Thirty-two studies (13 RCTs, 16 CCTs and three cluster RCTs) with 13,605 participants were included. Of the included trials, one involved homeless people, two involved indigenous populations, one involved prisoners, six involved at-risk youth, 12 involved low-income populations and ten involved people with a mental illness. Two trials were rated methodologically strong, 10 trials as moderate and 20 trials as weak. Follow-up ranged from four to six weeks to four years
Statistically significant beneficial effects of behavioural interventions were seen for low income women at short-term follow-up (RR 1.68, 95% CI 1.21 to 2.33; three trials) and for people with mental illness at long-term follow-up (RR 1.35, 95% CI 1.01 to 1.81; seven trials). Heterogeneity was low in these analyses (I2=0%).
Other meta-analyses showed statistically non-significant differences between behavioural intervention and control groups.