Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2019 Jul 1.
Published in final edited form as: Drug Alcohol Depend. 2018 May 12;188:232–238. doi: 10.1016/j.drugalcdep.2018.04.010

Initiating buprenorphine treatment prior to versus after release from prison: arrest outcomes

Michael S Gordon 1, Thomas R Blue 1, Kathryn Couvillion 1, Robert P Schwartz 1, Kevin E O’Grady 2, Terrence T Fitzgerald 3, Frank J Vocci 1
PMCID: PMC5999574  NIHMSID: NIHMS967491  PMID: 29783095

Abstract

Background

This secondary analysis of a randomized trial examines the association between initiation of buprenorphine treatment prior to, versus post-release, and rearrests during the 12-months following release.

Methods

Official rearrest data (N=199) for the 12-months post-release were examined. Four outcomes were measured: (1) rearrested (yes/no), (2) time to rearrest, (3) number of rearrests, and (4) severity of charges (less severe vs. severe).

Results

A minority (43.1%) of the sample were rearrested (N=91). There were no significant differences between study conditions in the proportion of rearrested participants [P = 0.28] nor in the mean number of arrests [P = 0.15]. Likewise, the condition was not a significant predictor of the hazard of rearrest [p = 0.10]. The mean number of days until rearrest for the in prison vs. post-release buprenorphine conditions were not significantly different (205.8 days (SD = 104.6) vs. 170.8 days (SD = 113.1), respectively; P = 0.13]. Treatment condition was not a significant predictor of the likelihood of rearrest for a severe crime compared to a less severe crime [P = 0.09].

Conclusion

Despite the parent study finding of higher rates of post-release drug treatment entry in the group assigned to start buprenorphine treatment prior to, compared to post-release, there were no significant differences in the proportion of individuals arrested, the mean number of arrests, the time to first arrest, or the severity of their charges.

Keywords: Buprenorphine Treatment, Prisoners, Opioid Use Disorder, Official Arrest Records, Criminal Activity

1. Introduction

Relapse to opioid use after release from prison is a serious problem among prisoners with opioid use disorder (OUD) throughout the world. Prisoners have a disproportionately higher rate of OUD than the general populations (Dolan et al., 2007; Fazel et al., 2006; Kanato, 2008; Kastelic et al., 2008; Kinlock et al., 2011). The United States (US) leads the world in the number of prisoners (2.17 million) and the rate of incarceration (670 per 100,000 population; (Kaeble et al., 2016)). It has been estimated that about 15% of US prisoners have histories of opioid addiction (Mumola and Karberg, 2006). Thus, there is a considerable need to deliver effective treatment for this population.

The three FDA-approved pharmacotherapies (e.g., methadone, buprenorphine, and naltrexone) used to treat OUD are infrequently used prior to release (Fox, 2015). Given their demonstrated efficacy in community settings (Lee et al., 2016; Mattick et al., 2014), this is a lost opportunity to prevent illicit opioid use in prison and relapse upon release for individuals who have maintained opioid abstinence during incarceration. Newly-released inmates have a high risk of opioid overdose death within their first month in the community (Binswanger et al., 2011; Binswanger et al., 2012; Binswanger et al., 2007; Bird and Hutchinson, 2003; Farrell and Marsden, 2008; Krinsky et al., 2009; Merrall et al., 2010;Marsden et al., 2017; Stewart et al., 2004) and are at risk for HIV and hepatitis B and C infections (Dolan et al., 2007; Inciardi, 2008; Kanato, 2008). Opioid use among newly-released inmates also has adverse public safety consequences, as it is associated with criminal activity (Hough, 2002; Inciardi, 2008; Kinlock et al., 2003) and re-incarceration (Dolan et al., 2005; Hough, 2002; Metz et al., 2010). Given the focus of correctional settings on reducing recidivism, research is needed on approaches, including pharmacotherapy, to reduce both the criminal behavior and the rearrest following release from prison.

Adults with opioid use disorder frequently engage in illegal activities, which are positively correlated with the frequency of illicit opioid use (Anglin and Speckart, 1988; Inciardi, 2008). These activities are most frequently related to possession and sale of illicit opioids and acquisitive crimes such as shoplifting and burglary to obtain funds for drugs (Ball and Ross, 2012). While self-reported criminal behavior is often considered accurate when reported under research confidential conditions (Chaiken and Chaiken, 1990; Nurco, 1985), it is subject to underreporting. An alternative to self-report measures of criminal behavior is objective data from official arrest records. These data are not subject to underreporting, although they underestimate the number of crimes committed (Nurco, 1998) and are subject to fluctuations in local police practices (Schwartz et al., 2017). Nevertheless, capturing arrest data is of some importance given the impact of the number and types of arrests on the police, the judiciary, the correctional system, and their associated costs to society. Arrest data are also important indicators of the success of correctional systems generally, and of treatment approaches for prisoners with opioid addictions in particular.

The published random assignment studies that have examined the association between initiating pharmacotherapy for opioid use disorder prior to versus after release from incarceration found no differences in self-reported arrests at post-release follow-up of varying durations (Kinlock et al., 2009; Lee et al., 2015; Magura et al., 2009; McKenzie et al., 2012).Two such trials found no difference in self-reported arrest rates between conditions in which participants were assigned to initiate methadone either during or after release from prison over a 6-month follow-up (McKenzie et al. 2012) and 12-month follow-up period (Kinlock et al., 2009). Magura and colleagues (2009) found no difference in self-reported arrest rates at 3-month follow-up between conditions in which participants were assigned to initiate either methadone or buprenorphine prior to release. Finally, one pilot study, which was able to examine the New York City jail admission database for the arrest status of participants lost to follow-up, found no difference in arrest rates within an 8-week follow-up between conditions in which participants were provided either extended-release naltrexone or no medication prior to release (Lee et al., 2015).

