Background Risk behaviours (e.g. unhealthy diet, lack of physical activity (PA)) are associated with increased risk of chronic disease and mortality. The majority of adults in England engage in two or more risk behaviours, which tend to co-occur in individuals. There are also strong socio-economic gradients in key health behaviours. A systematic review of interventions targeting multiple risk behaviours (MRBs) in general adult populations found small improvements in diet, PA and smoking. In this systematic review, we aimed to evaluate the effectiveness of MRB interventions in overweight or obese adults (PROSPERO CRD42016051589).

Methods Six databases were searched up to October 2016 with no language restrictions, and references of relevant systematic reviews were screened. Randomised controlled trials (RCTs) of interventions targeting two or more risk behaviours in obese or overweight adults (≥16 years) were eligible if they reported behavioural outcomes. Quality of RCTs was assessed with the Cochrane risk of bias tool. Where possible, random-effects meta-analyses were conducted. Data not included in the meta-analyses were narratively synthesised. Associations between study factors and effectiveness were explored through meta-regression.

Results In total 4392 records were identified, 219 full texts screened and 51 RCTs included. All studies targeted diet and PA, with weight management the most common objective. This contrasts with our review in general adult populations which found smoking and alcohol were often targeted alongside diet and PA and the most common aim was healthy lifestyle promotion. Few interventions included components other than education, training and enablement. There were small changes in fruit and vegetable (MD 27.88 grams, 95% CI 13.01 to 42.76, I2=51.5%), fat (SMD -0.33, 95% CI -0.51 to -0.15, I2=87.7%) and calorie intake (MD -282.89, 95% CI -426.01 to -139.78, I2=91.6) and total PA (SMD 0.21, 95% CI 0.01 to 0.41, I2=87.2%) at the end of interventions. Equivocal benefits were identified for other behavioural outcomes. Generally improvements declined at follow-up and when compared with an active control. We found no associations between study factors and effectiveness at the end of interventions.

Conclusion Interventions were effective for some behavioural outcomes in obese and overweight adults, but improvements were small and declined over time. Studies were limited in their approach, with most adopting ‘agentic’ strategies which make substantial demands on individuals to initiate and sustain behaviour change. Research is needed into alternative approaches which target the environmental conditions that influence health behaviours and make healthier choices easier for individuals.