Twelve studies were included.
Instructional programmes were assessed in 8 studies: 4 randomised controlled trials (RCTs) (3,115 individuals plus 87 classes), one non-randomised study (n=664), one before-and-after study with concurrent control (n=322), one time series with concurrent control (n=2,376), and one before-and-after study with no control (n=60).
Peer organisation programmes were assessed in 2 studies: one time series with concurrent control (4 schools) and one post intervention only study with concurrent control (n=17,187).
Social norming was assessed in 2 studies: one RCT (n=160) and one before-and-after study with no control (n at least 1,786).
In terms of study quality, the attrition rates ranged from zero for short-term follow-up to nearly two thirds.
Instructional programmes (8 studies).
The results for DD were mixed. The median change in self-reported DD (5 studies) was -0.10 SD (range: -0.22 to 0.04). Three of the 4 studies assessing RDD found a statistically significant benefit of the intervention. The median change in self-reported RDD (4 studies) was -0.18 SD (range: -0.72 to -0.10). The only study assessing moving violations and crashes found no significant difference between the intervention and control; the relative risk from 1 to 6 years ranged from 0.92 to 1.14.
Peer organisation programmes (2 studies).
One quasi-experimental time series with a concurrent control group found no significant difference for combined DD and RDD between a Students Against Destructive Decisions (SADD) programme and a control group. Although rated as "greatest" design quality, the programme was not fully implemented. The other study, classified as "least suitable" for design quality and with low power to detect an effect found non-statistically significant improvements in outcomes in SADD schools.
Social norming programmes (2 studies).
One before-and-after survey with breath alcohol tests (n=1,786) found that a campus-wide campaign appeared to reduce alcohol consumption. The other study found that groups exposed to a peer theatre intervention reported a statistically significant increased use of a designated driver and decreased DD compared with unexposed controls.