Our research group previously reported on a randomized clinical trial comparing post-release outcomes of prisoners with a pre-incarceration history of DSM-IV defined opioid dependence who were offered 12 weeks of counseling in prison and were randomly assigned to begin sublingual buprenorphine/naloxone prior to vs. post-release (Gordon et al., 2014; Gordon et al., 2017). This study found that participants randomly assigned to initiate buprenorphine in prison were significantly more likely to enter and remain in buprenorphine treatment post-release compared to those participants who were assigned to begin buprenorphine in the community (Gordon et al., 2017). However, notwithstanding greater community treatment exposure, there were no significant differences in heroin and cocaine use during the 12-month follow-up (Gordon et al., 2017).

1.1 The Present Study

Here we present findings from a secondary analysis from the above-mentioned clinical trial comparing the impact on official rearrest data during the 12-month post-prison release follow-up period of being assigned to initiate buprenorphine treatment prior to vs. post-release. We hypothesized that because of the advantages conferred by initiating buprenorphine treatment in prison (higher rates of treatment entry and retention in the community), participants who initiated buprenorphine in prison compared to post-release would, over the 12-month follow-up period: (1) be less likely to be rearrested in the 12-month post-release period, (2) have longer time to first rearrest, (3) be rearrested fewer times, and (4) be less likely to be rearrested for a severe crime (e.g., burglary, assault and battery, robbery, kidnapping, murder, rape)

2. Methods

2.1 Parent Study

Parent study participants were 211 adult prisoners with a history of DSM-IV defined opioid dependence in the year prior to their index incarceration. They were recruited between 2008 and 2012 within 3 to 9 months prior to release. All participants were offered 12 weekly group counseling sessions during prison and were randomly assigned to receive sublingual buprenorphine/naloxone (bup/nal), hereafter called buprenorphine, in prison or post-release and randomly assigned to receive buprenorphine treatment post-release in either an opioid treatment program (OTP) or non-OTP outpatient drug treatment program. Because most of the participants were not opioid-tolerant at the time of study recruitment, bup/nal was started at low doses (1/0.25 mg daily) and increased much more slowly (e.g., increase of 1/0.25 mg per week until reaching 4 mg/1mg with subsequent increases by 2 mg/0.5 mg per week to reach 8 mg/2mg) than would be used for opioid-tolerant patients in the community (see (Vocci et al., 2015) for details). A detailed description of methods and outcomes in the parent study can be found elsewhere (Gordon et al., 2014; Gordon et al., 2017). The study was approved by the Friends Research Institute’s Institutional Review Board and the Maryland Department of Public Safety and Correctional Services (DPSCS) Research Committee, and the Federal Office of Human Research Protections. The ClinicalTrials.gov identifier is NCT 00574067.

2.2. Arrest Data

As part of the study protocol, Maryland’s DPSCS provided the researchers the official arrest data for study participants for the first 12-months after release from each individual’s index incarceration. Data included dates of arrest and specific charges. Participants’ data were extracted for the 12-month period following their respective release date and were matched using their: (1) names, (2) social security numbers, (3) dates of birth, (4) DPSCS identification number, (5) gender; (6) and race. DPSCS records could not be located for 12 participants; consequently, they were excluded from the present study.

Individuals with at least one arrest within the post-release period were considered to have been rearrested. The number of times each was arrested was also recorded. As shown in Table 1, arrest data were coded for severity on a scale from lowest (1) to highest (7) (Schwartz et al., 2009) that was adapted from a crime severity rating scale developed by Nurco, Hanlon, Balter, Kinlock, and Slaght (Nurco et al., 1991). Offenses involving the use of violence (e.g., homicide, rape, assault) or that could quickly lead to violence (burglary, robbery) were considered the most severe followed by those involving a loss or destruction of property (e.g., theft, vandalism), and offenses for which there is no immediate victim (e.g., drug possession, prostitution). Previous research has found that rating crime severity in this manner produces a high degree of interrater agreement as measured by an intraclass correlation (ICC) (Shrout and Fleiss, 1979), ICC(2,k) = 0.97 (Schwartz et al., 2009). For individuals with multiple offenses in a single arrest, or multiple offenses resulting from more than one arrest, the rating of the most severe offense was used. In addition to the crime scale (1–7), offense ratings were dichotomized into classifications of severe and less severe (See Schwartz et al. 2009) for comparison between treatment conditions and participant genders. Arrests resulting from crimes rated 5 or higher (involving violence or the possibility of quickly leading to violence) were considered severe while arrests resulting from crimes rated 4 or lower (non-violent offenses and property crimes) were considered less severe.

Table 1.

Crime Severity Scale Scoring Examples

Severity Rating Examples
1 Prostitution, possession of a controlled, dangerous substance, violation of probation/parole
2 Violation of a protection order, trespassing, disorderly conduct, malicious destruction of property less than $500
3 Theft less than $500, malicious destruction of property greater than $500, forgery, uttering
4 Theft greater than $500, possession of a deadly weapon, drug distribution or manufacturing
5 Burglary, second degree assault, battery
6 Robbery, first degree assault, kidnapping
7 Attempted first degree murder, first degree murder, rape
Open in a new tab

Adapted from Schwartz et al. (2009). Scores greater than 4 are considered severe crimes. Scores less than or equal to 4 are considered less severe crimes.

2.3. Statistical Analysis

2.3.1. Predictor variables

The predictor variables can be considered either treatment or control variables. The treatment variable was assigned treatment condition: buprenorphine initiated in prison vs. buprenorphine initiated in the community. The control variable was participant gender.

