A Comparative Study of the Therapeutic Education System for Incarcerated Substance-Abusing Offenders

Abstract

Although many efficacious and cost-effective treatments have been established, very few substance-abusing offenders receive such treatment while incarcerated. This study compares the effectiveness of a computerized intervention, the Therapeutic Education System (TES), with Standard Care on measures of crime (including re-incarceration), drug use, and HIV risk behavior post prison release. Results show that TES and standard treatment were equally effective in reducing criminality, relapse to drug use, and HIV risk behavior. In prisons, where a majority of substance-using offenders do not receive treatment, identifying an equally effective high-volume alternative such as TES can greatly expand access to quality psychosocial interventions.

Keywords

computerized intervention substance abuse treatment correctional settings offenders

Introduction

Substance Use Among Incarcerated Offenders

Offenders engage in disproportionately high rates of substance use while in the community. Recent estimates indicate that approximately 53% of state and 45% of federal prisoners meet criteria for a substance use disorder (Mumola & Karberg, 2007) compared with only 3% of the general U.S. population (Substance Abuse and Mental Health Services [SAMHSA], 2010). In addition, 83% of prisoners report lifetime drug use and more than two thirds report regular use (Mumola & Karberg, 2006). Estimates also suggest that about 50% of male and 33% of female inmates with substance use disorders require substance abuse treatment services in prison (Belenko & Peugh, 2005). However, the best available estimates show that, while incarcerated, only about 20% to 25% of those in need of treatment actually receive it (Chandler, Fletcher, & Volkow, 2009; Taxman, Perdoni, & Harrison, 2007). This unmet need is of great concern as untreated substance-abusing offenders are more likely to relapse to drug use and recidivate (Knight, Simpson, & Hiller, 1999).

Barriers to the Delivery of Substance Abuse Treatment in Prison

Since the mid 1970s, many efficacious and cost-effective treatments for substance use have been established (National Institute on Drug Abuse [NIDA], 2012). However, significant barriers exist to the widespread adoption of science-based psychosocial treatment targeting substance use disorders in criminal justice settings, particularly in prisons, the most significant of which is often cost. As many prisons have cut costs in recent years to accommodate budget reductions (Scott-Hayward, 2009), the cost of offering science-based psychosocial treatments has become increasingly infeasible. These interventions can be very expensive to implement, often requiring financial and staffing resources not typically available in criminal justice settings. When evidence-based interventions are utilized, ensuring fidelity can be difficult, as many programs experience significant staff turnover and/or high patient caseloads. Thus, the limited compatibility of research-based interventions along with criminal justice system realities creates significant barriers to the transfer of evidence-based practice to criminal justice settings. Consequently, innovative approaches are needed to meet these challenges and foster widespread adoption of evidence-based interventions in prisons and other criminal justice settings.

Information Technologies as the Medium of Treatment Delivery

Making use of empirically supported technology can play a critical role in expanding access to effective interventions, thereby helping to prevent and treat substance use disorders in a variety of settings. Indeed, computer-based programs have been used effectively in various therapeutic contexts for more than a decade (Murphy & Mitchell, 1998; Newman, Consoli, & Taylor, 1997; Selmi, Klein, Greist, Sorrell, & Erdman, 1990). Computer or online programs have been reported for substance abuse (Marsch, Bickel, & Grabinski, 2007; Riper et al., 2008), mental health (Postel, de Haan, & De Jong, 2008), and HIV risk behavior (Kiene & Barta, 2006; Marsch & Bickel, 2004). Computerized interventions have the potential to reach a wider audience than is possible using traditional models alone (Marsch, 2011), and computerized interventions have been demonstrated to be as effective as counselor-delivered approaches for motivating, engaging, and treating drug-dependent persons (Bickel, Christensen, & Marsch, 2011). The following studies illustrate typical findings from computer-based treatments where subjects report greater engagement with the program, longer periods of abstinence, and greater satisfaction than more standard therapist-based treatments (Bickel et al., 2011).

Carroll et al. (2008) assigned substance-dependent individuals to a standard therapist-facilitated cognitive behavioral treatment or a standard treatment plus a six-module computer-based cognitive behavioral program. Results indicate that subjects assigned to the computer-based treatment submitted significantly more negative urine screens for any drug and tended to have longer continuous periods of abstinence. Moreover, participants stayed in treatment longer and evaluated the computer-based program more positively. The efficacy of computerized interventions is also supported in studies that have examined the drug use of new mothers. In one study, one session of a computer-delivered motivational intervention for post-partum women resulted in improvements in participant motivation to engage in treatment and address their substance abuse (Ondersma, Chase, Svikis, & Schuster, 2005). In a similar study, one session of computer-delivered motivational intervention combined with a voucher-based reward system to encourage attendance was found to significantly lower drug use for all substances except marijuana for post-partum women (Ondersma, Svikis, & Schuster, 2007).

Evidence Base for the Therapeutic Education System (TES)

Since the advent of computerized technologies for the treatment of substance abuse, perhaps the most widely researched intervention has been the TES, an interactive computerized psychosocial treatment program grounded in the community reinforcement approach (CRA) and cognitive behavioral therapy. Developed as a web-based version of the CRA (Onken, Blaine, & Boren, 1997), TES has prior efficacy data supporting CRA (Budney & Higgins, 1998; Hunt & Azrin, 1973) combined with abstinent-contingent incentives (Bickel, Amass, Higgins, Badger, & Esch, 1997; Higgins et al., 1994; Higgins et al., 2003; Smith, Meyers, & Miller, 2001). TES has been the subject of numerous clinical trials, which have provided mounting evidence of its effectiveness and additional support for the promise of computer-based therapy. One study showed that both the computer- and therapy-based interventions produced a similar total number of abstinent weeks and longer continuous abstinence than standard treatment (Bickel et al., 2008). Importantly, the computer-based treatment required much less therapist time compared with the therapist-led treatments, introducing potential cost savings. In addition, the authors found no difference between the three treatments when assessed by the helping alliance questionnaire, which suggests that computer-based approaches could be used in therapeutic settings without jeopardizing clinically important alliances between clients and counselors.

In a recently completed effectiveness trial conducted within NIDA’s clinical trials network (Campbell et al., 2014), 507 participants from 10 outpatient treatment programs were randomly assigned to 12 weeks of Treatment as Usual (TAU;standard care of at least two on-site sessions per week) or modified TAU + TES (whereby TES substituted for approximately 2 hr of clinician time per week plus motivational incentives contingent on abstinence and retention). Results showed significant positive results favoring TES. Compared with TAU, TES reduced dropout from treatment (hazard ratio [HR] = 0.72, 95% confidence interval [CI] = [0.57, 0.92], p = .01) and increased abstinence in the last 4 weeks of treatment (odds ratio [OR] = 1.62, 95% CI = [1.12, 2.35], p = .01); an effect that was more pronounced among patients with a positive urine drug and/or breath alcohol screen at baseline (n = 228; OR = 2.18, 95% CI = [1.30, 3.68], p = .003).

