Study 1
Wexler, De Leon, and colleagues (1999a) conducted a randomized controlled trial to evaluate the Amity In-Prison Therapeutic Community (TC). Volunteers were recruited from the general population of a medium-security prison in San Diego, Calif. The California Department of Corrections (DOC) and Amity staff reviewed incarcerated persons' records to identify eligible incarcerated persons. Incarcerated persons were eligible to participate in the program if they met the admission criteria of having a drug problem and being 9 to 14 months from parole. Incarcerated persons who were convicted of arson or sexual crimes involving minors were excluded. The remaining eligible incarcerated persons in the volunteer pool were randomly assigned to the treatment condition as bed space became available. Incarcerated persons who were not randomly selected to participate became members of the no-treatment control group. The control group did not receive formal substance abuse treatment while in prison, although limited drug education and 12-step groups were available.
The study sample included 715 males, of which 425 were in the treatment group, and 290 were in the no-treatment control group. The treatment group had an average age of 31.2 years and was 39.5 percent white, 28.4 percent Hispanic, 23 percent African American, and 9.2 percent was other. The control group had an average age of 30.5 years and was 35.3 percent white, 32.5 percent Hispanic, 21.6 percent African American, and 10.6 percent was other. The only statistically significant difference between the treatment and control groups at baseline was in educational level; the control group had a slightly greater number of participants who had continued their education beyond the high school level.
The outcome of interest was reincarceration, which was measured as a return to prison for either a parole violation or for a new arrest. The study looked at the 12- and 24-month periods following participants’ release from prison (the CrimeSolutions review of the study focused on the outcomes at the 24-month follow up). Follow up was conducted on all study participants for at least 12 months following their release from prison, and 263 study participants were observed for the 24-month follow-up period. Reincarceration data were obtained from the California DOC’s computerized Offender Based Information System (OBIS). Intent-to-treat analyses were conducted using bivariate analyses (chi square).
Subgroup analyses looked at reincarceration rates separately for the four treatment subgroups using multivariate analyses. The treatment group was divided into four subgroups: 1) incarcerated persons who entered the in-prison TC program, but who left early for disciplinary or personal reasons (prison TC dropouts); 2) incarcerated persons who completed the in-prison TC, but who did not choose to participate in the aftercare TC available in the community (prison TC treatment completers); 3) incarcerated persons who completed the in-prison TC and volunteered to attend the aftercare TC, but who dropped out in less than 3 months (aftercare TC treatment dropouts); and 4) incarcerated persons who completed the prison and Vista aftercare TC programs (aftercare TC completers).
Study 2
Wexler, Melnick, and colleagues (1999b) conducted a 36-month follow-up assessment using study participants from the original sample in Study 1 (Wexler et al. 1999a). Of the original sample of 715 study participants, this study included 478 participants (289 members in the treatment group and 189 members in the control group).
The average age of the treatment group members was 30.85 years, and they were 31 percent African American, 28 percent white, 27 percent Hispanic, and 4 percent was other. The average age of the control group members was 30.48 years, and they were 36 percent African American, 35 percent white, 22 percent Hispanic, and 7 percent was other. The only statistically significant difference between the treatment and control groups at baseline was age and race/ethnicity, which were statistically significant predictors for reincarceration but not controlled for in the analysis.
The outcome of interest was reincarceration within 3 years following release from prison. Reincarceration included a return to prison for either a parole violation or for a new arrest. Return-to-prison data were collected from the DOC’s OBIS. Intent-to-treat analyses was conducted using bivariate analyses (chi-square).
Subgroup analyses looked at reincarceration rates separately for the three subgroups using multivariate analyses. The treatment group consisted of three subgroups: 1) prison TC dropouts, 2) prison TC treatment completers, and 3) aftercare TC completers (aftercare TC treatment dropouts were not examined).
Study 3
Prendergast and colleagues (2004) conducted a 5-year follow-up assessment using study participants from the original sample in Study 1 (Wexler et al. 1999a). For those study participants who could be located, interviews were conducted in the community, in jails and prisons, and out of state. Of the original sample of 715 study participants, 81 percent were interviewed for the follow-up study, 5 percent were confirmed to have died, 3 percent refused to participate, 2 percent had been deported, and 10 percent could not be located. The final sample consisted of 576 participants, including 341 treatment group members and 235 control group members. There were no statistically significant differences between the study participants who were interviewed and the study participants who could not be interviewed, except for race. There were fewer white participants in the group that was interviewed, compared with the group of study participants that was not interviewed.
The outcomes of interest were reincarceration and heavy drug use. Reincarceration was defined as a return to prison in the 5 years since release. Heavy drug use was defined as self-report of having used drugs (excluding alcohol and marijuana) several times per week in the 12 months before the interview. To supplement the self-reported information on criminal justice system involvement, arrest records were also obtained from the California Department of Justice’s Automated Criminal History System (ACHS). Incarceration and parole records were also collected from the California DOC’s OBIS. Reincarceration data were censored at 1,825 days (5 years).
Intent-to-treat analyses were conducted using bivariate analyses (chi square). Subgroup analyses looked at reincarceration rates separately for the four subgroups of the treatment group using multivariate analyses: 1) prison TC dropouts; 2) prison TC treatment completers; 3) aftercare TC treatment dropouts; and 4) aftercare TC completers.