2.3.2. Outcome variables

There were four outcome measures: (1) arrested (yes vs. no), (2) time to first arrest (days), (3) number of arrests, and (4) crime severity of arrests (less severe vs. severe). The outcomes covered the 12-month post-release period.

2.3.3. Statistical approach

To evaluate the relationship of the predictor variables and the likelihood of rearrest at any given time during the 12-month post release period, a Cox Proportional Hazards (PH) regression was used. A Univariate Analysis of Variance (ANOVA) was conducted to test for differences between treatment conditions and genders in the mean time to the first arrest and for a potential interaction effect of treatment condition and gender on mean time to the first rearrest. Negative binomial regression was used to test the relationship of the predictor variables with the total number of arrests, which were count data. Lastly, crime severity was evaluated by a multiple linear regression examining the relationship of the predictor variables with the severity rating of the most severe criminal charge for which each participant was arrested.

3. Results

3.1. Participants

Participants in the parent study were 211 prisoners; however, this secondary analysis reports on 199 (12 participants had no DPSCS records and excluded from the analysis) of whom 100 were randomly assigned to begin buprenorphine treatment in prison and 99 to begin buprenorphine post-release. As shown in Table 2, there were 141 (70.1%) males and 58 (29.9%) females. Their mean age was 39.4 years old (SD = 8.5), 139 (69.8%) were African American, 51 (25.6%) were Caucasian, and 9 (4.5 %) identified as some other racial group. The average age of first use for heroin was 19.3 years old (SD = 6.0), and the average age of first crime was 13.7 years old (SD = 6.3). On average, participants reported being incarcerated 7.4 times (SD=7.2) prior to their index incarceration. During the past 30 days prior to their index incarceration, participants reported using heroin on 24.3 days (SD=10.2) and committing a crime on 20.7 days (SD = 13.7) on average. Of the characteristics described above and shown in Table 2, there were no significant differences between treatment conditions. A summary of the results of all major statistical tests is presented in Table 3.

Table 2.

Sample Characteristics (N=199)

Buprenorphine Initiated in community(n = 99) Buprenorphine Initiated in prison(n = 100) Full Sample (N = 199)
Categorical Variables n (%)
Gender
 Male 72 (72.7) 69 (69.0) 141 (70.9)
 Female 27 (27.3) 31 (31.0) 58 (29.1)
Race
 African American 72 (72.7) 67 (67.0) 139 (69.8)
 Caucasian 22 (22.2) 29 (29.0) 51 (25.6)
 Other 5 (5.0) 4 (4.0) 9 (4.5)
Prior drug treatment 83 (83.8) 81 (81.0) 164 (82.4)
Prior buprenorphine treatment 17 (17.2) 14 (14.0) 31 (15.6)
Continuous Variables M (SD)
Age 39.54 (8.9) 39.19 (8.2) 39.36 (8.5)
Age of first heroin use 19.46 (5.9) 19.10 (6.1) 19.28 (6.0)
Age of first crime 13.14 (4.7) 14.27 (7.6) 13.71 (6.3)
Number of prior incarcerations 6.41 (3.5) 8.34 (9.5) 7.38 (7.2)
Heroin use days1 24.38 (10.5) 24.28 (10.0) 24.33 (10.2)
Crime days1 22.45 (14.4) 18.97 (12.8) 20.71 (13.7)
Current incarceration (days) 655.89 (1125.4) 502.53 (815.8) 578.82 (982.6)
Open in a new tab
1

For the last 30 days prior to current incarceration.

Table 3.

Summary of findings for planned statistical tests

Outcome Sample and sample size Statistical Method Results for Test of Condition Effect Results of Test for Gender Differences
Parameter Estimate P Parameter Estimate P
Likelihood of rearrest Full Sample (N = 199) Cox PH Regression b = 0.34 SE = 0.21 0.10 b = 0.77
SE = 0.27		 0.005
Mean time to first rearrest Rearrested at least once (n = 91) 2 x 2 Factorial ANOVA F(1,87) = 0.01
ΔEMM = 3.63
ΔSE = 30.44		 0.91 F(1,87) = 0.13
ΔEMM = 11.16
ΔSE = 30.44		 0.72
Total number of arrests Full Sample (N = 199) Negative Binomial Regression b = 0.25 SE = 0.20 0.19 b = 0.77
SE = 0.25		 0.002
Severity rating of most severe criminal charge Rearrested at least once (n = 91) Multiple Linear Regression b = 0.46 SE = 0.35 0.19 b = 1.09
SE = 0.45		 0.02
Open in a new tab

Note. ΔEMM represents the differences in estimated marginal means and ΔSE represents the corresponding standard error.

3.2. Arrested (yes vs. no)

Although most individuals were not arrested, 43.1% (n = 91) were rearrested at least once during the 12-month post-release period. The proportion of arrested individuals who had initiated buprenorphine in prison (n = 50, 50.0%) did not differ significantly from that proportion of participants who had initiated buprenorphine in the community (n = 41, 41.4%; χ2(1) = 1.15, p = 0.28). However, there was a significantly smaller proportion of female participants (n = 17, 29.3%) who were arrested than male participants (n = 74, 52.5%; χ2(1) = 7.98, p = 0.004). The proportion of participants arrested at least once throughout the 12-month post-release period are shown in Figure 1.

Figure 1.

Figure 1

Open in a new tab

Time to first rearrest. The proportion of participants rearrested at least once throughout 12-month post-release period.

