey postrelease.

DOC staff conduct baseline assessments at SAP entry using computer assisted personal interview (CAPI) software. UK CDAR uses telephone CAPI software for a single follow-up survey 12-months postrelease. A random sample proportionate to the number of males and females released from prison, jails, and community custody programs are selected for inclusion with a yearly target sample of 350. Individuals were ineligible for follow-up if they moved out of state (n=31) or were deceased (n=13). However, individuals who did not complete SAP were eligible for follow-up. Generally, persons may terminate the program early if released earlier than expected or if removed from the program for disciplinary reasons. The study obtained a federal Certificate of Confidentiality and informed individuals that their follow-up information would not be shared with the DOC.

The current study utilizes data from a cohort that participated in follow-up surveys from 2015 to 2017 and who provided both baseline and follow-up data, which included 1,044 individuals. Rates of participation in the follow-up survey were high (80% in 2015, 83% in 2016, and 84% in 2017). To address the study aim of examining the associations specific to pre-incarceration PSU involving opioids, the study applied three inclusion criteria to the sample. The following three criteria ensured that the sample had pre-incarceration use (i.e., were not part of the community SAP sample) and had histories of both opioid use and PSU. The analytic sample was then limited to (1) individuals from prison or jail SAP only (n=982), (2) with a history of opioid use in the 12 months prior to incarceration (n=816), and (3) who reported using more than one substance on a given day (i.e., a PSU population), which resulted in a final sample of 501 individuals. In the final sample, 77.6% graduated from the program, meaning that they completed the full six-month program.

2.2. Variables

The primary outcome was a dichotomous measure of self-reported postrelease relapse. Individuals were asked, "Would you consider your [alcohol]/[substance] use a relapse?" (1=relapse, 0=no relapse or no use). An additional dependent variable measured days until relapse through self-report. Individuals were asked, "How many days were you out on the street before you first used any illegal substance/alcohol" (0–365). The study included this variable for a time-series model.

The study measured independent variables of polysubstance use patterns through assignment to identified latent profiles based on pre-incarceration substance use. Substance use questions came from the Addiction Severity Index (McLellan et al., 1992). At baseline, the study asked individuals if they had used a given substance in the 12 months prior to their incarceration. For each substance that an individual reported using, the study then asked how many days in the 30 days prior to incarceration they had used the substance (R: 0–30). Substance use frequency is not used in the current research to indicate the presence of a substance use disorder, but rather to determine the possible overlap of days of substance use (i.e., polysubstance use). Latent profiles were created based on 30-day use of alcohol, cocaine, marijuana, heroin, nonmedical use of buprenorphine, nonmedical use of prescription opioids (NMPO), amphetamines, and nonmedical use of tranquilizers. Amphetamines included use of methamphetamine and nonmedical use of prescription amphetamines. Tranquilizers included nonmedical use of benzodiazepines, ketamine, and muscle relaxers.

Although PSU is the substantive variable of interest, the models controlled for additional sociodemographic and health variables that have been associated with return to substance use in prior research (e.g., Binswanger et al., 2012; Chatav Schonbrun et al., 2013; Evans et al., 2009; Krishnan et al., 2013; Seal et al., 2001). The study drew all sociodemographic variables from baseline assessments. Variables included age (measured continuously in years), education (measured continuously with GED=12 years), race1 (1=white, 0=nonwhite), marital status (1=married, 0=single/divorced/widowed), gender (1=male), pre-incarceration employment (1=unemployed), homelessness in the 12 months prior to incarceration (1=homeless), and the number of years the individual was incarcerated. The study measured economic hardship using a summative scale adapted from the Survey of Income and Program Participation (R:0–8, α=0.87), which includes eight dichotomous measures of difficulty meeting needs of food, housing, clothing, and medical care (Beverly, 2001). Higher scores indicate greater economic hardship. The study coded the county that an individual lived in prior to incarceration using a rural-urban coding scheme (Ingram & Franco, 2014), collapsed to a dichotomous measurement such that 1=rural and 0=urban.

The study obtained all health variables from baseline assessments. A dichotomous variable measured individual’s chronic pain as described to them as pain persisting or recurring for 3 months or longer (1=yes). The study measured pre-incarceration physical and mental health using two continuous variables from the Behavioral Risk Factor Surveillance (BRFSS) that measure the number of self-reported poor physical and poor mental health days in the 30 days prior to incarceration (CDC, 2019; Hennessy et al., 1994). We also measured lifetime injection drug use history (1=yes) and history of the hepatitis C virus infection (HCV; 1=yes).

The study measured anxiety and depression symptoms in the 12 months prior to incarceration using an adapted version of the Patient Health Questionnaire-9 (PHQ-9; Spitzer, Kroenke, Williams, & Group, 1999) (R: 0–9), and the Generalized Anxiety Disorder-7 (GAD-7; Spitzer, Kroenke, Williams, & Löwe, 2006) (R:0–7). In the current study the internal reliability was strong for both scales (PHQ-9 α=0.93 and GAD-7 α=0.97).