Objective

The investigative team completed a NIDA-funded, multi-site randomized controlled trial comparing TES with standard care in a sample of substance-abusing offenders in prison. This is the first known clinical trial testing the effectiveness of a computer-based substance abuse treatment program in prison. We hypothesized that TES would be comparably effective with standard care for offenders with substance use disorders in prison settings. This report focuses on outcomes post prison release: self-reported crime, drug use, and HIV risk at 3 and 6 months and re-incarceration at 12 months. This research contributes new empirical information relevant to increasing the delivery of, and access to, science-based, psychosocial treatment with fidelity to individuals in criminal justice settings.

Materials and Method

Trial Design

Incarcerated men and women with substance use disorders were recruited from 10 prisons in four states (seven prisons in Colorado and one prison in each of Kentucky, Pennsylvania, and Washington). Subjects were randomly assigned to the experimental condition, Therapeutic Educational System (E-TES; n = 258), or to the control condition, Standard Care (C; n = 255). This was an open label trial, meaning that treatment assignment was known to both participants and research staff. Randomization to treatment conditions was stratified at the prison level to ensure that an equal number of research subjects were assigned to E-TES and C in each participating prison facility.

Subjects

Although the study was open to all substance-abusing inmates in need of treatment, both treatment interventions (E-TES and C) targeted inmates with low to moderate severity of substance abuse. That is, generally inmates with more severe substance use disorders are often prioritized and assigned to prison-based intensive outpatient or residential treatment. Male and female inmates were eligible for the study if they volunteered to participate in the study and were (a) identified as having a substance use disorder requiring treatment, and they were not already involved in treatment; (b) eligible for parole or had a mandatory release date within 4 to 6 months of recruitment (to allow sufficient time for participation in treatment); (c) at least 18 years of age; and (d) fluent in English. A research assistant obtained informed consent from each eligible inmate. The purpose of the study was reviewed, and the subject’s role in the study was explained. It was made clear that participation (or refusal to participate at any time) would not affect the inmate’s release or treatment status. Inmates who chose to participate in the study signed an itemized consent form to attest to his or her voluntary participation.

The study was reviewed and monitored by National Development & Research Institutes, Inc.’s (NDRI) institutional review board (IRB). The NDRI IRB requested approval from the Office of Human Research Protections (OHRP) to apply to all participating sites. The IRBs from each of the participating sites (Temple University, University of Kentucky, and University of California, Los Angeles) conducted both initial and annual reviews of the study protocols. In some cases, although a blanket approval was in place, participating study sites also requested and received supplementary OHRP approval at the suggestion of their individual IRBs.

A total of 660 eligible inmates were recruited to participate in the study; approximately 13% (n = 85) declined participation or were excluded because of changes in their parole eligibility or access. Inmates who declined participation cited reasons such as not wanting to participate in research, not wanting more treatment, concern that participation in the study would interfere with post release plans, or needing to address other competing priorities. Of the 575 inmates who consented to participate in the study, 89% (n = 513) completed a baseline interview (conducted immediately after informed consent was signed) and were randomly assigned to treatment. Nineteen (nine E-TES; 10 C) subjects were removed from the study post-randomization because they were not released from prison during the follow-up period as was expected. The final sample consisted of 494 offenders with substance use disorders randomly assigned to TES or Standard Care (249 E-TES; 245 C).

Treatment Conditions

TES

The TES intervention included 48 interactive, multimedia modules completed during a planned 12-week cycle. Participants attended TES sessions in a classroom setting located in the prison once a week for 2 hr or twice weekly for 1 hr, depending on lab availability. In some prison facilities, TES admission was rolling (i.e., participants were randomized into the study as they were identified). Thirty-two modules were considered “Core” and were completed in the first 8 weeks. In the remaining 4 weeks, participants explored the 16 optional modules or re-visited previously completed modules.

The core modules have been broadly classified as (a) Substance Use/Abuse (e.g., drug refusal skills, coping with thoughts about using); (b) Risk Reduction for HIV, AIDS, and Sexually Transmitted Infections (e.g., drug use, HIV and hepatitis, identifying/managing triggers for risky sex); (c) Cognitive and Emotional Regulation (e.g., managing negative thinking, anger management, etc.); and (d) Psychosocial Functioning (e.g., effective problem solving, communication skills). The optional modules provide more advanced information on risk reduction and psychosocial functioning. Because modules focus on developing improved approaches to decision-making skills, all have particular relevance for substance-abusing offenders. TES content is provided using informational technologies grounded in scientific principles of effective learning, including “fluency-based” computer-assisted instruction (CAI) and “precision teaching” (e.g., Binder, 1993), to promote mastery of the skills and information being taught, as well as interactive videos modeling various behaviors, and exercises to enhance learning. An electronic reporting system allows summaries of participants’ TES activity.

Psycho-educational/psychosocial substance abuse services-Standard Care (C)

Substance abuse treatment in the Standard Care C condition consisted of the prevailing treatment approach that was being implemented at the time of the study in each of the four state Department of Corrections and 10 prison sites. Although treatment in the C condition was not identical across sites, it was administered using a psycho-educational format that was generally focused on relapse prevention. Inmates in the Standard Care group attended group activities 1 day per week for 2 hr a day over 8 to 12 weeks. Staff members certified as addictions counselors (minimum CAC [Certified Addictions Counselor]) who specialized in substance abuse treatment conducted Standard Care groups in a building at the correctional facility. The Standard Care curriculum covered inmate recognition of the inter-connections among feelings, thoughts, and behavior, and aimed to improve coping mechanisms to handle relapse triggers. In addition, the Standard Care program was designed to increase awareness of the connections between substance use and health/HIV risk behaviors to the individual.