There was no significant effect of treatment condition on hazard of arrest [b = 0.34, SE = 0.21, hazards ratio (HR) = 1.41, 95% Confidence Interval (95% CI): (0.93, 2.13), p = 0.10]. Moreover, male participants were more than twice as likely to be arrested as female participants at any given time [b = 0.77, SE = 0.27, HR = 2.16, 95% CI: (1.27, 3.66), p = 0.005]. The survival function given by the Cox PH regression at each level of participant gender is presented in Figure 2.

Figure 2.

Figure 2

Open in a new tab

Survival function by participant gender (controlling for treatment condition). The probability of survival throughout the 12-month post-release period given by the survival function of the Cox PH regression at each level of participant gender when controlling for treatment condition.

3.3. Time to First Arrest

The mean number of days until the first arrest for the entire sample was 186.6 (SD=110.1, Mdn = 190). The range of time to arrest was from 0 to 363 days. The mean number of days until the first arrest was 205.8 days (SD = 104.6) and 170.82 days (SD = 113.1) for participants who initiated buprenorphine in prison vs. post-release, respectively. Mean time to arrest for male participants was 187.2 days (SD = 113.3) compared to 183.9 days (SD = 98.3) for female participants. A 2 x 2 (treatment condition x gender) factorial ANOVA was conducted to test the relationships of treatment condition and gender with the time to first arrest. The results revealed that there were no significant relationships between condition [F(1,87) = 0.01, p = 0.91] or gender [F(1,87) = 0.13, p = 0.72] and the time to first arrest; nor was there a significant interaction effect between the two factors [F(1,87) = 2.50, p = 0.12].

3.4. Number of Arrests

Arrest count data are summarized in Table 4. The number of arrests is a more sensitive measure of criminal activity than a categorical measure of arrest (Yes/No) because a single individual may have been arrested multiple times during the 12-month post-release period. Among the 91 individuals rearrested at least once, there were a total of 154 arrests. On average, participants were charged with 2.5 (SD = 2.2; Min = 1; Max = 12) crimes per arrest. Over half of all arrests were related to drug possession (n = 89, 57.8%). Property crimes such as theft, destruction of property, and fraud were related to the second largest proportion of arrests at 24.7% (n = 38) followed by arrests related to drug distribution and manufacturing (n = 25, 16.2%), crimes of force or violence such as assault and robbery (n = 18, 11.7%), burglary (n = 10, 6.5%), prostitution and indecent acts (n = 8, 5.2%), and possession of deadly weapons (n = 6, 3.9%). The mean number of charges per arrest did not significantly differ between treatment conditions [t(89) = −0.35, P = 0.73] or participant genders [t(89) = −1.00, P = 0.32].

Table 4.

Summary of Arrest Data (N=199)

Rearrested at Least Once (N = 91) Buprenorphine initiated in Community (n = 41) Buprenorphine Initiated in Prison (n = 50) Female Participants (n = 17) Male Participants (n = 74)
Arrest Counts, Frequencies n (%)
1 52 (57.1) 22 (53.7) 30 (60.0) 12 (70.6) 40 (54.1)
2 23 (25.3) 13 (31.7) 10 (20.0) 2 (11.8) 21 (28.4)
3 9 (11.0) 4 (9.8) 5 (10.0) 3 (17.6) 6 (8.1)
4 5 (5.5) 2 (4.9) 4 (8.0) 0 (0) 6 (8.1)
5 1 (1.1) 0 (0) 1 (2.0) 0 (0) 1 (1.4)
Crime Severity Ratings, Frequencies n (%)
1 23 (25.3) 12 (29.3) 11 (22.0) 7 (41.2) 16 (21.6)
2 6 (6.6) 2 (4.9) 4 (8.0) 1 (5.9) 5 (6.8)
3 5 (5.5) 2 (4.9) 3 (6.0) 3 (17.6) 2 (2.7)
4 32 (35.2 17 (41.5) 15 (30.0) 4 (23.5) 28 (37.8)
5 16 (17.6) 6 (14.6) 10 (20.0) 2 (11.8) 14 (18.9)
6 9 (9.9) 2 (4.9) 7 (14.0) 0 (0) 9 (12.2)
7 0 (0) 0 (0) 0 (0) 0 (0) 0 (0)
Open in a new tab

Note. Data for participants rearrested at least once. Counts only the most severe rating for each participant.

The mean number of arrests per person was 0.8 (SD= 1.1). Among participants who were arrested at least once, the mean number of arrests was 1.7 (SD = 1.0). The mode of arrests (for those arrested at least once) was 1 (n = 52), and the most times an individual was arrested was 5 (n = 1).

The average number of arrests for participants who initiated buprenorphine in prison versus post-release were 0.9 (SD = 1.1, Mdn = 0.50) and 0.7 (SD = 0.99, Mdn = 0), respectively. Male participants were arrested 0.9 times on average (SD = 1.1, Mdn = 1) while female participants were only arrested 0.4 times on average (SD = 0.8, Mdn = 0).

To test the relationships of treatment condition and gender with arrest count, a negative binomial regression analysis was used. Normally, Poisson regression analyses are used to model count data; however, inspection of the arrest count data suggested that there may be significant levels of over-dispersion [index of dispersion, D = 1.5, χ2(1) = 293.1, P < 0.0001; (Frome, 1982)]. Negative binomial models can be used to analyze over-dispersed data by estimating a dispersion parameter that is held constant by Poisson models. Treatment condition and participant gender were entered as predictors to the negative binomial model and a Poisson model for comparison. Results of a likelihood ratio test confirmed that the use of the negative binomial model was more appropriate than a Poisson model for the current dataset [χ2(1) = 11.60, P = 0.001]. The results of the negative binomial model revealed that there was no significant effect of treatment condition on the number of arrests [b = 0.25, SE = 0.20, Rate Ratio (RR) = 1.29, 95% CI: (0.88, 1.89), P = 0.19]. However, participant gender was related to the number of arrests when controlling for treatment condition [b = 0.77, SE = 0.25, RR = 2.15, 95% CI: (1.32, 3.50), P = 0.002] such that male participants were expected to be arrested 2.15 times often as female participants.