2.3. Analytic plan

Previously, research has found latent profiles of polysubstance using the baseline sample of the CJKTOS population from 2015 to 2017, which included 6,569 individuals and yielded a six-profile model solution (Bunting et al., 2020). The current study of 501 individuals with both baseline and 12-month postrelease follow-up data replicated the latent profile analysis. The process of replication, often referred to in latent modeling literature as the process of validation, is a common sensitivity analysis to determine if profiles exist in different samples (Collins & Lanza, 2010). To validate the six-profile solution previously found, this study applied the posterior probabilities of profile membership from the baseline study to the current sample of 501. We then entered the 30-day substance use indicators in the current sample into a latent profile model with 6 profiles. The baseline study provided posterior probabilities (Bunting et al., 2020) as starting values for the current model using Stata’s gsem function. This function first refines starting values using an expectation maximization (EM) algorithm, a complete-data likelihood, before maximizing the likelihood. In this way, providing starting values assists in model convergence. The validation model successfully converged. Additional analysis of selection criteria indicated that the six-profiles were the best fit for the data compared to models fitting 1–5 profiles. We based selection criteria on standard fit indices of AIC, BIC, and entropy. Based on these results, the study determined the baseline sample to have utility at classifying the follow-up sample. Next, the study compared latent indicators of mean number of prior 30-day use to the previous analysis (Bunting et al., 2020) and the follow-up profiles remained the same in terms of substantive meaning and structure. After the study determined substantive meanings to remain consistent across the samples, the study examined specific means of prior 30-day use. The current sample had fewer days of mean use of amphetamines, but this did not affect profile structure or substantive meaning. The entropy value for the six-profile solution was 0.94, indicating excellent model classification.

The study used a latent profile analysis with a distal outcome to examine the dependent variables. Given modeling complexity with continuous profile indicators and that entropy in the current data was excellent (0.94), indicating profile membership probabilities were high, the study selected the classify-analyze approach to latent modeling with distal outcomes. In this approach, the study assigned individuals to their latent profile based on their maximum posterior probability in the first step, and the outcome analysis is completed in the second step (Nagin, 2005). Latent profiles were validated, then assigned to their most likely profile membership. Next a multivariate logistic regression estimated the dependent variable of self-reported relapse. A second model examined the days until relapse using a Cox-proportional hazard model. In the Cox-proportional hazard model, individuals were right censored such that those not reporting relapse were given a value of 365 days. Five individuals reported a relapse at 0 days, indicating they relapsed immediately after release. To be included in the Cox model which requires positive integers, the study coded these individuals as having a relapse event at day 1. Diagnostics for collinearity revealed no issues with variance inflation factors less than 2.0. The study performed all analyses with Stata (SE) 15.1 (StataCorp, College Station, TX).

3. Results

3.1. Demographics

Table 2 details sample characteristics. The total sample relapse rate was 40%, with an average of 244 days until relapse (right-censored). In a truncated sample (i.e., only those who self-report a relapse) the average was 156 days until relapse. The sample was on average age 33 years old, with 13 years of education at baseline, and consisted of predominantly White, nonmarried males. The majority of participants reported unemployment prior to incarceration (64.9%). Nearly one in five participants were homeless prior to incarceration. Individuals were incarcerated an average of two years before entering the treatment program. On average, they reported two instances of economic hardship (R:0–8). The sample was nearly evenly split between those who resided in rural and urban counties prior to incarceration.

Table 2.

Latent profile conditional means for polysubstance opioid use (n=501).

Profile 1 Profile 2 Profile 3 Profile 4 Profile 5 Profile 6
Descriptive profile abbreviation Primarily Alcohol Primarily Heroin Less PSU Tranquilizer PSU Primarily Buprenorphine Stimulant-Opioid
Latent Profile indicators:
Prior 30- day use
 Alcohol 27.66 3.81 2.80 10.27 6.15 8.68
 Cocaine 3.08 1.25 0.62 1.97 1.22 28.13
 Marijuana 13.95 11.24 11.98 13.68 13.67 14.90
 Heroin 1.77 28.85 0.89 7.45 1.94 13.55
 Buprenorphine 1.45 2.71 1.41 9.36 28.49 7.97
 NMPO 11.14 15.14 16.51 19.79 13.30 17.21
 Amphetamines 6.82 5.77 8.09 9.84 8.37 6.09
 Tranquilizers 2.85 4.29 2.77 28.71 4.23 11.11
Profile Prevalence 6.2% 15.9% 42.9% 15.2% 8.6% 11.2%
Open in a new tab

Note: 30-day use refers to the 30-days prior to incarceration; substances with use above 50% of the month are shaded

Approximately one-third of the sample reported chronic pain, with an average of a week of poor physical health and 11 days of poor mental health in the 30 days prior to incarceration. A majority (61.9%) reported lifetime IDU history and 14% reported HCV positive serostatus. On average, individuals met the criteria for mild to moderate depression and anxiety at baseline as indicated by average scores above the value of two.

3.2. Latent profiles

The six PSU profiles (see Table 2) were given descriptive profile titles consistent with the previous study. The primarily alcohol profile consisted of individuals with near daily use of alcohol and co-use of marijuana and NMPO on about 40% of days during the month. The primarily heroin profile was characterized by near daily use of heroin along with co-use of NMPO 50% of the month and marijuana about 40% of the month. Individuals in the less PSU profile had less diverse patterns of use than other profiles, but had NMPO and marijuana use 40– 50% of the month. The tranquilizer PSU profile was differentiated by diverse PSU to include near daily use of tranquilizers and co-use of marijuana and NMPO 50–70% of the month. The primarily buprenorphine profile was unique in that individuals had near daily use of nonmedical buprenorphine and co-use of marijuana and amphetamines about 40% of the month. The stimulant-opioid profile was characterized by diverse PSU patterns to include near daily use of cocaine and co-use of marijuana, heroin, and NMPO 50–60% of the month. Additional information on the differences among profiles on the dependent variables and controls are in Table 1.

Table 1.

Characteristics of study population (n=501).