Measures

The primary outcomes of interest for this study represented three domains: criminality, substance use, and HIV risk behavior. Re-incarceration measures were drawn from official Department of Corrections (DOC) records and reflect activity during the 12-month period after release from prison. Re-incarceration outcomes include any return to prison and, more specifically, re-incarceration for a new offense or re-incarceration for a parole/probation technical violation. Self-reported measures for criminal activity, substance use, and HIV risk were collected at 3 and 6 months post prison release. Measures of criminal activity were drawn from data collected for 21 different offense types and include any criminal activity, any drug-related criminal activity, the highest frequency of illegal activity reported for an offense type, and committing a parole/probation violation. Five self-reported outcomes are included for substance use: any drug use, highest frequency of any drug used, any alcohol intoxication, highest frequency of alcohol intoxication reported, and the number of days abstinent. Two measures of HIV risk are also presented: any intravenous drug use and any HIV sexual risk behavior (unprotected sex while high/partner high, when exchanging sex for money/drugs, with an intravenous drug user, with someone HIV+, with partner who refused to use condom, when afraid to ask partner to use a condom). Self-reported data were collected during an interviewer-administered survey (lasting approximately 1 hr) conducted by a trained research assistant.

Re-incarceration outcomes gathered from DOC data were available for 98% of the sample, whereas self-reported outcomes are based on subjects retrieved at 3- and 6-month follow-up. Follow-up interview data were available for 86% (429/498) of the sample at 3 months (88% [221/250] for E-TES and 83% [208/248] for C) and 80% of the sample at 6 months (81% [203/250] for E-TES and 78% [194/248] for C). Retrieval bias was assessed for the total sample, resulting in several differences between inmates who completed both follow-up interviews (3 and 6 months) and those who did not. Apart from the difference in age (M = 37 retrieved; M = 35 not retrieved; p < .02), the differences that emerged suggest greater dysfunction for the retrieved sample. That is, retrieved inmates were more likely to report lifetime mental health symptoms (61% vs. 50%; p < .05), especially depression (46% vs. 32%; p < .02) and anxiety (49% vs. 33%; p < .01). They were also more likely to report a lifetime arrest for a violent offense (61% vs. 45%; p < .01), lifetime marijuana use (88% vs. 78%; p < .03), and alcohol intoxication (84% vs. 74%; p < .03). Differences also emerged for cocaine/crack use (40% vs. 29%; p < .04), alcohol use (77% vs. 57%; p < .001), alcohol use to intoxication (60% vs. 44%; p < .004), and frequency of alcohol intoxication (3.8 vs. 3.1; p < .05) in the 6 months prior to baseline.

Analytic Plan

The primary aim was to test the hypothesis that the E-TES (experimental) condition had comparable effectiveness with the C condition. This was examined by pooling data across all 10 participating prisons. First, intra-class correlations (ICC) were computed for profile and outcome measures at baseline to determine the facility-level effect. ICCs describe the proportion of the observed variance that can be explained by between-facility differences (Marsh et al., 2012). ICCs that are close to zero suggest a lack of variation by facility and can justify aggregating facility-level data (Hofmann, Griffin, & Gavin, 2000). The average of the ICCs in the current study was .040 with a range from .003 to .120, suggesting that facility-level data were in agreement and that there were no significant differences across facilities. Therefore, analyses were based on aggregated site data.

A profile comparison between subjects in the two treatment conditions was conducted to achieve a better understanding of the population and to detect any between-group baseline differences (see Table 1). Chi-square tests were used to compare groups on categorical variables and independent samples t tests were used for interval-level variables. Logistic regression was used to compare rates of re-incarceration, criminal activity, drug use, and HIV risk behavior for the two study conditions. The logistic regression model predicted outcomes post prison release (dependent variables) using treatment condition (key independent variable) and two control variables collected at baseline—alcohol most problematic substance in the last 6 months and opiates most problematic substance in the last 6 months. Control variables were chosen based on group differences detected at baseline. For outcomes other than re-incarceration, the corresponding outcome measure collected at baseline was also included as a covariate. Cox regression was used to compare the number of days to re-incarceration for the two treatment conditions, and ordinary least squares regression was used to compare groups for continuous self-reported outcomes collected at 3 and 6 months post prison release. All analyses included the same two control variables and scored the C condition as the reference group.

Table 1.

Demographic and Other Background Characteristics.

Domain/measure	E-TES	C	Total	χ²
	(n = 249)	(n = 245)	(n = 494)	t test
	%/M (SD)	%/M (SD)	%/M (SD)	p
Demographics
Age	36.7 (9.4)	36.5 (9.8)	36.6 (9.6)	.799
Gender—Male	69.5	69.8	69.6	.939
Race/ethnicity				.130
Black/African American	21.8	21.2	21.5
White	44.9	53.3	49.1
Hispanic	13.2	12.5	12.8
Other/multiple	20.2	12.9	16.6
High school diploma/General	77.5	82.0	79.7	.219
Equivalency Diploma (GED)
L6 employment	67.9	70.2	69.0	.575
Never married	46.0	45.7	45.8	.407
Children	76.2	72.2	74.2	.314
Motivation for treatment	39.0 (10.2)	40.5 (9.4)	39.7 (9.8)	.091
Mental Health
LT mental health treatment	37.6	34.0	35.8	.415
LT any mental health symptom	59.4	57.8	58.6	.710
Depression	45.0	40.6	42.8	.323
Anxiety	47.0	44.7	45.8	.606
Hallucinations	13.7	14.8	14.2	.727
LT prescribed medication	29.3	30.7	30.0	.731
Criminality
Age first arrest	17.4 (6.9)	17.2 (6.1)	17.3 (6.5)	.772
LT number of arrests	18.2 (18.7)	18.3 (21.5)	18.2 (20.1)	.994
LT arrest—Substance-related offense	71.1	72.2	71.7	.775
LT arrest—Property offense	76.3	74.7	75.5	.677
LT arrest—Violent offense	59.8	58.0	58.9	.671
Substance use
Age first alcohol/drug use	14.1 (5.6)	13.8 (4.7)	14.0 (5.2)	.509
LT crack/cocaine	70.3	71.0	70.6	.857
LT opiates	35.7	44.1	39.9	.058
LT meth/amphetamines	54.2	59.6	56.9	.228
L6 drug use	77.0	81.6	79.3	.214
L6 substance most problematic				.009**
Alcohol	38.7	25.2	32.0
Cannabis	12.6	12.8	12.7
Crack/cocaine	18.5	17.1	17.8
Opiates	4.6	11.1	7.8
Meth/amphetamines	17.2	23.9	20.6
Prior substance abuse treatment	59.5	60.3	59.9	.854
HIV Risk
LT injection drug use	22.3	28.0	25.1	.145
L6 injection drug use	14.5	20.9	17.6	.061
L6 number of sex partners	4.1 (7.8)	3.7 (7.7)	3.9 (7.8)	.567
L3 unprotected sex risk^a	63.2	68.7	65.9	.194

Note. E-TES = experimental condition, Therapeutic Educational System; C = control condition; LT = lifetime; L6 = 6 months prior to criminal justice involvement; L3 = 3 months prior to criminal justice involvement.

