Study 1
Gottfredson and colleagues (2003) used an experimental design to test the effectiveness of the Baltimore City Drug Treatment Court (BCDTC) to reduce crime among severely drug-addicted populations. The study looked at the outcome results of 235 individuals randomly assigned to BCDTC or to a treatment-as-usual control group, from February 1997 to August 1998, over a 24-month follow up period. Study participants were identified and referred by Baltimore Court personnel if they met the following criteria: 1) were at least 18 years of age, 2) resided in Baltimore City (Md.), and 3) did not have any past or current convictions for violent offenses.
Treatment group participants were processed in a court that implemented the drug treatment model. Those in the control group underwent the adjudication process in a traditional court, without treatment services. There were 139 people in the treatment group, of which 89.2 percent were African American, 74.1 percent were male, and had an average age of 34.8 years. The control group consisted of 96 individuals who were 89.6 percent African American, 74 percent male, and had an average age of 34.7 years. There were no statistically significant differences between the groups at baseline. Although they were randomly assigned, records indicated that only 91 percent of the treatment group participated in the BCDTC, and 7 percent of the control group assigned to treatment as usual also participated in the BCDTC. The reasons for these mis-assignments were not known; however, the study authors speculated that they may be due to judicial overrides of the randomization process and clerical errors. As a result, the study authors used an intent-to-treat (ITT) analysis, which means that the study participants were treated as randomized, regardless of their actual treatment. This strategy was adopted to preserve the comparability of study groups.
Outcomes of interest were measured through various recidivism reports of rearrests, reconvictions, at least one new drug charge, number of new arrests, number of new charges, and number of new convictions over a 24-month period following the date of randomization. Outcome data were collected from the Maryland Department of Public Safety and Correctional Services and the Baltimore Substance Abuse Services (BSAS), an organization that coordinates drug treatment services in Baltimore. The collected information included demographic characteristics and prior offense history, as well as information on recidivism, drug treatment, drug testing, probation supervision, judicial monitoring, and time spent behind bars over 24-months. The study authors used chi-square tests and t tests to compare the BCDTC participants with control group members. The study authors did not conduct subgroup analyses.
Study 2
Gottfredson and colleagues (2005) conducted a follow up evaluation to Study 1 (Gottfredson et al. 2003) to examine the long-term effects of the BCDTC model. The study authors conducted interviews with 157 of the 235 original Study 1 participants, which was 3 years after they were randomly assigned to either the drug treatment court (n = 93) or the treatment-as-usual control group (n = 64). Participants’ demographic characteristics were the same as the sample in Study 1. In this study, interviews took place between February 2000 and November 2001, and each lasted approximately 90 minutes. There were no statistically significant differences between the interviewed group and the non-interviewed group, except in recidivism. At the 3-year follow up, the interviewed group had more arrests, compared with the non-interviewed group. This finding reflects the fact that jails and prisons were reliable locations to find individuals who were otherwise difficult to track.
Outcomes of interest were recidivism, maximum crime seriousness, crime variety, severity of alcohol and drug addiction, employment, and global mental health severity. Data were collected using the past 12 months as a reference period and at the 3-year follow up. However, employment was measured at the immediate time of the interview. Recidivism was measured by arrests reported through official records. Maximum crime seriousness and crime variety were defined through criminal activity using two scales: the Maximum Crime Seriousness Scale, which combines data regarding 10 property, public order, and violent crimes (with each crime given a numeric value based on the severity of the offense) and the Crime Variety Scale (based on the same 10 crimes), which produces a score based on the total number of different types of crime an individual committed. Alcohol and drug use was measured using three scales, including the Alcohol Addictions Severity Scale (15 questions), the 13-item Drug Addiction Severity Scale, and the Drug Variety Scale (which combines data for 13 drugs/drug types and produces an aggregated score of different drugs/drug types consumed by the individual). Finally, the Brief Symptom Inventory (BSI) Global Mental Health Severity Scale measured the participant’s current level of symptomatology relating to a host of psychological disorders, including somatization, obsession-compulsion, interpersonal sensitivity, depression, anxiety, hostility, phobic anxiety, paranoid ideation, and psychoticism.
The study authors used chi-square tests and t tests to compare the drug court and control group members. In instances in which multiple tests were conducted on a single measurement (i.e., criminal activity and drug use), a Bonferroni adjustment was performed. An ITT analysis was conducted to preserve the comparability of the treatment and control groups. The study authors did not conduct subgroup analyses.
Study 3
Kearley and Gottfredson (2019) conducted a follow up evaluation to Study 1 and Study 2 (Gottfredson et al. 2003; 2005) to examine longer-term effects of the BCDTC model across a 15-year period, as well as the impact of court case assignment (Baltimore City District Court versus Baltimore City Circuit Court). Study authors used the same samples from Study 1 and Study 2 that randomly assigned participants to receive either BCDTC services (n = 139) or treatment as usual in the traditional court (n = 96). The BCDTC treatment group consisted of 139 individuals(84 people who had their cases processed in the District Court and 55 whose cases were processed in the Circuit Court). There were 96 participants in the treatment-as-usual control group (42 individuals had their cases processed in the District Court and 54 had their cases processed in the Circuit Court). All 235 participants were included in interviews at the 15-year follow up.
At intake, treatment group participants averaged 12 prior arrests and 5.3 prior convictions. The majority of the treatment group participants had a current offense charge as a result of a drug crime (71.2 percent), followed by property offense (20.9 percent), public order offense (10.8 percent), prostitution or solicitation (5.8 percent), violation of probation (2.2 percent), or violent or sex offense (1.4 percent); less than 1 percent had a weapons offense (0.7 percent). Control group participants averaged 11.3 prior arrests and 4.6 prior convictions. The majority of the control group participants had a current offense charge as a result of a drug crime (71.9 percent), followed by a property offense (18.8 percent), public order offense (4.2 percent), prostitution or solicitation (4.2 percent), violation of probation (4.2 percent), violent or sex offense (4.2 percent), or weapons crime (3.1 percent). Demographic characteristics (i.e., race and gender) of the total sample were the same as in Study 1 and Study 2. There were no statistically significant differences between treatment and control groups.
Follow up data were collected at the 15-year period from official records as previously mentioned in Study 1 (Gottfredson et al. 2003), from the Maryland Department of Public Safety and Correctional Services and BSAS. Recidivism outcomes of interest at the 15-year follow up included 1) total arrest charges, 2) drug charges, 3) unique arrests (i.e., instances in which arrests do not necessarily lead to the same number of convictions), 4) total number of convicted charges, 5) convicted drug charges, and 6) unique arrests resulting in at least one conviction. All recidivism outcomes were measured cumulatively across the 15-year period and for each year of the follow up.
Negative binomial regression estimates were used, controlling for race, gender, age at randomization, and the number of prior convictions indicated in official records, to determine group differences on outcomes of interest over the 15-year follow up between BCDTC and control conditions The study authors conducted subgroup analyses within the treatment and control groups to determine whether court case assignment (District Court versus Circuit Court) impacted recidivism outcomes.