Total Sample Primarily Alcohol Primarily Heroin Less PSU Tranquilizer PSU Primarily Buprenorphine Stimulant-Opioid
Dependent Variable
Post-release relapse 40.5 58.1 41.2 34.4 40.8 58.1 39.3
 Days until relapse Right-censored 244.35 (153.15) 173.64 (169.01) 245.79 (152.60) 261.37 (148.07) 249.43 (148.87) 191.28 (156.07) 249.95 (154.42)
 Days until relapse Truncated 156.56 (149.06) 35.44 (46.71) 162.08 (150.47) 171.25 (153.45) 174.06 (149.08) 107.41 (119.01) 180.91 (159.67)
Sociodemographic
 Age 32.55 (7.67) 33.84 (8.61) 30.54 (6.71) 32.99 (7.70) 33.33 (8.82) 31.81 (6.97) 32.55 (6.91)
 Education level 11.87 (2.13) 12.06 (1.65) 11.77 (2.10) 11.83 (2.17) 12.21 (2.53) 11.65 (1.97) 11.78 (1.81)
 White 84.8 77.4 90.0 86.0 86.8 83.7 75.0
 Unemployed 64.9 71.0 38.7 34.4 35.5 20.9 46.4
 Male 72.6 80.6 68.7 72.1 67.1 79.1 78.6
 Married 21.4 12.9 22.5 20.9 19.7 30.2 21.4
 Homeless 20.2 16.1 28.7 14.4 23.7 16.3 30.4
 Years incarcerated 2.23 (1.74) 2.60 (2.36) 1.71 (1.25) 2.21 (1.72) 2.61 (1.54) 2.17 (1.46) 2.41 (2.23)
 Economic hardship (R:0–8) 2.10 (2.52) 2.42 (2.78) 2.24 (2.75) 1.91 (2.40) 2.25 (2.57) 1.70 (2.11) 2.61 (2.67)
 Rural 52.7 58.1 23.7 59.5 56.6 79.1 39.3
Health
 Chronic pain 30.3 32.3 26.2 29.3 44.7 23.3 25.0
 Poor physical health days past month 7.40 (11.97) 9.42 (13.45) 7.15 (12.12) 6.47 (11.35) 9.25 (12.88) 8.32 (12.03) 6.96 (12.01)
 Poor mental health days past month 11.54 (13.80) 13.52 (14.15) 12.52 (14.30) 10.12 (13.32) 15.31 (14.21) 9.51 (13.83) 10.91 (13.54)
 IDU 61.9 48.4 81.2 50.7 71.0 65.1 69.6
 HCV 14.2 9.7 23.7 9.8 18.4 18.6 10.7
 Depression (R:0–9) 4.44 (3.52) 4.26 (3.55) 5.07 (3.43) 4.01 (3.53) 5.61 (3.31) 3.35 (3.48) 4.55 (3.46)
 Anxiety (R:0–7) 3.71 (3.23) 3.84 (3.36) 3.51 (3.18) 3.30 (3.22) 4.68 (3.12) 3.74 (3.32) 4.14 (3.15)
Open in a new tab

Note: Means (standard deviations) presented for continuous variables and percentages presented for nominal variables.

In analyses, we selected the less PSU profile as the reference category. The study team chose this profile so that analyses would indicate how more risk-engaging profiles differed. Additionally, it was the most prevalent profile, representing 42.9% of the sample, making it an ideal reference category.

3.3. Multivariate models

A logistic regression model of self-reported relapse is presented in model 1 of Table 3. Compared to the reference category of less PSU, individuals categorized by the primarily alcohol (OR: 2.93, p<.01) and primarily buprenorphine (OR: 2.32, p<.05) profiles were more likely to report a postrelease relapse. Age was significant, such that older individuals were less likely to report a relapse (OR: 0.93, p<.001). One health variable was significant—self-reported positive HCV status—indicating that those with HCV were more likely to report relapse (OR: 2.32, p<.01).

Table 3.

Multivariate models predicting self-reported post-release relapse; (n=501).

Model 1 Model 2
Odds Ratio 95% CI Hazard Ratio 95% CI
Latent Profiles
Primarily Alcohol 2.93** 1.30–6.61 2.65*** 1.57–4.48
Primarily Heroin 1.09 0.60–1.98 1.09 0.69–1.72
Tranquilizer PSU 1.18 0.66–2.12 1.08 0.69–1.67
Primarily Buprenorphine 2.32* 1.14–4.73 1.84* 1.15–2.95
Stimulants-Opioid 1.22 0.64–2.33 1.27 0.77–2.08
Sociodemographic
Age 0 93*** 0.91–0.96 0.95*** 0.93–0.97
Education level 0.94 0.86–1.03 0.96 0.90–1.03
White 1.39 0.79–2.43 1.31 0.84–2.02
Unemployed 0.94 0.62–1.44 0.91 0.67–1.24
Male 1.60 0.99–2.57 1.47* 1.02–2.11
Married 1.04 0.65–1.67 0.97 0.68–1.37
Homeless 0.74 0.44–1.24 0.75 0.50–1.11
Years incarcerated 1.11 0.98–1.25 1.07 0.99–1.16
Economic hardship (R:0–8) 0.98 0.90–1.07 0.97 0.93–1.06
Rural 1.05 0.69–1.60 1.05 0.77–1.45
Health
Chronic pain 1.48 0.93–2.35 1.40* 1.00–1.95
Poor physical health days past month 0.99 0.97–1.01 1.00 0.98–1.01
Poor mental health days past month 1.00 0.98–1.02 1.00 0.99–1.01
Lifetime IDU 1.16 0.75–1.80 1.12 0.80–1.56
HCV 2.32** 1.28–4.21 1.67* 1.12–2.50
Depression (R:0–9) 0.98 0.91–1.06 0.99 0.94–1.05
Anxiety (R:0–7) 1.04 0.97–1.13 1.03 0.97–1.09
Model Fit
LR Chi-square 54.04*** 55.84***
Pseudo R2 0.0799 N/A
Open in a new tab

Notes: Model 1 reports results from a logistic regression of self-reported relapse. Model 2 reports hazard ratios of days in the community until self-reported relapse. Latent profile 3, representing less PSU, is reference group.