Includes unprotected sex while high/partner high, when exchanging sex for money/drugs, with an intravenous drug user, with someone HIV+, with partner who refused to use condom, when afraid to ask for condom use.

p < .05. **p < .001.

Results

Profiles

Overall

As indicated in Table 1, the majority of the full sample was male (70%), in their mid-30s (M = 36.6 years; SD = 9.6 years), with a high school diploma/GED (80%), and 69% were employed in the 6 months prior to incarceration. Nearly half were Caucasian (49%), 46% had never been married, and three fourths (74%) had children. Early onset for both substance use (M = 14.0 years; SD = 5.2 years) and arrest (M = 17.3 years; SD = 6.5 years) was evident. On average, inmates had 18.2 lifetime arrests (SD = 20.1) and were most often arrested for drug-related offenses (72%), property offenses (76%), and violent crimes (59%). The majority of inmates reported lifetime use of cocaine (71%) or methamphetamines/amphetamines (57%), and 40% reported opioid use. In the 6 months prior to incarceration, 79% had used illegal drugs, with 32% reporting alcohol as the most problematic, followed by amphetamines/methamphetamines (21%), crack/cocaine (18%), cannabis (13%), and opioids (8%). Most (60%) had received some drug treatment prior to their current incarceration.

Group differences

Treatment groups were significantly different on the substance cited as most problematic during the 6 months prior to incarceration; 39% of E-TES participants versus only 25% of C participants reported alcohol as most problematic, while C participants were more likely to report opiates (11%) and methamphetamines (24%) as the most problematic compared with E-TES participants (5% and 17%, respectively). To account for potential effects from these differences, both these measures were included as statistical controls in the outcome analyses that are presented later in this article.

Comparable Treatment Effects for E-TES

Primary hypothesis: Re-incarceration at 12-months post prison release

Table 2 examines the comparative effectiveness of participation in E-TES on re-incarceration at 12 months post prison discharge. Analysis indicated similar rates of re-incarceration for the two groups overall and separately for new offenses and for parole/technical violations. Approximately one quarter (27.5% E-TES; 21.4% C) of inmates were re-incarcerated for any offense, though less than 10% (7.4% E-TES; 6.7% C) were re-incarcerated for a new offense. Thus, the large majority of subjects (20.1% E-TES; 15.1% C) that were re-incarcerated within the year were due to parole technical violations.

Table 2.

Re-incarceration for a New Offense at 12 Months Post Release (Logistic/Cox Regression).

	E-TES (n = 244)		C (n = 238)		E-TES vs. COR (p)
	%	Minimum-Maximum^a	%	Minimum-Maximum	E-TES vs. COR (p)
Re-incarceration—Any offense	27.5	14.3-36.7	21.4	15.4-33.3	1.41 (.13)
Re-incarceration—New offense	7.4	0.0-9.1	6.7	0.0-13.0	1.13 (.74)
Reincarceraion—Parole/technical violation	20.1	9.1-36.7	15.1	11.9-33.3	1.42 (.17)
	M (SD)	Minimum-Maximum	M (SD)	Minimum-Maximum
Days to re-incarceration—Any offense	181.8 (91.7)	156.1-218.8	225.1 (97.0)	211.7-244.0	0.84 (.08)
Days to re-incarceration—New offense	186.6 (85.7)	103.5-210.0	244.8 (101.9)	234.8-352.0	0.93 (.70)
Days to re-incarceration—Parole/technical violation	180.1 (94.6)	156.1-245.0	220.1 (96.2)	183.8-244.0	0.84 (.13)

Note. Model prison Modified Therapeutic Community (MTC) = 1, an odds ratio less than one indicates a greater likelihood for activity by the C group. Covariates include: L6 alcohol most problematic and L6 opiates most problematic. E-TES = experimental condition, Therapeutic Educational System; C = control condition; OR = odds ratio; L6 = 6 months prior to CJ involvement.

The minimum and maximum values are for the percentages that resulted across the four research sites.

The number of days to re-incarceration for any offense, for a new offense, and for a parole/technical violation was also compared for the two treatment groups. Differences between the two conditions did not reach statistical significance for the number of days to re-incarceration. On average, inmates returned to prison after 181.8 days for E-TES and 225.1 days for C. The average days to re-incarceration were not significantly different for new offenses (186.6 days E-TES; 244.8 days C) or for re-incarcerations as a result of parole/technical violations (180.1 days E-TES; 220.1 days C).

Primary hypothesis: Self-reported outcomes at 3 and 6 month post release

As shown in Table 3, inmates assigned to both treatment groups reported significant and comparable improvement for criminal activity, substance use, and HIV risk at 3 and 6 months post prison release. Inmates in both groups reported a 42% reduction in criminal activity with the average frequency of criminal activity for those reporting involvement in crime decreasing from once a week to once a month. An even greater reduction (53% E-TES; 64% C) was reported for drug-related criminal activity, where any activity fell from about half of the sample to a fifth. A significant reduction (33% E-TES; 37% C) was also present with regard to the occurrence of parole/technical violations.

Table 3.

Self-Reported Outcomes: 3 and 6 Months Post Release Combined (Logistic/OLS Regression).

	E-TES(n = 201)		C(n = 192)		E-TES vs. C
	Baseline 6 months%/M (SD)	3 and 6 months PP%/M (SD)	Baseline 6 months%/M (SD)	3 and 6 months PP%/M (SD)	OR/B (p)
Crime
Criminal activity^a	79.6	45.8	80.2	45.8	1.01 (.97)
Drug-related crime	46.8	21.9	50.0	18.2	0.84 (.51)
Highest frequency activity^b	3.7 (3.0)	1.5 (2.1)	3.9 (3.0)	1.6 (2.4)	−0.14 (.58)
Parole/probation violation	50.7	33.8	54.7	34.4	0.97 (.89)
Drug use
Illegal drug use	78.0	36.0	81.7	38.2	0.99 (.98)
Highest frequency drug use^c	5.1 (2.9)	1.9 (2.7)	5.2 (2.8)	1.9 (2.7)	−0.04 (.91)
Alcohol intoxication	60.2	27.9	59.4	25.5	0.91 (.71)
Highest frequency intoxication^d	3.7 (2.4)	1.3 (2.0)	3.5 (2.3)	1.4 (2.1)	0.03 (.93)
Days abstinent	79.1 (74.2)	159.1 (42.5)	71.6 (72.0)	156.1 (45.5)	4.52 (.30)
HIV risk
Needle use	16.5	2.5	23.4	6.2	1.98 (.24)
Sexual risk^e	64.1	18.2	71.7	19.9	1.13 (.61)

Note. TES = 1, an odds ratio less than one or a (−) B value indicates a greater likelihood for activity by the C group. Covariates include L6 alcohol most problematic and L6 opiates most problematic. OLS = ordinary least squares; E-TES = experimental condition, Therapeutic Educational System; C = control condition; OR = odds ratio; PP = Post Prison (release); L6 = 6 months prior to CJ involvement.

Criminal activity does not include parole or technical violations.

Based on a reduced sample of subjects who reported criminal activity (168 C and 176 E-TES). Frequency codes: 0 = none, 1 = 1 to 3 times, 2 = 1 time/month, 3 = 2 to 3 times/month, 4 = 1 time/week, 5 = 2 to 6 times/week, 6 = 1 time/day, 7 = 2 to 3 times/day, and 8 = 4+ times/day.

Based on a reduced sample of subjects who reported drug use (168 C and 165 E-TES). Frequency codes: 0 = none, 1 = 1 to 3 times, 2 = 1 time/month, 3 = 2 to 3 times/month, 4 = 1 time/week, 5 = 2 to 6 times/week, 6 = 1 time/day, 7 = 2 to 3 times/day, and 8 = 4+ times/day.

Based on a reduced sample of subjects who reported alcohol intoxication (118 C and 130 E-TES). Frequency codes: 0 = none, 1 = 1 to 3 times, 2 = 1 time/month, 3 = 2 to 3 times/month, 4 = 1 time/week, 5 = 2 to 6 times/week, 6 = 1 time/day, 7 = 2 to 3 times/day, and 8 = 4+ times/day.

Baseline measure based on activity in last 3 months. Sexual risk includes unprotected sex while high/partner high, when exchanging sex for money/drugs, with an intravenous drug user, with someone HIV+, with partner who refused to use condom, when afraid to ask for condom use.

The degree of change was statistically similar for measures of substance use. Illegal drug use (54% E-TES; 53% C) and alcohol intoxication (54% E-TES; 57% C) fell by more than half, and the number of days abstinent increased by nearly 3 months (80 E-TES; 85 C). For inmates reporting drug use, the average frequency declined from 4 to 6 times a week to once a month. Likewise, for those reporting alcohol intoxication, intoxication decreased from weekly to 1 to 3 times during the 6-month period. Finally, the vast majority of intravenous drug use did not reoccur at follow-up, and HIV sexual risk behavior was reduced by 72% in both treatment groups. Ultimately, there were no significant differences between groups, indicating that E-TES is equally effective compared with Standard Care.

Discussion

Summary of Results

This study constitutes the first randomized experiment comparing the effectiveness of a computerized intervention, the TES, targeting substance abuse and HIV risk among incarcerated offenders with substance use disorders. Results from this study demonstrate that (a) it is feasible to implement a computerized treatment intervention in a prison setting; (b) participants in both treatment conditions reported substantial improvements for criminality, substance use, and HIV risk post prison release; and (c) TES and standard treatment were equally effective in reducing criminality (re-incarceration and criminal activity), relapse to substance use, and HIV risk behavior.

TES implementation feasibility

During the course of this study, TES was successfully implemented in 10 different prisons operated by the State Department of Corrections in four states, each presenting some challenges regarding installation and implementation. Barriers included, but were not limited to, (a) availability of computer lab space—in some prisons, space had to be retrofitted (e.g., building computer desks and installing wiring); (b) restrictions on inmate access to the Internet—TES, which is most conveniently networked using a web browser, had to be networked to a stand-alone server that would restrict access to all online content other than TES; (c) computer equipment specifications for DOC limited the use of certain hardware or software; (d) competition for lab time from other prison programs that utilized computers; (e) prison security protocols required that inmates be supervised while in the computer labs despite the fact that TES is intended to be an entirely self-directed program; (f) TES was not seen as sufficiently intensive by some prison administrators so participation was limited to inmates with a more moderate substance use problem or those who had already satisfied their treatment mandate and were volunteering for additional services; and (g) expanded access to treatment strained prison staff resources that had been allocated to delivering Standard Care—some prisons did not have enough counseling staff to accommodate inmates who were identified for the study and subsequently assigned to Standard Care so that the research project had to allocate resources (i.e., counselors) to these prisons.

Despite these and other barriers to implementation, TES was eventually made available to 258 inmates across 10 prisons during a 6-month period. Most of these barriers were addressed using resources that were available via the research grant. Specifically, the costs of retrofitting lab space, networking TES to a stand-alone server, purchasing computer equipment, hiring additional counselors to accommodate the increased demand for standard care treatment, and securing proctors for TES lab sessions were either paid for with grant funds or were covered using existing staff resources (e.g., research assistants proctored some TES lab sessions). Thus, the sustainability of TES in participating prisons as well as the adoption of TES in other prisons would require that DOC secure management support as well as the resources necessary to support TES infrastructure and implementation. The majority of costs for TES are incurred at start-up (e.g., equipment, licensing fees, space), yet given the capacity of a computerized intervention, the cost per client and/or per treatment episode can be greatly minimized in prisons where a large number of inmates are in need of substance abuse treatment.

Substantial improvement at follow-up for all study participants

Study participants in both treatment conditions demonstrated substantial and significant improvements on key study outcomes, including several measures of self reported drug use (e.g., any illegal drug use, frequency of illegal drug use, alcohol to intoxication, and days abstinent), crime (e.g., any criminal activity, any drug-related crime, frequency of criminal activity, and parole/probation violations), HIV drug risk (i.e., needle use), and sexual risk behaviors (i.e., unprotected sex with risky partners). These outcomes suggest that both interventions (i.e., computerized and clinician-delivered therapy) can be used with incarcerated individuals to reduce these maladaptive behaviors post prison release. Furthermore, comparable effectiveness between interventions suggests that TES can be used as a stand-alone intervention—an alternative to standard care—to effectively treat substance-abusing offenders. Furthermore, study results extend the evidence base for TES by demonstrating its feasibility and effectiveness with offenders and within criminal justice settings. Prior studies have focused almost entirely on evaluating the impact of TES in outpatient substance abuse treatment populations. The present project is an extension of the evidence base to show the impact of TES on recidivism rates; a key outcome variable for any agency that works with a criminal justice population.