Although there was no significant effect of treatment condition on the number of arrests, there were differences between conditions in the proportion of arrests involving various crimes. The results of a z-test for independent proportions revealed that there was a significantly smaller proportion of arrests involving drug possession for the group that initiated buprenorphine in prison (50.0%) compared with the group that initiated buprenorphine in the community (67.6%; z = −2.20, P = 0.03). However, the proportion of arrests involving burglary was significantly larger for the buprenorphine in prison condition (10.5%) than it was for the buprenorphine in the community condition (1.5%; z = 2.25, P = 0.02). The proportions of arrests involving other crimes were not significantly different between treatment conditions.

In addition to the significant relationship between participant gender and the number of arrests, there were significant gender differences in the proportion of arrests related to different crimes. The number of arrests for crimes involving drug possession was significantly larger for male participants (62.8%) than it was for female participants (32.0%; z = −2.85, P = 0.004). However, the number of arrests for crimes involving prostitution and indecent exposure was significantly larger for female participants (32.0%) than it was for male participants (0%; z = 6.60, P < 0.0001). The proportions of arrests involving other crimes were not significantly different between genders.

3.5. Crime Severity

The crime severity ratings data are summarized in Table 4. Of the 91 participants arrested within the 12-month post-release period, 27.5% (n = 25) were arrested for a severe crime at least once. Of participants who initiated buprenorphine in prison and were arrested at least once, 34.0% (n = 17) were arrested for at least one severe crime; 19.5% (n = 8) of participants who initiated buprenorphine in the community and were arrested at least once were arrested for at least one severe crime. Of female participants arrested at least once, 11.8% (n = 2) were arrested for a severe crime at least once. Of male participants arrested at least once, 31.1% (n = 23) were arrested for at least one severe crime. A multiple linear regression was performed to test the relationships of treatment condition and gender with the severity rating of the most severe criminal charge for which each participant was arrested. The results revealed that although there was no significant effect of treatment condition on severity ratings [b = 0.46, SE = 0.35, 95% CI: (−0.22, 1.14), t(88) = 1.32, P = 0.19], there was a significant relationship between gender and severity ratings [b = 1.09, SE = 0.45, 95% CI: (0.22, 1.96), P = 0.02] such that the mean severity rating of males was higher than that of females by 1.1 units when controlling for treatment condition.

4. Discussion

This is a secondary analysis of official arrest data from our previously reported randomized clinical trial conducted with non-opioid-tolerant prisoners with a history of opioid use disorder who were randomly assigned to begin buprenorphine in prison prior to release compared to post-release (Gordon et al., 2014; 2017). Those participants assigned to buprenorphine treatment in prison were more likely to enter and remain in buprenorphine treatment in the community (Gordon et al., 2014). Notwithstanding longer treatment exposure to community treatment of those who were assigned to start buprenorphine prior to release, there were no significant differences in illicit opioid or cocaine use as measured by self-report and urine testing or in self-reported criminal activity between study conditions (Gordon et al., 2017). In this secondary analysis, no significant differences were found between conditions in the likelihood of being arrested; the time to first arrest; the number of arrests; and the likelihood of being arrested for a severe crime as measured by official arrest records during the 12-month post-release follow-up. One possible explanation for the failure to find a difference between rearrests after release from prison is that neither condition had much treatment exposure (mean of 65.9 days vs. 21.8 days for prior to be compared to after release, respectively) during the one-year follow-up (Gordon et al., 2017). Such limited treatment exposure in the present study may have been insufficient to positively impact criminal behavior and arrest outcomes over a 12-month period.

There are multiple factors that likely impact any single arrest. It has been noted that arrest underrepresents the amount of criminal behavior (Nurco, 1998). Policing policy and practices have an important influence on arrest rates, thus exposing research outcomes to secular trends (Schwartz et al., 2017). The population under study can also have an impact on the likelihood of arrest. In prison populations, there is a strong likelihood of rearrest regardless of their opioid addiction status. One Bureau of Justice Statistics report found that within 3 years of release, 67.8% of prisoners were rearrested and within 5 years 76.6% were re-arrested (Durose et al., 2014). Furthermore, nearly three-quarters of prisoners with opioid use disorder relapse within three-month of release (Kinlock et al. 2008). Many of these former inmates return to environments that trigger relapse to drug use (Binswanger et al. 2012). Moreover, regardless of intervention in prison these individuals mostly return to the same environment. Thus, examining the association between arrest and criminogenic and environmental issues is a complicated area that would benefit from additional, carefully conducted research.

We are not aware of any other research from random assignment studies in the US that examined the impact on official arrest data, of in-prison vs. post-release initiation of buprenorphine, methadone, or naltrexone. Indeed, there is a paucity of research reports using official arrest data in community trials as well. In one such study in a randomized trial of interim methadone (methadone without counseling) compared to waiting list (N=319), Schwartz, Jaffe, O’Grady et al., 2009 found a lower percentage of participants who were arrested in the 12-months post-study enrollment compared to the present study (26.5% versus 43.1%). In that study, there was a significant difference in the mean number of arrests in the 6- but not 12-months following study enrollment, favoring the interim methadone condition. Clearly, more work is needed using official arrest data in trials in both the community and in correctional settings.