*

p<.05

**

p<.01

***

p<.001

Model 2 of Table 3 reports the hazard ratios for a Cox-proportional hazard model examining time in the community until self-reported relapse. Individuals were right-censored such that those who did not experience a relapse were given the value of 365 (max number of days in follow-up period). In the time-series model, individuals characterized by primarily alcohol (HR: 2.65, p<.001) and primarily buprenorphine (HR: 1.84, p<.05) PSU patterns were at accelerated risk for relapse. Men (HR: 1.47, p<.05), those with chronic pain (HR: 1.40, p<.05), and persons with HCV (HR: 1.67, p<.05) were also at accelerated risk for relapse, meaning that the time to first use occurred sooner for men, those with chronic pain, and those with a history of HCV.

4. Discussion

The current study explored the association of pre-incarceration PSU opioid patterns with postrelease substance use as self-defined relapse by persons who enrolled in a prison substance use treatment program. The relapse rate in the current justice-involved population was 40% and ranged from 34% to 58% by PSU profiles indicating heterogeneity by substance patterns. This rate is higher than general prison populations postrelease (Chamberlain et al., 2019 [18–23%]), but similar to postrelease relapse rates found in other research of persons who use opioids (Fox et al., 2015a [>66%]; Kinlock et al., 2008 [44%]; Lee et al., 2015 [38–88%]).

The study found two pre-incarceration PSU profiles, primarily alcohol and primarily buprenorphine, to be at increased and accelerated risk for relapse. Although some research has indicated risk factors are similar for returning to use and proposed that the process is similar across substances (Witkiewitz & Marlatt, 2004, 2007), these findings support other studies that have found that recovery paths differ by substances (Castro et al., 2000; Hser et al., 2007).

Individuals characterized by pre-incarceration primarily alcohol pattern of use, with near daily drinking and substantial co-use of nonmedical opioids and marijuana, had lower than average risk factors such as pre-incarceration homelessness, IDU histories, and HCV, as would be expected given the substances regularly used. However, individuals in this group were more likely to self-report relapse even when controlling for other risk factors. Individuals’ return to alcohol use may be more influenced by proximal risks (e.g., immediate triggers) compared to distal risks (e.g., predispositions) (Witkiewitz & Masyn, 2008). For many individuals posttreatment, alcohol use tends toward the dichotomy of abstinence or excessive use, with the return to heavy drinking patterns tending to occur quickly among those who return to use (Hufford et al., 2003). In the current research, individuals in the primarily alcohol profile were not only likely to report a relapse, but to have a relapse event soon after returning to the community (x¯=35.44). Returning to the community encompasses myriad difficulties that may serve as a proximal trigger. Alcohol use is the most commonly reported substance used by men returning to communities from prison (Mallik-Kane & Visher, 2008), perhaps owing to its legal status and the ease with which it can be obtained.

Considering individuals self-reported if they experienced a relapse, individuals with this pattern of pre-incarceration primarily alcohol use may be more likely to consider their use a "relapse" event. Indeed, peer-recovery groups, typically based on twelve-step models, are prevalent in prisons and jails and mandated as part of parole (Leverentz, 2013), where any use is typically framed as relapse. The celebration of sobriety birthdays in programs such as Alcoholics Anonymous indicate a preference for abstinence, and individuals who have been involved in these programs may be more likely to self-cite relapse. However, differences in perception may not be the only issue, as the primarily alcohol profile had a significantly shorter time to first use after being released.

Individuals characterized by the primarily buprenorphine profile were also more likely to report a postrelease relapse event at an accelerated rate than those with less substantial PSU. This profile was unique in individual’s near daily use of nonmedical use of buprenorphine as well as co-use of marijuana and nonmedical use of opioids prior to incarceration. Individuals in this profile were the most likely to live in rural areas compared to the other profiles. Their location may have influenced their preferred substance patterns, as nonmedical use of buprenorphine and prescription opioids is more likely in rural Kentucky compared to urban populations (Young et al., 2010).

Research on nonmedical use of buprenorphine indicates that individuals report both euphoric effects as well as using it to self-treat opioid use disorder, particularly to reduce withdrawal symptoms (Bazazi et al., 2011; Hakansson et al., 2007; Yokell et al., 2011). This profile reported considerable nonmedical opioid use as well, so the buprenorphine use may reflect efforts to mitigate withdrawal. Since buprenorphine’s primary use is a treatment modality for opioid use disorder, we know less about the recovery trajectories of individuals who report primarily nonmedical use of buprenorphine without engaging in formal treatment through an office-based buprenorphine provider or specialty treatment setting. Given the association with rurality for this profile, if individuals return to rural areas postrelease, they may continue to experience a lack of treatment service availability, including lack of access to opioid use disorder treatment providers (Bunting et al., 2018). Alternatively, use and subsequent return to use may be associated with local drug market availability and unrelated to treatment or access to services.

While both the primarily alcohol and primarily buprenorphine profiles were more likely to self-report relapse early than other profiles, this is not necessarily an indicator of the severity of their substance use. However, in the context of justice-involved populations early return to use may be associated with worse outcomes as individuals are at an elevated risk for overdose in the first weeks after release (Binswanger et al., 2007).