Group differences

Comparable results for the computerized intervention lend additional support for TES as an effective approach that can serve as an alternative to standard prison treatment for incarcerated individuals. In this setting, comparable effectiveness is considered to be a positive outcome given the goal of expanding access to treatment. That is, given the trend that so few individuals in need of treatment receive it while incarcerated, an equally effective alternative that provides opportunities for expanding treatment services (e.g., computerized and self-administered) should be considered as a means for improving access to treatment in prisons and other treatment settings. In this regard, prisons would have to determine the best way to utilize TES in terms of identifying specific target populations (e.g., drug use severity), treatment modalities, as well as the overall implementation approach.

Furthermore, findings of comparable effectiveness are consistent with several prior studies demonstrating the positive impact of TES on substance use behaviors. This research has typically shown that TES is equally effective to interventions of similar content (i.e., CRA delivered by a trained therapist), but more effective when compared with Standard Care. Although this study did not produce significantly better outcomes for TES participants compared with Standard Care, unique design features could explain this finding. For example, this study did not make use of the voucher incentive system that is available with TES and has been routinely used in the context of many of the prior studies conducted in outpatient substance abuse treatment settings. This decision was made because the study took place in prisons where payment to offenders for participation in treatment is not feasible. Thus, the effects of TES were examined without the benefits of contingency management. Only one prior study has been able to demonstrate the effectiveness of computerized CRA in the absence of contingency management (Marsch et al., 2014); however, it did not constitute a randomized controlled trial that was testing the efficacy of TES with and without prize-based incentives.

Although in this study TES was implemented as a stand-alone intervention, there are many potential ways to utilize computerized interventions to expand access to treatment and improve clinical outcomes. For example, wait lists have been established in many prisons because the availability of outpatient slots is limited. Thus, TES could be offered to those who will not get access to treatment or in preparation for placement in other programs (e.g., intensive outpatient or residential) as a way to increase knowledge, engagement, and motivation. TES can also be used as a “clinician extender,” whereby offenders currently in treatment can complete assigned modules as an adjunct to treatment, which can be integrated as a component of the clinical approach. This will free up additional face-to-face time for clinicians and allow for the delivery of more services. TES could also be assigned to all newly identified inmates in need of treatment or to all inmates as a pre-release refresher. All these represent valuable uptakes of TES and would be dependent on the particular needs of each prison.

Regardless of the way in which it is used, TES should be considered for its potential to improve access to treatment for offenders, many of whom would not otherwise receive any treatment while incarcerated. Based on study observations, TES was responsive to the more moderately impaired clients, thereby freeing the limited number of trained clinicians to deal with more severely impaired offenders. That is, an effective equivalent to standard care sets up the potential to triage substance-using offenders so that more moderate users can be referred to TES while the more severely impaired users can be reserved for more intensive interpersonal services delivered by trained clinicians.

Finally, although many of the inmates who participated in TES as part of this study would be classified as having mild to moderate substance abuse issues, TES has proven to be effective with a variety of substance users, in a number of different settings, and when used in a variety of ways (e.g., as a stand-alone intervention as well as an adjunct to treatment or what has often been termed as a “clinician extender”). For example, TES has also proven to be effective with more severe users that include cocaine and opioid dependent individuals receiving outpatient treatment (see, for example, Bickel et al., 2008; Campbell et al., 2014). As was previously mentioned, TES also has a built-in incentives system that has proven effective with substance users, which is consistent with the vast literature available on the effectiveness of contingency management approaches (Petry, 2013).

Limitations

A number of limitations affect the interpretation of results presented in this article. First, without the inclusion of a no-treatment control group, it cannot be concluded definitively that these substantial improvements evidenced at follow-up for both treatment conditions are in fact attributable to treatment participation. Instead, it could be argued that other factors (e.g., imprisonment) contributed to such improvements. Nevertheless, findings of comparable effectiveness between interventions suggest that TES participants do no worse, thereby offering a viable alternative to the standard treatment services.

Second, although a CAC minimum administered substance abuse treatment with a psycho-educational format at each site to all offenders in the C condition over 12 weeks (2 hr per week), each of the four DOCs (CO, KY, PA, and WA) delivered their own standard treatment curricula (e.g., SAMHSA, 2006). A standardized approach for the C condition was not feasible, which made it difficult to determine whether pooling of the data confounded the analysis.

Third, because the study expanded access to treatment, Colorado prisons had difficulty accommodating inmates that were being identified as in need of treatment but did not meet the criteria for participation in existing DOC interventions (e.g., intensive outpatient, residential or therapeutic communities) because their drug abuse was not severe enough. In such instances, the research study provided standard care interventions in these prisons, which were arguably more comprehensive than that which is typically administered in many prison settings, including prisons participating in Kentucky, Pennsylvania, and Washington. This created the possibility that standard care as a hybrid comparison was more effective than would have been the case given a more natural experiment.

Fourth, neither site differences nor treatment dosage effects were examined as part of the analysis because the primary focus of this article is on the first aim of the research in accordance with the intent-to-treat design of the clinical trial. With regard to treatment dosage, although both interventions were measured and compared in terms of session attendance, TES dosage is best measured in terms of the number of modules completed. The data demonstrate that although participants in both groups completed an average of six sessions, TES subjects completed three quarters of the intervention (i.e., an average of 24 of 32 TES modules) while more than half (56%) completed all 32 core modules (only 15% of control subjects attended 12 sessions). Thus, it would be important to explore outcomes based on comparable intervention doses and to differentiate the impact of increased utilization of TES modules. These secondary analyses will be the subject of subsequent articles by the investigative team.

Finally, as reported for both the 3- and 6-month follow-up, outcomes are available for more than 80% of the sample, but those that were not included in the analysis tend to be more functional (i.e., the analysis sample was more dysfunctional). This certainly raises some questions as to the potential impact on the aggregate outcome had those more functional offenders been available for analysis. However, these findings, which are best extrapolated to those offenders reporting more dysfunction (e.g., mental health, severity of substance use), certainly suggest that TES could be effective with complex and higher risk substance-using populations if this need were to arise in other treatment settings.

Future Research

Although effective technology-based interventions such as TES have been identified and rigorously studied, moving services from carefully controlled research trials into “real world” practice is often complicated. Ultimately, despite the potential of TES to achieve a significant expansion of services to substance-using offenders as well as at least one state (Colorado Department of Corrections[CO-DOC]) having the infrastructure necessary to support expansion already in place, widespread implementation of TES in prisons remains a challenge and somewhat unlikely. As a direct follow-up to this study, a supplementary study could examine factors that influence implementation to identify effective strategies for promoting technology adoption. For the CO-DOC in particular, such a study would help to identify the ideal application of TES to increase access to treatment for substance-abusing offenders in prison.

Conclusion

In prisons across the United States, where resource constraints mean that a majority of substance-using offenders do not receive treatment, identifying an effective alternative such as TES that can be delivered in large volumes with high fidelity (due to its computerized format) is one method that can potentially remedy this gap. In this regard, TES could advance treatment in correctional settings by improving access and exposing more in-need offenders to quality psychosocial interventions.

Footnotes

Declaration of Conflicting Interests

The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: In addition to her academic affiliation, Dr. Marsch is affiliated with HealthSim, LLC, the health-promotion software development organization that developed the web-based Therapeutic Education System referenced in this manuscript. Dr. Marsch has worked extensively with her institutions to manage any potential conflict of interest. All research data collection, data management, and statistical analyses were conducted by individuals with no affiliation to HealthSim, LLC. The remaining author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Author Biographies

Michael Chaple, PhD, a criminologist, is the deputy director of the Center for the Integration of Research and Practice (CIRP) at National Development & Research Institutes, Inc. (NDRI) in New York City. He has served as project director on numerous NIDA-funded multi-site clinical trials and has more than a decade of experience coordinating and delivery training/technical assistance initiatives designed to translate research findings into practical guidelines for immediate clinical application.

Stanley Sacks, PhD, a clinical psychologist and research scientist, is the director of CIRP at NDRI in New York City. The author of numerous publications, he has been a principal investigator on NIDA-funded studies and has also served as the chair and lead writer for Substance Abuse and Mental Health Services’ (SAMHSA) Treatment Improvement Protocol (TIP) 42, the expert leader for SAMHSA’s Co-Occurring Center of Excellence (COCE), director of the New York State Health Foundation’s Center for Excellence in Integrated Care, and director of the Northeast and Caribbean Addiction Technology Transfer Center (NeC-ATTC) serving Health and Human Services (HHS) Region 2.

Karen McKendrick, MHP, data systems and analysis director in CIRP, has provided statistical and data management services to NDRI for more than 15 years and has made significant contributions to all CIRP research studies. She has developed extensive detailed data management protocols that govern all CIRP studies and has authored and co-authored several papers relating to the treatment of substance use and co-occurring disorders.

Lisa A. Marsch, PhD, is the director of the Center for Technology and Behavioral Health at Dartmouth College and a faculty member of Dartmouth’s Department of Psychiatry. With funding from the National Institutes of Health (NIH), she has led a line of research focused on the development and evaluation of state of the art, technology-based (mobile-, and Internet-delivered) interventions targeting substance abuse and mental health.

Steven Belenko, PhD, is professor in the Temple University Department of Criminal Justice, Adjunct Professor in the Department of Psychiatry, University of Pennsylvania School of Medicine, and director of the Methods Core for the Center for Behavioral Health Services and Criminal Justice Research, Rutgers University. His research interests focus on strategies for improving the implementation of evidence-based drug treatment, HIV service interventions, and other health services into the criminal justice and juvenile justice systems.

Carl Leukefeld, DSW, is professor and chair of the Department of Behavioral Science and founding director of the Center on Drug and Alcohol Research, University of Kentucky. He has published more than 200 articles, chapters, books, and monographs. His research interests include treatment interventions, HIV prevention and treatment, criminal justice sanctions, and health services.

Michael Prendergast, PhD, is the director of the Criminal Justice Research Group at the University of California, Los Angeles (UCLA) Integrated Substance Abuse Program (ISAP). He has been a principal investigator of numerous state- and federally funded projects to study drug treatment strategies in the criminal justice system. His research interests include interventions for drug-abusing offenders, systematic reviews of treatment effectiveness, and strategies for implementing evidence-based practices.

Michael French, PhD, is a health economist and professor in the Department of Sociology at the University of Miami. He has been a principal investigator on numerous federal, state, and foundation grants and has published more than 150 peer-reviewed articles. His research interests include health policy, program evaluation, substance abuse, pharmaco-economics, alternative health care delivery systems, and the economics of crime.

References

Belenko

Peugh

(2005). Estimating drug treatment needs among state prison inmates. Drug and Alcohol Dependence, 77, 269-281. doi:10.1016/j.drugalcdep.2004.08.023

Bickel

W. K.

Amass

Higgins

S. T.

Badger

G. J.

Esch

R. A.

(1997). Effects of adding behavioral treatment to opioid detoxification with buprenorphine. Journal of Consulting and Clinical Psychology, 65, 803-810.

Bickel

W. K.

Marsch

L. A.

Buchhalter

Badger

(2008). Computerized behavior therapy for opioid-dependent outpatients: A randomized controlled trial. Experimental & Clinical Psychopharmacology, 16, 132-143. doi:10.1037/1064-1297.16.2.132

Bickel

W. K.

Christensen

D. R.

Marsch

L. A.

(2011). A review of computer-based interventions used in the assessment, treatment, and research of drug addiction. Substance Use & Misuse, 46, 4-9. doi:10.3109/10826084.2011.521066

Binder

L. M.

(1993). An abbreviated form of the Portland Digit Recognition Test. The Clinical Neuropsycholgoist, 7, 104-107. doi:10.1080/13854049308401892

Budney

A. J.

Higgins

S. T.

(1998). Therapy manuals for drug addiction, a community reinforcement plus vouchers approach: Treating cocaine addiction. Rockville, MD: National Institute on Drug Abuse.

Campbell

A. N.

Nunes

E. V.

Matthews

A. G.

Stitzer

Miele

G. M.

. . . Ghitza

U. E.

(2014). Internet-delivered treatment for substance abuse: A multisite randomized controlled trial. American Journal of Psychiatry, 171(6), 683-690. doi:10.1176/appi.ajp.2014.13081055

Carroll

K. M.

Ball

S. A.

Martino

Nich

Babuscio

T. A.

Nuro

K. F.

. . .Rounsaville

B. J.

(2008). Computer-assisted delivery of cognitive-behavioral therapy for addiction: A randomized trial of CBT4CBT. American Journal of Psychiatry, 165, 881-888. doi:10.1176/appi.ajp.2008.07111835

Chandler

R. K.

Fletcher

B. W.

Volkow

N. D.

(2009). Treating drug abuse and addiction in the criminal justice system. The Journal of American Medical Association, 301, 183-190. doi:10.1001/jama.2008.976

10.

Higgins

S. T.

Budney

A. J.

Bickel

W. K.

Foerg

F. E.

Donham

Badger

G. J.

(1994). Incentives improve outcome in outpatient behavioral treatment of cocaine dependence. Archives of General Psychiatry, 51, 568-576.

11.

Higgins

S. T.

Sigmon

S. C.

Wong

C. J.

Heil

S. H.

Badger

G. J.

. . . Anthony

(2003). Community reinforcement therapy for cocaine-dependent outpatients. Archives of General Psychiatry, 60, 1043-1052. doi:10.1001/archpsyc.60.9.1043

12.

Hofmann

D. A.

Griffin

M. A.

Gavin

M. B.

(2000). The application of hierarchical linear modeling to organizational research. In Klein

K. J.

Kozlowski

S. J.

(Eds.), Multilevel theory, research, and methods in organizations: Foundations, extensions, and new directions (pp. 467-511). San Francisco, CA: Jossey-Bass.

13.

Hunt

G. M.

Azrin

N. H.

(1973). A community reinforcement approach to alcoholism. Behavioral Research and Therapy, 11, 91-104.

14.

Kiene

S. M.

Barta

W. D.

(2006). A brief individualized computer-delivered sexual risk reduction intervention increases HIV/AIDS preventive behavior. Journal of Adolescent Health, 39, 404-410.

15.

Knight

Simpson

D. D.

Hiller

M. L.

(1999). Three-year reincarceration outcomes for in-prison therapeutic community treatment in Texas. The Prison Journal, 79, 337-351. doi:10.1177/0032885599079003004

16.

Marsch

L. A.

(2011). Technology-based interventions targeting substance use disorders and related issues: An editorial. Substance Use & Misuse, 46, 1-3. doi:10.3109/10826084.2011.521037

17.

Marsch

L. A.

Bickel

W. K.

(2004). Efficacy of computer-based HIV/AIDS education for injection drug users. American Journal of Health Behavior, 28, 316-327. doi:10.5993/AJHB.28.4.3

18.

Marsch

L. A.

Bickel

W. K.

Grabinski

M. J.

(2007). Application of interactive, computer technology to adolescent substance abuse prevention and treatment. Adolescent Medicine State of the Art Reviews, 18(2), 342-356.

19.

Marsch

L. A.

Guarino

Acosta

Aponte-Melendez

Cleland

Grabinski

. . . Edwards

(2014). Web-based behavioral treatment for substance use disorders as a partial replacement of standard methadone maintenance treatment. Journal of Substance Abuse Treatment, 46, 43-51. doi:10.1016/j.jsat.2013.08.012

20.

Marsh

H. W.

Ludtke

Nagengast

Trautwein

Morin

A. S.

Abduljabbar

A. S.

Koller

(2012). Classroom climate and contextual effects: Conceptual and methodological issues in the evaluation of group-level effects. Educational Psychologist, 47, 106-124. doi:10.1080/00461520.2012.670488

21.

Mumola

Karberg

(2007). Drug use and dependence, state and federal prisoners, 2004 (Special report). Bureau of Justice Statistics, Office of Justice Programs. Retrieved from http://www.bjs.gov/content/pub/pdf/dudsfp04.pdf

22.

Murphy

Mitchell

D. L.

(1998). When writing helps to heal: E-mail as therapy. British Journal of Guidance & Counseling, 26, 21-32. doi:10.1080/03069889808253835

23.

National Institute on Drug Abuse. (2012). Principles of drug addiction treatment: A research-based guide. Retrieved from http://www.drugabuse.gov/sites/default/files/podat_1.pdf

24.

Newman

M. G.

Consoli

Taylor

C. B.

(1997). Computers in assessment and cognitive behavioral treatment of clinical disorders: Anxiety as a case in point. Behavior Therapy, 28, 211-235.

25.

Ondersma

S. J.

Chase

S. K.

Svikis

D. S.

Schuster

C. R.

(2005). Computer-based brief motivational intervention for perinatal drug use. Journal of Substance Abuse Treatment, 28, 305-313. doi:10.1016/j.jsat.2005.02.004

26.

Ondersma

S. J.

Svikis

D. S.

Schuster

C. R.

(2007). Computer-based brief intervention: A randomized trial with postpartum women. American Journal of Preventive Medicine, 32, 231-238. doi:10.1016/j.amepre.2006.11.003

27.

Onken

L. S.

Blaine

J. D.

Boren

J. J.

(1997). Treatment for drug addiction: It won’t work if they don’t receive it. In Onken

L. S.

Blaine

J. D.

Boren

J. J.

(Eds.), Beyond the therapeutic alliance: Keeping the drug-dependent individual in treatment (Vol. 165, pp. 1-3). Rockville, MD: National Institute on Drug Abuse.

28.

Petry

(2013). Contingency management for substance abuse treatment: A guide to implementing this evidence-based practice. New York, NY: Routledge.

29.

Postel

M. G.

de Haan

H. A.

De Jong

C. A.

(2008). E-therapy for mental health problems: A systematic review. Telemedicine Journal and e-Health, 14, 707-714. doi:10.1089/tmj.2007.0111

30.

Riper

Kramer

Smit

Conijn

Schippers

Cuijpers

(2008). Web-based self-help for problem drinkers: A pragmatic randomized trial. Addiction, 103, 218-227. doi:10.1111/j.1360-0443.2007.02063.x

31.

Scott-Hayward

C. S.

(2009). The fiscal crisis in corrections: Rethinking policies and practices. New York, NY: The Center on Sentencing and Corrections, VERA Institute of Justice.

32.

Selmi

P. M.

Klein

M. H.

Greist

J. H.

Sorrell

S. P.

Erdman

H. P.

(1990). Computer-administered cognitive-behavioral therapy for depression. American Journal of Psychiatry, 147, 51-56.

33.

Smith

J. E.

Meyers

R. J.

Miller

W. R.

(2001). The community reinforcement approach to the treatment of substance use disorders. The American Journal of Addictions, 10(Suppl. s1), s51-s59.

34.

Substance Abuse and Mental Health Services. (2010). Results from the 2010 national survey on drug use and health: Summary of national findings. Retrieved from http://www.samhsa.gov/data/NSDUH/2k10NSDUH/2k10Results.htm

35.

Substance Abuse and Mental Health Services Administration (SAMHSA). (2006). Counselors manual for relapse prevention with chemically dependent criminal offenders. Technical Assistance Publication (TAP) Series 19. Rockville, MD: SAMHSA, Center for Substance Abuse Treatment (CSAT). Retrieved from http://adaiclearinghouse.org/downloads/TAP-19-Counselors-Manual-for-Relapse-Prevention-with-Chemically-Dependent-Criminal-Offenders-109.pdf (accessed 11 January 2014).

36.

Taxman

F. S.

Perdoni

M. L.

Harrison

L. D.

(2007). Drug treatment services for adult offenders: The state of the state. Journal of Substance Abuse Treatment, 32, 239-254. doi:10.1016/j.jsat.2006.12.019