Mentoring adolescents to prevent drug and alcohol use

Many adolescents receive mentoring. There is no systematic review if mentoring prevents alcohol and drug use.

Assess effectiveness of mentoring to prevent adolescent alcohol/drug use.

Cochrane CENTRAL (issue 4), MEDLINE (1950-to July 2011), EMBASE (1980-to July 2011), 5 other electronic and 11 Grey literature electronic databases, 10 websites, reference lists, experts in addictions and mentoring.

Randomised controlled trials (RCTs) of mentoring in adolescents to prevent alcohol/drug use.

We identified 2,113 abstracts, independently assessed 233 full-text articles, 4 RCTs met inclusion criteria. Two reviewers independently extracted data and assessed risks of bias. We contacted investigators for missing information.

We identified 4 RCTs (1,194 adolescents). No RCT reported enough detail to assess whether a strong randomisation method was used or allocation was concealed. Blinding was not possible as the intervention was mentoring. Three RCTs provided complete data. No selective reporting.Three RCTs provided evidence about mentoring and preventing alcohol use. We pooled two RCTs (RR for mentoring compared to no intervention = 0.71 (95% CI = 0.57 to 0.90, P value = 0.005). A third RCT found no significant differences.Three RCTs provided evidence about mentoring and preventing drug use, but could not be pooled. One found significantly less use of "illegal" drugs," one did not, and one assessed only marijuana use and found no significant differences.One RCT measured "substance use" without separating alcohol and drugs, and found no difference for mentoring.

All four RCTs were in the US, and included "deprived" and mostly minority adolescents. Participants were young (in two studies age 12, and in two others 9-16). All students at baseline were non-users of alcohol and drugs. Two RCTs found mentoring reduced the rate of initiation of alcohol, and one of drug usage. The ability of the interventions to be effective was limited by the low rates of commencing alcohol and drug use during the intervention period in two studies (the use of marijuana in one study increased to 1% in the experimental and to 1.6% in the control group, and in another study drug usage rose to 6% in the experimental and 11% in the control group). However, in a third study there was scope for the intervention to have an effect as alcohol use rose to 19% in the experimental and 27% in the control group. The studies assessed structured programmes and not informal mentors.