Effects of parental alcohol rules on risky drinking and related problems in adolescence: Systematic review and meta-analysis

- Parents' rules about their children and alcohol may affect drinking in adolescence.
- We conducted a systematic review and meta-analysis of longitudinal studies.
- Setting rules about alcohol predicted less risky drinking in adolescence.
- The risk of bias in many studies was high and more rigorous research is needed.
- On current evidence, parents should be advised not to allow children to use alcohol.

BackgroundIt is unclear what effect parents' rules about their children's alcohol use have on drinking in adolescence. This review and meta-analysis investigated associations between prospectively measured parental alcohol rules and later adolescent risky drinking.MethodsUsing the PRISMA guidelines, we searched eight electronic databases for a variety of terms up to 10 September 2016. We imposed no restrictions on publication year. We assessed the risk of bias and conducted a meta-analysis.ResultsWe identified 13 eligible studies in four groups of specific exposures for meta-analysis. The pooled overall estimate showed that when parents set rules concerning alcohol, their children were less likely to develop risky drinking and related problems (OR = 0.64, 95% CI = 0.48, 0.86). Pooled estimates illustrate that parental alcohol rules were significantly negatively associated with adolescent risky drinking and related problems (OR = 0.73, 95% CI = 0.53, 0.99), as was parental approval of alcohol use (inverse OR = 0.41, 95% CI = 0.34, 0.50). Neither parental permissiveness (inverse OR = 0.83, 95% CI = 0.59, 1.19) nor parental disapproval of alcohol use (OR = 0.49, 95% CI = 0.20, 1.20) was significantly associated with alcohol-related problems. However, the small number of studies and variability in the point estimates in these latter two groups of studies limits inferences.ConclusionsParents' restrictiveness of their children's drinking was associated with lower risky drinking, but the risk of bias in the existing literature precludes strong inferences about the association. Further longitudinal studies with prospective measurement of parent behaviour, low attrition, and control for likely confounders, are needed.