It is worthy of note that there were significant gender differences in the present study. A significantly higher percentage of men compared to women were arrested, and women were more likely to be arrested for prostitution while men were likely to be arrested for drug possession. Finally, men were more likely to be arrested for a more serious crime. This is consistent with the literature as there is some variation in the frequency, type, and severity of the crime committed by gender among persons with opioid use disorders (Hser et al., 2003; Inciardi, 2008). Males typically have earlier onsets of criminal activity and greater involvement in more serious offenses (Inciardi, 2008; Gordon et al., 2013; Kinlock et al., 2003), whereas females are more likely to commit prostitution and nonviolent crimes at high rates. Moreover, men with heroin use disorders typically have more legal problems (Pelissier and Jones, 2005) and greater involvement with the criminal justice system (Inciardi, 2008).

The present study adds to the extant self-report research findings that indicate that starting methadone or buprenorphine prior to release rather than after release from prison does not have a significant impact on the likelihood of arrest. The present study’s strengths including random assignment, relatively large sample size, and analysis of official arrest records in support this finding. There are no such comparative data with extended-release naltrexone. It also appears that neither starting methadone (Kinlock et al., 2012) nor extended-release naltrexone (Lee et al., 2015) prior to release is superior to no medication in terms of reducing arrest, although the naltrexone data are from a small pilot study.

The present study has a few important limitations that should be mentioned. First, we were missing official arrest data for 12 participants. Second, the results of the negative binomial regression examining number of arrests may have been biased by the fact that individuals who are arrested inherently have fewer opportunities for arrest during the remainder of the 12-month follow-up period due to spending time detained or in prison (if found guilty). Third, the data were gathered from Baltimore city, which may limit the generalizability of the results to other cities within the U.S. and other countries.

5. Directions for Future Research

Overall, in spite of the failure of the current study to support the stated hypotheses, the initiation of buprenorphine treatment in prison is a promising intervention due to the higher rates of future buprenorphine treatment entry in both prison and in the community post-release and therefore needs further examination. Future research should address the limitations of the present study by using larger sample and understanding other criminogenic risk factors (e.g., anti-social values, low self-control, dysfunctional family ties, substance abuse, criminal personality, criminal peers, housing, employment, education) that may impact on arrest and criminal activity. This can be obtained by replicating the study in another city with a larger sample size and examining other potential moderating variables. In addition, future studies might examine longer duration dosage forms of buprenorphine (injectables, implant) and long-acting naltrexone which might make a difference in treatment retention and arrests.

Highlights.

  • Forty-three percentof participantswere rearrested within 12 -month of release.

  • Buprenorphine in prison vs.after prison was not associated with fewer rearrests.

  • Condition was not a significant predictor of the likelihood of rearrest.

Acknowledgments

We would like to thank the Maryland Department of Public Safety and Correctional Services (DPSCS) and specifically Mrs. Sandra Davis-Hart, Director of Treatment Services for DPSCS. We also wish to thank all of the participants in this study.

Role of Funding Source

This study was supported by the National Institute on Drug Abuse (NIDA), Buprenorphine for Prisoners (PI: Kinlock; R01DA021579). NIDA staff had no further role in study design; in the collection, analysis, and interpretation of data; in the writing of this manuscript and the decision to submit it for publication.

Footnotes

Contributors

All authors contributed to and approved the final manuscript. The manuscript is the work of the authors alone. Authors Gordon, Schwartz, and O’Grady designed the study and wrote the protocol. Authors Gordon and Blue wrote the first draft of the manuscript. Authors Blue and Gordon wrote the first draft of the tables and figures. Authors Blue and O’Grady made the primary contribution to the statistical analyses.

Conflict of Interest

This study was supported by an unrestricted, unsolicited investigator-initiated request from Reckitt Benckiser Pharmaceuticals, Inc. (provided study drug only) who had no role in study design; collection, analysis and interpretation of data; in the writing of the manuscript; or in the decision to submit the manuscript for publication. The authors alone are responsible for the content and writing of this manuscript. Drs. Gordon and Fitzgerald received investigator-initiated funding and study drug from Alkermes. Dr. Blue reports no conflicts of interest. Dr. Schwartz did a one-time consultation for Reckitt-Benckiser on behalf of his employer (the Friends Research Institute). Dr. O’Grady has in the past received funding for his time from Reckitt-Benckiser. Dr. Vocci reports personal fees and other from Braeburn Pharmaceuticals; personal fees and other from Pinney Associates; personal fees and other from Indivior, personal fees and other from Demerx, personal fees from Alkermes, personal fees and other from Insys Pharmaceuticals, and stock ownership from Intratab Labs, Inc.

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 citable 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.