The study found sociodemographic and health variables to be associated with postrelease relapse. Age was significantly associated with both measures of postrelease relapse. Substantial literature exists to support this relationship (Hser et al., 2007; Kopak et al., 2016). In both models, HCV status was a robust predictor of postrelease relapse. Complex physical health problems can exacerbate reentry difficulties (Binswanger et al., 2012). HCV is prevalent among individuals with more extensive substance use histories and IDU (Klinkenberg et al., 2003; Rosenberg et al., 2005; Shapatava et al., 2006). In this case, HCV may be a proxy to a severe drug use history, a significant factor in predicting return to use (Kopak et al., 2016; Nordfjærn, 2011). Further, there are risk behaviors correlated with HCV serostatus (Koblin et al., 2003; Vescio et al., 2008; Willner-Reid et al., 2008), and the significance of HCV in multivariate models may reflect underlying latent variables.

There are limitations of the current study to consider. The study operationalized PSU as regular interval use (i.e., use of two or more substances within the same timeframe) among a PSU sample. That is, while individuals had self-reported histories of using more than one substance on a given day, the latent profiles captured regular interval PSU and were not able to measure simultaneous PSU patterns. Future research should examine both regular interval and simultaneous use and their subsequent effects on postrelease behaviors. Future research may also consider how polysubstance use behaviors vary during different time periods prior to incarceration (e.g., at time of arrest, during release on bond, etc.). While the available data allowed for an analysis of patterns of regular interval PSU, they cannot speak to rates of substance use disorder as the dataset did not include DSM-V measures. Future research of PSU should consider measuring formal diagnosis of a substance use disorder. Moreover, these findings may not replicate across justice systems more broadly, as the sample consisted of treatment-seeking individuals in a single state. The data that this study included are based on self-reports from a justice-involved sample. While university research staff oversee the follow-up data collection, individuals may still have feared negative consequences from the DOC for reporting illegal behavior at follow-up and, therefore, underreported postrelease substance use.

The current research contributes to previous literature exploring the concept of "relapse" by using a dependent variable that allowed individuals to self-report if they had a relapse event during the follow-up period. This operationalization allowed for a unique definition of relapse that provided individuals with more autonomy over their substance use, allowing them to define a momentary postrelease lapse compared to a return to problematic or undesired use. Future research should continue to explore varying definitions of returning to substance use, including collecting data on adverse outcomes related to use. This study was also unique given its focus on PSU opioid patterns and highlighting the importance of PSU in the relapse literature.

4.1. Conclusions

The findings of the current study indicate that among users of opioids who return to the community after prison, there is a continuum of risk of relapse based on pre-incarceration polysubstance use. This study found an elevated risk of relapse among individuals who co-used marijuana and opiates with alcohol or nonmedical use of buprenorphine prior to incarceration. Understanding individual’s pre-incarceration patterns of PSU may assist in mitigating risk postrelease if appropriate interventions are made available. The period of incarceration provides opportunity to implement targeted interventions and provide aftercare planning, and this research indicates that tailored release planning that considers pre-incarceration substance use patterns may be warranted. Interventions that reduce postrelease substance use can mitigate adverse outcomes for formerly justice-involved individuals, and provide cost-savings for communities and justice systems.

Highlights.

  • Two pre-incarceration use polysubstance profiles were at increased risk of post-release relapse

  • Patterns of risk were unique in their near daily use of alcohol and use of nonmedical buprenorphine

  • Both at-risk profiles co-used with marijuana and nonmedical opioids

  • Polysubstance use is a unique risk factor for post-release return to substance use

Role of Funding Source

Supported by National Institute on Drug Abuse Grants T32DA035200, R25DA037190 (Bunting; PI: Rush, PI: Beckwith) and K02DA035116 (PI: Oser) and Agency for Healthcare Research and Quality Grant T32HS026120 (Bunting; PI: Schwartz). The opinions expressed are those of the authors.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

1

The study dichotomized race as prevalence in the current sample was 84.8% White, 8.4% Black, 6.8% multiracial/other.

References

  1. Bazazi AR, Yokell M, Fu JJ, Rich JD, & Zaller ND (2011). Illicit Use of Buprenorphine/Naloxone Among Injecting and Noninjecting Opioid Users. Journal of Addiction Medicine, 5(3), 175–180. 10.1097/ADM.0b013e3182034e31 [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Betts KS, Chan G, McIlwraith F, Dietze P, Whittaker E, Burns L, & Alati R (2016). Differences in polysubstance use patterns and drug-related outcomes between people who inject drugs receiving and not receiving opioid substitution therapies. Addiction, 111(7), 1214–1223. 10.1111/add.13339 [DOI] [PubMed] [Google Scholar]
  3. Beverly S (2001). Material hardship in the United States: Evidence from the Survey of Income and Program Participation. Social Work Research, 25(3), 143–151. [Google Scholar]
  4. Binswanger IA, Blatchford PJ, Mueller SR, & Stern MF (2013). Mortality After Prison Release: Opioid Overdose and Other Causes of Death, Risk Factors, and Time Trends From 1999 to 2009. Annals of Internal Medicine, 159(9), 592–600. 10.7326/0003-4819-159-9-201311050-00005 [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Binswanger IA, Nowels C, Corsi KF, Glanz J, Long J, Booth RE, & Steiner JF (2012). Return to drug use and overdose after release from prison: A qualitative study of risk and protective factors. Addiction Science & Clinical Practice, 7(1), 3. 10.1186/1940-0640-7-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Binswanger IA, Stern MF, Deyo RA, Heagerty PJ, Cheadle A, Elmore JG, & Koepsell TD (2007). Release from prison—A high risk of death for former inmates. The New England Journal of Medicine, 356(2), 157–165. 10.1056/NEJMsa064115 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Bradizza CM, Stasiewicz PR, & Paas ND (2006). Relapse to alcohol and drug use among individuals diagnosed with co-occurring mental health and substance use disorders: A review. Clinical Psychology Review, 26(2), 162–178. 10.1016/j.cpr.2005.11.005 [DOI] [PubMed] [Google Scholar]
  8. Bronson J, Stroop J, Zimmer S, & Berzofsky M (2017). Drug Use, Dependence, and Abuse Among State Prisoners and Jail Inmates, 2007–2009. 27. [Google Scholar]
  9. Bunting AM, Oser CB, Staton M, Eddens KS, & Knudsen H (2018). Clinician identified barriers to treatment for individuals in Appalachia with opioid use disorder following release from prison: A social ecological approach. Addiction Science & Clinical Practice, 13(1), 23. 10.1186/s13722-018-0124-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Bunting AM, Oser CB, Staton M, & Knudsen H (2020). Polysubstance use patterns among justice-involved individuals who use opioids. Substance Use & Misuse, 55(13), 2165–2174. 10.1080/10826084.2020.1795683 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Calabria B, Degenhardt L, Briegleb C, Vos T, Hall W, Lynskey M, Callaghan B, Rana U, & McLaren J (2010). Systematic review of prospective studies investigating "remission" from amphetamine, cannabis, cocaine or opioid dependence. Addictive Behaviors, 35(8), 741–749. 10.1016/j.addbeh.2010.03.019 [DOI] [PubMed] [Google Scholar]
  12. Castro FG, Barrington EH, Walton MA, & Rawson RA (2000). Cocaine and Methamphetamine: Differential Addiction Rates. Psychology of Addictive Behaviors, 14(4), 390–396. [PubMed] [Google Scholar]
  13. CDC. (2019, September 27). HIV among people who inject drugs | HIV by Group j HIV/AIDS. https://www.cdc.gov/hiv/group/hiv-idu.html
  14. Chalana H, Kundal T, Gupta V, & Malhari AS (2016, September 18). Predictors of Relapse after Inpatient Opioid Detoxification during 1-Year Follow-Up [Research Article]. Journal of Addiction; Hindawi. 10.1155/2016/7620860 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Chamberlain A, Nyamu S, Aminawung J, Wang EA, Shavit S, & Fox AD (2019). Illicit substance use after release from prison among formerly incarcerated primary care patients: A cross-sectional study. Addiction Science & Clinical Practice, 14(1), 7. 10.1186/s13722-019-0136-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Chatav Schonbrun Y, Kuo CC, Nargiso EJ, Zlotnick C, Williams AC, Shefner TR, & Johnson EJ (2013). "I know if I drink I won’t feel anything": Substance use relapse among depressed women leaving prison. International Journal of Prisoner Health, 9(4), 169–186. 10.1108/IJPH-02-2013-0009 [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Cicero TJ, Ellis MS, & Kasper ZA (2019). Polysubstance Use: A Broader Understanding of Substance Use During the Opioid Crisis. American Journal of Public Health, 110(2), 244–250. 10.2105/AJPH.2019.305412 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Darke S, & Hall W (1995). Levels and correlates of polydrug use among heroin users and regular amphetamine users. Drug and Alcohol Dependence, 39(3), 231–235. 10.1016/0376-8716(95)01171-9 [DOI] [PubMed] [Google Scholar]
  19. De Leon G The therapeutic community: Theory, model, and method. Springer Publishing Co; 2000. [Google Scholar]
  20. Evans JL, Hahn JA, Lum PJ, Stein ES, & Page K (2009). Predictors of injection drug use cessation and relapse in a prospective cohort of young injection drug users in San Francisco, CA (UFO Study). Drug and Alcohol Dependence, 101(3), 152–157. 10.1016/j.drugalcdep.2008.12.007 [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Fogger SA (2019). Methamphetamine Use: A New Wave in the Opioid Crisis? Journal of Addictions Nursing, 30(3), 219–223. 10.1097/JAN.0000000000000298 [DOI] [PubMed] [Google Scholar]
  22. Fox AD, Maradiaga J, Weiss L, Sanchez J, Starrels JL, & Cunningham CO (2015a). Release from incarceration, relapse to opioid use and the potential for buprenorphine maintenance treatment: A qualitative study of the perceptions of former inmates with opioid use disorder. Addiction Science & Clinical Practice, 10(1), 2. 10.1186/s13722-014-0023-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Fox AD, Maradiaga J, Weiss L, Sanchez J, Starrels JL, & Cunningham CO (2015b). Release from incarceration, relapse to opioid use and the potential for buprenorphine maintenance treatment: A qualitative study of the perceptions of former inmates with opioid use disorder. Addiction Science & Clinical Practice, 10(1), 2. 10.1186/s13722-014-0023-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Green TC, Black R, Serrano JMG, Budman SH, & Butler SF (2011). Typologies of Prescription Opioid Use in a Large Sample of Adults Assessed for Substance Abuse Treatment. PLOS ONE, 6(11), e27244. 10.1371/journal.pone.0027244 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Håkansson A, & Berglund M (2012). Risk factors for criminal recidivism – a prospective follow-up study in prisoners with substance abuse. BMC Psychiatry, 12, 111. 10.1186/1471-244X-12-111 [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Hakansson A, Medvedeo A, Andersson M, & Berglund M (2007). Buprenorphine misuse among heroin and amphetamine users in Malmo, Sweden: Purpose of misuse and route of administration. European Addiction Research, 13(4), 207–215. 10.1159/000104883 [DOI] [PubMed] [Google Scholar]
  27. Hennessy CH, Moriarty DG, Zack MM, Scherr PA, & Brackbill R (1994). Measuring health-related quality of life for public health surveillance. Public Health Reports (Washington, D.C.: 1974), 109(5), 665–672. [PMC free article] [PubMed] [Google Scholar]
  28. Hser Y-I, Longshore D, & Anglin MD (2007). The Life Course Perspective on Drug Use: A Conceptual Framework for Understanding Drug Use Trajectories. Evaluation Review, 31(6), 515–547. 10.1177/0193841X07307316 [DOI] [PubMed] [Google Scholar]
  29. Hufford MR, Witkiewitz K, Shields AL, Kodya S, & Caruso JC (2003). Relapse as a nonlinear dynamic system: Application to patients with alcohol use disorders. Journal of Abnormal Psychology, 112(2), 219–227. 10.1037/0021-843X.112.2.219 [DOI] [PubMed] [Google Scholar]
  30. Ingram D, & Franco SJ (2014). 2013 NCHS urban–rural classification scheme for counties. Vital and Health Statistics, Series, 2(166), 173. [PubMed] [Google Scholar]
  31. James D, & Glaze L (2006). Mental Health Problems of Prison. 12. [Google Scholar]
  32. Jones CM, Einstein EB, & Compton WM (2018). Changes in Synthetic Opioid Involvement in Drug Overdose Deaths in the United States, 2010–2016. JAMA, 319(17), 1819–1821. 10.1001/jama.2018.2844 [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Kinlock TW, Gordon MS, Schwartz RP, & O’Grady KE (2008). A Study of Methadone Maintenance For Male Prisoners. Criminal Justice and Behavior, 35(1), 34–47. 10.1177/0093854807309111 [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Klinkenberg WD, Caslyn RJ, Morse GA, Yonker RD, McCudden S, Ketema F, & Constantine NT (2003). Prevalence of human immunodeficiency virus, hepatitis b, and hepatitis c among homeless persons with co-occurring severe mental illness and substance use disorders. Comprehensive Psychiatry, 44(4), 293–302. 10.1016/S0010-440X(03)00094-4 [DOI] [PubMed] [Google Scholar]
  35. Koblin BA, Factor SH, Wu Y, & Vlahov D (2003). Hepatitis C virus infection among noninjecting drug users in New York City. Journal of Medical Virology, 70(3), 387–390. 10.1002/jmv.10407 [DOI] [PubMed] [Google Scholar]
  36. Kopak AM, Hoffmann NG, & Proctor SL (2016). Key Risk Factors for Relapse and Rearrest Among Substance Use Treatment Patients Involved in the Criminal Justice System. American Journal of Criminal Justice, 41(1), 14–30. 10.1007/s12103-015-9330-6 [DOI] [Google Scholar]
  37. Krishnan A, Wickersham JA, Chitsaz E, Springer SA, Jordan AO, Zaller N, & Altice FL (2013). Post-Release Substance Abuse Outcomes Among HIV-Infected Jail Detainees: Results from a Multisite Study. AIDS and Behavior, 17(2), 171–180. 10.1007/s10461-012-0362-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Lee JD, McDonald R, Grossman E, McNeely J, Laska E, Rotrosen J, & Gourevitch MN (2015). Opioid treatment at release from jail using extended-release naltrexone: A pilot proof-of-concept randomized effectiveness trial. Addiction, 110(6), 1008–1014. 10.1111/add.12894 [DOI] [PubMed] [Google Scholar]
  39. Leverentz A (2013). The Ex-Prisoner’s Dilemma. New Brunswick: Rutgers University Press. [Google Scholar]
  40. Leukefeld CG, & Tims FM (1986). Relapse and recovery: Some directions for research and practice. In Tims F & Luekefeld C (Eds.), 17 Relapse and Recovery in Drug Abuse (NIDA Monograph 72, pp. 86–103). Rockville, MD: National Institute on Drug Abuse; [PubMed] [Google Scholar]
  41. Mallik-Kane K, & Visher CA (2008). Health and Prisoner Reentry: How Physical, Mental, and Substance Abuse Conditions Shape the Process of Reintegration: (719772011–001) [Data set]. American Psychological Association. 10.1037/e719772011-001 [DOI] [Google Scholar]
  42. Manuel JI, Yuan Y, Herman DB, Svikis DS, Nichols O, Palmer E, & Deren S (2017). Barriers and facilitators to successful transition from long-term residential substance abuse treatment. Journal of Substance Abuse Treatment, 74, 16–22. 10.1016/j.jsat.2016.12.001 [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. McLellan AT, Kushner H, Metzger D, Peters R, Smith I, Grissom G, Pettinati H, & Argeriou M (1992). The Fifth Edition of the Addiction Severity Index. Journal of Substance Abuse Treatment, 9(3), 199–213. 10.1016/0740-5472(92)90062-s [DOI] [PubMed] [Google Scholar]
  44. Miller WRM (1996). What is a relapse? Fifty ways to leave the wagon. 14. [PubMed]
  45. Nagin DS (2005). Group-based modeling of development. Cambridge, MA: Harvard University Press. [Google Scholar]
  46. Nordfjærn T (2011). Relapse patterns among patients with substance use disorders. Journal of Substance Use, 16(4), 313–329. 10.3109/14659890903580482 [DOI] [Google Scholar]
  47. O’Donnell J (2020). Vital Signs: Characteristics of Drug Overdose Deaths Involving Opioids and Stimulants — 24 States and the District of Columbia, January–June 2019. MMWR. Morbidity and Mortality Weekly Report, 69. 10.15585/mmwr.mm6935a1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Rosenberg SD, Drake RE, Brunette MF, Wolford GL, & Marsh BJ (2005). Hepatitis C virus and HIV co-infection in people with severe mental illness and substance use disorders. AIDS, 19, S26. 10.1097/01.aids.0000192067.94033.aa [DOI] [PubMed] [Google Scholar]
  49. Ruhm CJ (2017). Drug involvement in fatal overdoses. SSM - Population Health, 3(Supplement C), 219–226. 10.1016/j.ssmph.2017.01.009 [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Seal KH, Kral AH, Gee L, Moore LD, Bluthenthal RN, Lorvick J, & Edlin BR (2001). Predictors and Prevention of Nonfatal Overdose Among Street-Recruited Injection Heroin Users in the San Francisco Bay Area, 1998–1999. American Journal of Public Health, 91(11), 1842–1846. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Shapatava E, Nelson KE, Tsertsvadze T, & Rio C. del. (2006). Risk behaviors and HIV, hepatitis B, and hepatitis C seroprevalence among injection drug users in Georgia. Drug and Alcohol Dependence, 82, S35–S38. 10.1016/S0376-8716(06)80006-2 [DOI] [PubMed] [Google Scholar]
  52. Spitzer RL, Kroenke K, Williams JBW, & Group PHQPCS (1999). Validation and Utility of a Self-report Version of PRIME-MD: The PHQ Primary Care Study. JAMA, 282(18), 1737–1744. 10.1001/jama.282.18.1737 [DOI] [PubMed] [Google Scholar]
  53. Spitzer RL, Kroenke K, Williams JBW, & Löwe B (2006). A Brief Measure for Assessing Generalized Anxiety Disorder: The GAD-7. Archives of Internal Medicine, 166(10), 1092–1097. 10.1001/archinte.166.10.1092 [DOI] [PubMed] [Google Scholar]
  54. Substance Abuse and Mental Health Services Administration [SAMHSA]. (2012). SAMHSA's working definition of recovery. Report PEP12-RECDEF.
  55. Sung H-E, Mahoney AM, & Mellow J (2011). Substance Abuse Treatment Gap Among Adult Parolees: Prevalence, Correlates, and Barriers. Criminal Justice Review, 36(1), 40–57. 10.1177/0734016810389808 [DOI] [Google Scholar]
  56. Vescio MF, Longo B, Babudieri S, Starnini G, Carbonara S, Rezza G, & Monarca R (2008). Correlates of hepatitis C virus seropositivity in prison inmates: A meta-analysis. Journal of Epidemiology and Community Health; London, 62(4), 305. 10.1136/jech.2006.051599 [DOI] [PubMed] [Google Scholar]
  57. Western B (2006). Punishment and Inequality in America. Russell: Sage Foundation. [Google Scholar]
  58. White W & Ali S (2010). Lapse and relapse: Is it time for a new language. Posted at www.facesandvoicesofrecovery.org. Available at: http://www.williamwhitepapers.com/pr/2010%20Rethinking%20the%20Relapse%20Language.pdf
  59. White WL (2007). Addiction recovery: Its definition and conceptual boundaries. Journal of Substance Abuse Treatment, 33(3), 229–241. 10.1016/j.jsat.2007.04.015 [DOI] [PubMed] [Google Scholar]
  60. Willner-Reid J, Belendiuk KA, Epstein DH, Schmittner J, & Preston KL (2008). Hepatitis C and human immunodeficiency virus risk behaviors in polydrug users on methadone maintenance. Journal of Substance Abuse Treatment, 35(1), 78–86. 10.1016/j.jsat.2007.08.011 [DOI] [PMC free article] [PubMed] [Google Scholar]
  61. Winter RJ, Stoové M, Degenhardt L, Hellard ME, Spelman T, Jenkinson R, McCarthy DR, & Kinner SA (2015). Incidence and predictors of non-fatal drug overdose after release from prison among people who inject drugs in Queensland, Australia. Drug and Alcohol Dependence, 153, 43–49. 10.1016/j.drugalcdep.2015.06.011 [DOI] [PubMed] [Google Scholar]
  62. Witkiewitz K, & Marlatt GA (2004). Relapse Prevention for Alcohol and Drug Problems: That Was Zen, This Is Tao. American Psychologist, 59(4), 224–235. 10.1037/0003-066X.59.4.224 [DOI] [PubMed] [Google Scholar]
  63. Witkiewitz K, & Marlatt GA (2007). Modeling the complexity of post-treatment drinking: It’s a rocky road to relapse. Clinical Psychology Review, 27(6), 724–738. 10.1016/j.cpr.2007.01.002 [DOI] [PMC free article] [PubMed] [Google Scholar]
  64. Witkiewitz K, & Masyn KE (2008). Drinking trajectories following an initial lapse. Psychology of Addictive Behaviors, 22(2), 157. [DOI] [PMC free article] [PubMed] [Google Scholar]
  65. Yokell MA, Zaller ND, Green TC, & Rich JD (2011). Buprenorphine and Buprenorphine/Naloxone Diversion, Misuse, and Illicit Use: An International Review. Current Drug Abuse Reviews, 4(1), 28–41. [DOI] [PMC free article] [PubMed] [Google Scholar]
  66. Young AM, Havens JR, & Leukefeld CG (2010). Route of administration for illicit prescription opioids: A comparison of rural and urban drug users. Harm Reduction Journal, 7(1), 24. 10.1186/1477-7517-7-24 [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES