Interventions for female drug-using offenders

Background This review represents one in a family of three reviews focusing on the effectiveness of interventions in reducing drug use and criminal activity for offenders.

Objectives To assess the effectiveness of interventions for female drug‐using offenders in reducing criminal activity, or drug use, or both.

Search methods We searched 12 electronic bibliographic databases up to February 2019.

Selection criteria We included randomised controlled trials (RCTs).

Data collection and analysis We used standard methodological procedures expected by Cochrane.

Main results We included 13 trials with 2560 participants. Interventions were delivered in prison (7/13 studies, 53%) and community (6/13 studies, 47%) settings. The rating of bias was affected by the lack of clear reporting by authors, and we rated many items as 'unclear'.

In two studies (190 participants) collaborative case management in comparison to treatment as usual did not reduce drug use (risk ratio (RR) 0.65, 95% confidence interval (CI) 0.20 to 2.12; 1 study, 77 participants; low‐certainty evidence), reincarceration at nine months (RR 0.71, 95% CI 0.32 to 1.57; 1 study, 77 participants; low‐certainty evidence), and number of subsequent arrests at 12 months (RR 1.11, 95% CI 0.83 to 1.49; 1 study, 113 participants; low‐certainty evidence).

One study (36 participants) comparing buprenorphine to placebo showed no significant reduction in self‐reported drug use at end of treatment (RR 0.57, 95% CI 0.27 to 1.20) and three months (RR 0.58, 95% CI 0.25 to 1.35); very low‐certainty evidence. No adverse events were reported.

One study (38 participants) comparing interpersonal psychotherapy to a psychoeducational intervention did not find reduction in drug use at three months (RR 0.67, 95% CI 0.30 to 1.50; low‐certainty evidence).

One study (31 participants) comparing acceptance and commitment therapy (ACT) to a waiting list showed no significant reduction in self‐reported drug use using the Addiction Severity Index (mean difference (MD) ‐0.04, 95% CI ‐0.37 to 0.29) and abstinence from drug use at six months (RR 2.89, 95% CI 0.73 to 11.43); low‐certainty evidence.

One study (314 participants) comparing cognitive behavioural skills to a therapeutic community programme and aftercare showed no significant reduction in self‐reported drug use (RR 0.86, 95% CI 0.58 to 1.27), re‐arrest for any type of crime (RR 0.73, 95% CI 0.52 to 1.03); criminal activity (RR 0.80, 95% CI 0.63 to 1.03), or drug‐related crime (RR 0.95, 95% CI 0.68 to 1.32). A significant reduction for arrested (not for parole) violations at six months follow‐up was significantly in favour of cognitive behavioural skills (RR 0.43, 95% CI 0.25 to 0.77; very low‐certainty evidence). A second study with 115 participants comparing cognitive behavioural skills to an alternative substance abuse treatment showed no significant reduction in reincarceration at 12 months (RR 0.70, 95% CI 0.43 to 1.12; low certainty‐evidence.

One study (44 participants) comparing cognitive behavioural skills and standard therapy versus treatment as usual showed no significant reduction in Addiction Severity Index (ASI) drug score at three months (MD 0.02, 95% CI ‐0.05 to 0.09) and six months (MD ‐0.02, 95% CI ‐0.09 to 0.05), and incarceration at three months (RR 0.46, 95% CI 0.04 to 4.68) and six months (RR 0.51, 95% CI 0.20 to 1.27); very low‐certainty evidence.

One study (171 participants) comparing a single computerised intervention versus case management showed no significant reduction in the number of days not using drugs at three months (MD ‐0.89, 95% CI ‐4.83 to 3.05; low certainty‐evidence).

One study (116 participants) comparing dialectic behavioural therapy and case management (DBT‐CM) versus a health promotion intervention showed no significant reduction at six months follow‐up in positive drug testing (RR 0.67, 95% CI 0.43 to 1.03), number of people not using marijuana (RR 1.23, 95% CI 0.95 to 1.59), crack (RR 1.00, 95% CI 0.87 to 1.14), cocaine (RR 1.02, 95% CI 0.93 to 1.12), heroin (RR 1.05, 95% CI 0.98 to 1.13), methamphetamine (RR 1.02, 95% CI 0.87 to 1.20), and self‐reported drug use for any drug (RR 1.20, 95% CI 0.92 to 1.56); very low‐certainty evidence.

One study (211 participants) comparing a therapeutic community programme versus work release showed no significant reduction in marijuana use at six months (RR 1.03, 95% CI 0.19 to 5.65), nor 18 months (RR 1.00, 95% CI 0.07 to 14.45), heroin use at six months (RR 1.59, 95% CI 0.49 to 5.14), nor 18 months (RR 1.92, 95% CI 0.24 to 15.37), crack use at six months (RR 2.07, 95% CI 0.41 to 10.41), nor 18 months (RR 1.64, 95% CI 0.19 to 14.06), cocaine use at six months (RR 1.09, 95% CI 0.79 to 1.50), nor 18 months (RR 0.93, 95% CI 0.64 to 1.35). It also showed no significant reduction in incarceration for drug offences at 18 months (RR 1.45, 95% CI 0.87 to 2.42); with overall very low‐ to low‐certainty evidence.

One study (511 participants) comparing intensive discharge planning and case management versus prison only showed no significant reduction in use of marijuana (RR 0.79, 95% CI 0.53 to 1.16), hard drugs (RR 1.12, 95% CI 0.88 to 1.43), crack cocaine (RR 1.08, 95% CI 0.75 to 1.54), nor positive hair testing for marijuana (RR 0.75, 95% CI 0.55 to 1.03); it found a significant reduction in arrests (RR 0.19, 95% CI 0.04 to 0.87), but no significant reduction in drug charges (RR 1.07, 95% CI 0.75 to 1.53) nor incarceration (RR 1.09, 95% CI 0.86 to 1.39); moderate‐certainty evidence.

One narrative study summary (211 participants) comparing buprenorphine pre‐ and post‐release from prison showed no significant reduction in drug use at 12 months post‐release; low certainty‐evidence. No adverse effects were reported.

Authors' conclusions The studies showed a high degree of heterogeneity for types of comparisons, outcome measures and small samples. Descriptions of treatment modalities are required. On one outcome of arrest (no parole violations), we identified a significant reduction when cognitive behavioural therapy (CBT) was compared to a therapeutic community programme. But for all other outcomes, none of the interventions were effective. Larger trials are required to increase the precision of confidence about the certainty of evidence.