Background Multiple risk behaviour (MRB) refers to two or more risk behaviours including smoking, drinking alcohol, poor diet and unsafe sex. Such behaviours are known to co-occur in adolescence.There are increasing public health interventions that address MRB as opposed to isolated behaviours. However, little is known about differential intervention effects by socioeconomic status (SES). There is a need to examine these effects in order to reduce health inequalities. The aim of this study was to examine universal public health interventions targeting adolescent multiple risk behaviour forsubgroup effects by SES.

Methods Two Cochrane systematic reviews that focused on adolescent MRB were screened to identify universal interventions that reported SES. Study authors were contacted, and outcome data requested stratified by SES and intervention status. Risk behaviour outcomes: alcohol use, smoking, substance use, unsafe sex, overweight/obesity, sedentarism, peer violence and dating violence were examined in random effects meta-analyses and subgroup analyses performed to explore differences between high SES and low SES adolescents.

Results Of 50 studies reporting universal interventions, 15 also reported having measured SES. Of these 15 studies, four study authors provided additional data for subgroup analyses. For alcohol use, the point estimates suggest that SES does not explain the effect of the intervention, as the direction of effect is the same for both high SES (RR 1.28, 95% CI 0.97, 1.69) and low SES (RR 1.14, 95% CI 0.97, 1.34). The point estimates for smoking behaviour are indicative of a differential intervention effect in favour of the low SES group (RR 0.83, 95% CI 0.66, 1.03) versus the high SES group (RR 1.10, 95% CI 0.78, 1.56). SES was not an explanatory factor for the intervention effect on substance use as the direction of effect in the high SES group (RR 1.13, 95% CI 0.83, 1.53) and the low SES group (RR 1.26, 95% CI 0.83, 1.92) was the same. Tests for subgroup differences showed no evidence of difference for all behaviour outcomes.

Conclusion The majority of studies identified did not report having measured SES. Findings from the four studies included in the subgroup analysis indicate the potential for interventions to differentially effect different SES groups. There is a need for routine reporting of demographic information within studies so that stronger evidence of effect by SES can be demonstrated and that interventions can be evaluated for their impact on health inequalities.