References

  1. Anglin MD, Speckart G. Narcotics use and crime: A multisample, multimethod analysis. Criminol. 1988;26:197–233. [Google Scholar]
  2. Ball JC, Ross A. The effectiveness of methadone maintenance treatment: Patients, programs, services, and outcome. SSBM; 2012. [Google Scholar]
  3. Binswanger IA, Blatchford PJ, Lindsay RG, Stern MF. Risk factors for all-cause, overdose and early deaths after release from prison in Washington state. Drug Alcohol Depend. 2011;117:1–6. doi: 10.1016/j.drugalcdep.2010.11.029. [DOI] [PubMed] [Google Scholar]
  4. Binswanger IA, Nowels C, Corsi KF, Glanz J, Long J, Booth RE, Steiner JF. Return to drug use and overdose after release from prison: A qualitative study of risk and protective factors. Addict Sci Clin Pract. 2012;7:3. doi: 10.1186/1940-0640-7-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Binswanger IA, Stern MF, Deyo RA, Heagerty PJ, Cheadle A, Elmore JG, Koepsell TD. Release from prison--A high risk of death for former inmates. N Engl J Med. 2007;356:157–165. doi: 10.1056/NEJMsa064115. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Bird SM, Hutchinson SJ. Male drugs-related deaths in the fortnight after release from prison: Scotland, 1996–99. Addiction. 2003;98:185–190. doi: 10.1046/j.1360-0443.2003.00264.x. [DOI] [PubMed] [Google Scholar]
  7. Chaiken JM, Chaiken MR. Drugs and predatory crime. In: Tonry M, Wilson JQ, editors. Drugs and Crime [Crime and Justice: A Review of Research] University of Chicago Press; Chicago: 1990. pp. 203–239. [Google Scholar]
  8. Dolan K, Khoei EM, Brentari C, Stevens A. Prisons and drugs: A global review of incarceration, drug use and drug services. Beckley Foundation; Oxford: 2007. [Google Scholar]
  9. Dolan KA, Shearer J, White B, Zhou J, Kaldor J, Wodak AD. Four-year follow-up of imprisoned male heroin users and methadone treatment: mortality, re-incarceration and hepatitis C infection. Addiction. 2005;100:820–828. doi: 10.1111/j.1360-0443.2005.01050.x. [DOI] [PubMed] [Google Scholar]
  10. Durose MR, Cooper AD, Snyder HN. Recidivism of Prisoners Released in 30 States in 2005: Patterns from 2005 to 2010. Washington, DC: US Department of Justice, Office of Justice Programs, Bureau of Justice Statistics; 2014. Retrieved from: https://www.bjs.gov/content/pub/pdf/rprts05p0510.pdf. [Google Scholar]
  11. Farrell M, Marsden J. Acute risk of drug–related death among newly released prisoners in England and Wales. Addiction. 2008;103:251–255. doi: 10.1111/j.1360-0443.2007.02081.x. [DOI] [PubMed] [Google Scholar]
  12. Fazel S, Bains P, Doll H. Substance abuse and dependence in prisoners: A systematic review. Addiction. 2006;101:181–191. doi: 10.1111/j.1360-0443.2006.01316.x. [DOI] [PubMed] [Google Scholar]
  13. Fox AD. Opioid addiction and criminal justice systems: Opportunities to break the cycle of incarceration. SGIM Forum. 2015:38. Retrieved from: https://www.sgim.org/File%20Library/SGIM/Resource%20Library/Forum/2015/Jan2015-04.pdf.
  14. Gordon MS, Kinlock TW, Schwartz RP, Fitzgerald TT, O'Grady KE, Vocci FJ. A randomized controlled trial of prison-initiated buprenorphine: prison outcomes and community treatment entry. Drug Alcohol Depend. 2014;142:33–40. doi: 10.1016/j.drugalcdep.2014.05.011. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Gordon MS, Kinlock TW, Schwartz RP, O'Grady KE, Fitzgerald TT, Vocci FJ. A randomized clinical trial of buprenorphine for prisoners: Findings at 12-months post-release. Drug Alcohol Depend. 2017;172:34–42. doi: 10.1016/j.drugalcdep.2016.11.037. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Hough M. Drug user treatment within a criminal justice context. Subst Use Misuse. 2002;37:985–996. doi: 10.1081/ja-120004162. [DOI] [PubMed] [Google Scholar]
  17. Hser Y, Huang D, Teruya C, Douglas AM. Gender comparisons of drug abuse treatment outcomes and predictors. Drug Alcohol Depend. 2003;72:255–264. doi: 10.1016/j.drugalcdep.2003.07.005. [DOI] [PubMed] [Google Scholar]
  18. Inciardi JA. The war on drugs IV: The continuing saga of the mysteries and miseries of intoxication, addiction, crime and public policy. Allyn and Bacon; Boston, MA: 2008. [Google Scholar]
  19. Kaeble D, Glaze L, Tsoutis A, Minton T. Correctional populations in the United States, 2014. Washington, DC: US Department of Justice, Office of Justice Programs Bureau of Justice Statistics; 2016. Retrieved from https://www.bjs.gov/content/pub/pdf/cpus14.pdf. [Google Scholar]
  20. Kanato M. Drug use and health among prison inmates. Curr Opin Psychiatry. 2008;21:252–254. doi: 10.1097/YCO.0b013e3282fc985c. [DOI] [PubMed] [Google Scholar]
  21. Kastelic A, Pont J, Stöver H. A practical guide. Oldenburg: BIS-Verlag; 2008. Opioid substitution treatment in custodial settings. Retrieved from: https://www.unodc.org/documents/hiv-aids/OST_in_Custodial_Settings.pdf. [Google Scholar]
  22. Kinlock TW, Gordon MS, Schwartz RP. Incarcerated populations. In: Ruiz P, Strain E, editors. Lowinson and Ruiz’s Substance Abuse: A comprehensive textbook. Lippincott, Williams and Wilkins; Philadelphia, PA: 2011. pp. 881–891. [Google Scholar]
  23. Kinlock TW, Gordon MS, Schwartz RP, Fitzgerald TT, O'Grady KE. A randomized clinical trial of methadone maintenance for prisoners: Results at 12-months post–release. J Subst Abuse Treat. 2009;37:277–285. doi: 10.1016/j.jsat.2009.03.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Kinlock TW, Gordon MS, Schwartz RP, O’Grady KE. A study of methadone maintenance for male prisoners: 3-month post-release outcomes. Crim Justice Behav. 2008;35:34–47. doi: 10.1177/0093854807309111. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Kinlock TW, O'Grady KE, Hanlon TE. Prediction of the criminal activity of incarcerated drug-abusing offenders. J Drug Issues. 2003;33:897–920. [Google Scholar]
  26. Krinsky CS, Lathrop SL, Brown P, Nolte KB. Drugs, detention, and death: A study of the mortality of recently released prisoners. Am J Forensic Med Pathol. 2009;30:6–9. doi: 10.1097/PAF.0b013e3181873784. [DOI] [PubMed] [Google Scholar]
  27. Lee JD, Friedmann PD, Kinlock TW, Nunes EV, Boney TY, Hoskinson RA, Jr, Wilson D, McDonald R, Rotrosen J, Gourevitch MN, Gordon M, Fishman M, Chen DT, Bonnie RJ, Cornish JW, Murphy SM, O’Brien CP. Extended-release naltrexone to prevent opioid relapse in criminal justice offenders. N Eng J Med. 2016;374:1232–1242. doi: 10.1056/NEJMoa1505409. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Lee JD, McDonald R, Grossman E, McNeely J, Laska E, Rotrosen J, Gourevitch MN. Opioid treatment at release from jail using extended-release naltrexone: A pilot proof-of-concept randomized effectiveness trial. Addiction. 2015;110:1008–1014. doi: 10.1111/add.12894. [DOI] [PubMed] [Google Scholar]
  29. Magura S, Lee JD, Hershberger J, Joseph H, Marsch L, Shropshire C, Rosenblum A. Buprenorphine and methadone maintenance in jail and post-release: A randomized clinical trial. Drug Alcohol Depend. 2009;99:222–230. doi: 10.1016/j.drugalcdep.2008.08.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Marsden J, Stillwell G, Jones H, Cooper A, Eastwood B, Farrell M, Lowden T, Maddalena N, Metcalfe C, Shaw J, Hickman M. Does exposure to opioid substitution treatment in prison reduce the risk of death after release? A national prospective observational study in England. Addiction. 2017;112:1408–1418. doi: 10.1111/add.13779. [DOI] [PubMed] [Google Scholar]
  31. Mattick RP, Breen C, Kimber J, Davoli M. Buprenorphine maintenance versus placebo or methadone maintenance for opioid dependence. Cochrane Database Syst Rev. 2014:1–84. doi: 10.1002/14651858.CD002207.pub4. [DOI] [Google Scholar]
  32. McKenzie M, Zaller N, Dickman SL, Green TC, Parihk A, Friedmann PD, Rich JD. A randomized trial of methadone initiation prior to release from incarceration. J Subst Abuse. 2012;33:19–29. doi: 10.1080/08897077.2011.609446. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Merrall EL, Kariminia A, Binswanger IA, Hobbs MS, Farrell M, Marsden J, Hutchinson SJ, Bird SM. Meta-analysis of drug-related deaths soon after release from prison. Addiction. 2010;105:1545–1554. doi: 10.1111/j.1360-0443.2010.02990.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Metz V, Matzenauer C, Kammerer K, Winklbaur B, Ebner N, Radler D, Fischer G. Evaluation of opioid-dependent prisoners in oral opioid maintenance therapy. Heroin Addict Relat Clin Probl. 2010;12:5–16. [Google Scholar]
  35. Mumola CJ, Karberg JC. Drug use and dependence, State and Federal Prisoners, 2004. U.S. Department of Justice, Bureau of Justice Statistics (NCJ 213530); Washington, D.C: 2006. https://www.bjs.gov/content/pub/pdf/dudsfp04.pdf. [Google Scholar]
  36. Nurco DN. A discussion of validity, self-report methods of estimating drug use: Meeting current challenges to validity. Natl. Inst. Drug Abuse; Rockville, MD: 1985. pp. 4–11. http://archivetemp.iqscloud.net/pdf/monographs/57.pdf. [Google Scholar]
  37. Nurco DN. A long-term program of research on drug use and crime. Subst Use Misuse. 1998;33:1817–1837. doi: 10.3109/10826089809059323. [DOI] [PubMed] [Google Scholar]
  38. Nurco DN, Hanlon TE, Balter MB, Kinlock TW, Slaght E. A classification of narcotic addicts based on type, amount, and severity of crime. J Drug Issues. 1991;21:429–448. [Google Scholar]
  39. Pelissier B, Jones N. A review of gender differences among substance abusers. J Res Crime Delinq. 2005;51:343–372. [Google Scholar]
  40. Schwartz RP, Jaffe JH, O'Grady KE, Kinlock TW, Gordon MS, Kelly SM, Wilson ME, Ahmed A. Interim methadone treatment: impact on arrests. Drug Alcohol Depend. 2009;103:148–154. doi: 10.1016/j.drugalcdep.2009.03.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Schwartz RP, Kelly SM, Mitchell SG, Gryczynski J, O'Grady KE, Jaffe JH. When does methadone treatment reduce arrest and severity of arrest charges? An analysis of arrest records. Drug Alcohol Depend. 2017;180:385–390. doi: 10.1016/j.drugalcdep.2017.08.025. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Shrout PE, Fleiss JL. Intraclass correlations: Uses in assessing rater reliability. Psychol Bull. 1979;86:420–428. doi: 10.1037//0033-2909.86.2.420. [DOI] [PubMed] [Google Scholar]
  43. Stewart LM, Henderson CJ, Hobbs MS, Ridout SC, Knuiman MW. Risk of death in prisoners after release from jail. Aust N Z J Public Health. 2004;28:32–36. doi: 10.1111/j.1467-842x.2004.tb00629.x. [DOI] [PubMed] [Google Scholar]
  44. Vocci FJ, Schwartz RP, Wilson ME, Gordon MS, Kinlock TW, Fitzgerald TT, O'Grady KE, Jaffe JH. Buprenorphine dose induction in non-opioid-tolerant pre-release prisoners. Drug Alcohol Depend. 2015;156:133–138. doi: 10.1016/j.drugalcdep.2015.09.001. [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES