BackgroundMass media campaigns can be used to communicate public health messages at the population level. Although previous research has shown that they can influence health behaviours in some contexts, there have been few attempts to synthesise evidence across multiple health behaviours.ObjectivesTo (1) review evidence on the effective use of mass media in six health topic areas (alcohol, diet, illicit drugs, physical activity, sexual and reproductive health and tobacco), (2) examine whether or not effectiveness varies with different target populations, (3) identify characteristics of mass media campaigns associated with effectiveness and (4) identify key research gaps.DesignThe study comprised (1) a systematic review of reviews, (2) a review of primary studies examining alcohol mass media campaigns, (3) a review of cost-effectiveness evidence and (4) a review of recent primary studies of mass media campaigns conducted in the UK. A logic model was developed to inform the reviews. Public engagement activities were conducted with policy, practitioner and academic stakeholders and with young people.ResultsThe amount and strength of evidence varies across the six topics, and there was little evidence regarding diet campaigns. There was moderate evidence that mass media campaigns can reduce sedentary behaviour and influence sexual health-related behaviours and treatment-seeking behaviours (e.g. use of smoking quitlines and sexual health services). The impact on tobacco use and physical activity was mixed, there was limited evidence of impact on alcohol use and there was no impact on illicit drug behaviours. Mass media campaigns were found to increase knowledge and awareness across several topics, and to influence intentions regarding physical activity and smoking. Tobacco and illicit drug campaigns appeared to be more effective for young people and children but there was no or inconsistent evidence regarding effectiveness by sex, ethnicity or socioeconomic status. There was moderate evidence that tobacco mass media campaigns are cost-effective, but there was weak or limited evidence in other topic areas. Although there was limited evidence on characteristics associated with effectiveness, longer or greater intensity campaigns were found to be more effective, and messages were important, with positive and negative messages and social norms messages affecting smoking behaviour. The evidence suggested that targeting messages to target audiences can be effective. There was little evidence regarding the role that theory or media channels may play in campaign effectiveness, and also limited evidence on new media.LimitationsStatistical synthesis was not possible owing to considerable heterogeneity across reviews and studies. The focus on review-level evidence limited our ability to examine intervention characteristics in detail.ConclusionsOverall, the evidence is mixed but suggests that (1) campaigns can reduce sedentary behaviour, improve sexual health and contribute to smoking cessation, (2) tobacco control campaigns can be cost-effective, (3) longer and more intensive campaigns are likely to be more effective and (4) message design and targeting campaigns to particular population groups can be effective.Future workFuture work could fill evidence gaps regarding diet mass media campaigns and new-media campaigns, examine cost-effectiveness in areas other than tobacco and explore the specific contribution of mass media campaigns to multicomponent interventions and how local, regional and national campaigns can work together.Study registrationThis study is registered as PROSPERO CRD42015029205 and PROSPERO CRD42017054999.FundingThe National Institute for Health Research Public Health Research programme.