Background
Health risk behaviours (such as unhealthy diet, physical inactivity and smoking) are common and there is good evidence that they tend to cluster and co-occur. Reference Meader, King, Moe-Byrne, Wright, Graham and Petticrew1–Reference Noble, Paul, Turon and Oldmeadow2 For example, 68% of adults in England and 52% in the USA were found to engage in two or more health risk behaviours. Unhealthy diet is particularly common. For example, in the UK a national survey found that 79% of men and 73% of women ate less than five portions of fruit or vegetables per day. Reference Poortinga3 Unhealthy behaviour accounts for 28% of all health burden (measured in disability-adjusted life-years) and unhealthy diet was the largest contributor of all behaviour risks to health burden. Reference Newton, Briggs, Murray, Dicker, Foreman and Wang4 People with severe mental illness (SMI) are more likely than the general population to engage in health risk behaviours. Reference Dipasquale, Pariante, Dazzan, Aguglia, McGuire and Mondelli5 In addition, antipsychotic medication is associated with increased hunger and unhealthy eating. Reference Treuer, Hoffmann, Chen, Irimia, Ocampo and Wang6 Given the high prevalence and health burden of risk behaviours in the general population, improving diet in people with SMI is an important priority.
Meta-analyses of interventions to reduce health risk behaviours in people with SMI
Health risk behaviour interventions in people with SMI are still in a relatively early stage of development. For example, a recent meta-analysis on the effectiveness of physical activity interventions found limited data and the need for further research. Reference Pearsall, Smith, Pelosi and Geddes7 Teasdale et al's systematic review in this issue shows a promising literature on the effectiveness of nutrition interventions in people with SMI. Weight loss (2.7 kg compared with controls) following nutrition intervention in people with SMI compares well with weight-loss interventions in other populations. Reference Teasdale, Ward, Rosenbaum, Samaras and Stubbs8 For example, a recent meta-analysis in overweight and obese adults found a weight loss of 2.8 kg compared with controls. Reference Hartmann-Boyce, Johns, Jebb and Aveyard9 Teasdale et al also found statistically significant reductions in BMI, waist circumference and blood glucose levels in intervention participants compared with controls. Reference Teasdale, Ward, Rosenbaum, Samaras and Stubbs8 In addition, they found that dietitian-led interventions and those delivered at antipsychotic initiation were associated with more improved outcomes.
Future research
Replication in larger trials and the need for further exploration of intervention content
Most trials included in the meta-analysis by Teasdale et al were relatively small and there remain uncertainties regarding risk of bias, with almost half not reporting allocation concealment and more than a third not reporting masking of outcome assessors. Reference Teasdale, Ward, Rosenbaum, Samaras and Stubbs8 There was also potential for publication bias identified in the meta-analysis. Therefore, it is currently unclear whether these promising findings will be replicated in large well-conducted randomised controlled trials (RCTs). A further limitation is the lack of data on diet outcomes. To evaluate the benefits of nutrition interventions there is a need to be able to show the extent to which they improve nutrition, which is unclear based on current data. Future trials should report diet outcomes more transparently and consistently across trials. In addition, there is great variability in the content of nutrition interventions and therefore there is a need to identify whether particular types of behaviour change techniques are associated with improved outcomes using validated taxonomies.
Multiple risk behaviour interventions
A further area of investigation that has received less attention is interventions for multiple risk behaviours in people with SMI. As noted above, people with unhealthy diets are more likely to also be physically inactive or to smoke. Therefore, change in one risk behaviour (such as diet) has potential to have an impact on other risk behaviours (such as physical activity or smoking). Further investigation should also examine whether there are additional gains in targeting diet and physical activity together in order to increase weight loss or if making sequential changes in behaviours is a more effective long-term strategy.
Conclusions
In summary, there are promising data on the potential effectiveness of nutrition interventions for improving weight management in people with SMI. Current evidence suggests such interventions appear as effective in people with SMI as in other populations such as people with obesity. Therefore, nutrition interventions have great potential to improve the lives of people with SMI and may reduce their risk of chronic disease. Future research should seek to replicate these findings in larger well- conducted RCTs. There are also a number of further questions to explore including whether particular intervention content and other components of intervention delivery are associated with improved outcomes. In addition, whether interventions for unhealthy diet should be considered in isolation or within the context of other health risk behaviours.
eLetters
No eLetters have been published for this article.