BackgroundWorldwide, there are an estimated 15 million individuals with drug use disorders and over five times as many with alcohol use disorders. Most individuals with substance misuse have family who are affected, regardless of whether they remain close and provide caring roles or have become estranged. Initial scoping searches identified an expanse of broad and disparate studies and reviews on the topic. We did a systematic review of reviews to bring together the expanse of research on the effectiveness of family-based interventions in substance misuse. Evidence from the highest quality systematic reviews was synthesised to inform future research and evidence-based commissioning. MethodsExtensive electronic (inception to April, 2012) and manual (April, 2012) citation searches were undertaken. Medline, Embase, CINAHL, Cochrane Library, PsycINFO, IBSS, Campbell Collaboration, and CPSI-S were searched via two parallel searches: one alcohol and one drug specific. Screening, data extraction, and quality appraisal were undertaken by two reviewers, with disagreements resolved through discussion. The inclusion criteria were that the study be a systematically undertaken review, the population be individuals with substance misuse problems, and the interventions include a family-focused component. Studies that focused on prevention rather than treatment were excluded. Quality assessment of the identified reviews was undertaken using the AMSTAR framework, and those fulfilling at least half the quality criteria were deemed to be of sufficiently high quality to draw evidence from. The level of evidence available supporting the interventions was categorised into three areas: (1) working with family members to promote user entry into treatment; (2) joint involvement of user and family members; and (3) interventions for family members in their own right. Evidence was classed as sufficient if the effectiveness was consistently supported across reviews, tentative if the evidence was consistent across less robust studies, and no evidence if no results were available in that area. Because there was substantial heterogeneity between systematic reviews, a meta-analysis was inappropriate. FindingsThe search identified 1065 alcohol-related and 2977 drug-related reviews. After removal of duplicates and title and abstract screening, 83 alcohol-related and 214 drug-related full text articles were reviewed. After removal of overlap between both searches, 25 reviews were quality assessed. Evidence was drawn from the eight systematic reviews judged to be of sufficiently high quality. Robust evidence for effectiveness of joint involvement was found for family-based interventions for adult alcohol misusers for both substance-related and relationship outcomes (three reviews) and tentative evidence for behavioural couples therapy in adults who misused drugs (two reviews). However, no review-level evidence supported or refuted the effectiveness of family interventions to promote engagement into treatment of adolescents who misuse substances or adults who misuse drugs. Responding to the needs of the family members of substance misusers in their own right was found in only one review. No review-level evidence for joint involvement of family in the treatment of adolescents who misuse substances was found, despite six reviews in this area. However, positive evidence was found in support of family interventions for engagement of users in treatment for adults who misuse alcohol and for multidimensional family therapy for improvement of substance-use-related outcomes for adolescents. InterpretationEvidence exists for family-based interventions in improvement of substance-use-related outcomes and some relationship-based outcomes. However, the conclusions that can be drawn from the existing research are limited owing to gaps in the evidence base. An absence of evidence does not mean an absence of effect. Weaker study types identified during the research do show encouraging results for family interventions across all areas; however, more high-quality reviews are needed to quantify these results with certainty. FundingNone.