Recommendations for research

Research councils, national and local research commissioners and funders, research workers and journal editors.

Include as standard in research reports:
- a description of what was delivered, over what period, to whom and in what setting
- information on the impact on health
- clear definitions of the 'health outcomes' measured
- a report of differences in access, recruitment, and (where relevant data are available) uptake, according to socio-economic and cultural variables such as social class, education, gender, income or ethnicity
- a description and rationale of the research methods and forms of interpretation used, and where relevant the reliability and validity of the measures of behaviour change adopted.
Ensure research studies on behaviour change always:
- identify and account for the different components of change among different social groups
- pay attention to minority ethnic and religious groups
- include social variables wherever possible (for example, social class or education) in every study
- consider the impact of age and gender on the effectiveness of interventions and programmes.
Promote the inclusion of process as well as outcome data.
Encourage those in charge of randomised controlled trials on health-related behaviour change to register with a trial register.

Research commissioners and funders.

Encourage research that takes into account the social and cultural contexts in which people adapt or change their behaviour and the factors that encourage or inhibit change. These include:
- the role of support networks, neighbourhood resources and community action
- the relationships that help protect and build people's resilience
- the way people adapt positively to adverse socio-structural conditions
- social processes that strengthen the mutual support provided by families and other forms of households
- the clustering of health behaviours
- the material circumstances in which people live, including income levels, environmental characteristics of neighbourhoods and work-related factors.
Use embedded process evaluations that include the perspectives of recipients.
When studying the mechanisms of adaptation and change, use mixed method ethnographic research, longitudinal studies and qualitative approaches, as well as multivariate and interactive statistical models.
Support development of new methods for collating and synthesising a range of evidence on effectiveness. These methods should meet the highest scientific standards.

Policy makers, research commissioners and local service providers.

Collect baseline data at the outset of interventions or policy changes and allow for an adequate length of time for evaluation.
Develop evaluative approaches which can accommodate the complexities inherent in community and population-level interventions or programmes, including multiple and confounding factors.
Develop methods for synthesising and interpreting results across studies conducted in different localities, policy environments and population groups.
Formulate rigorous and transparent methods for assessing external validity and for translating evidence into practice.

Policy makers, research funders and health economists.

As a matter of urgency, commission research on the cost-effectiveness of behaviour change interventions. This should cover:

interventions over the mid to long term
interventions aimed at specific population groups (for example, low-income groups, men versus women, young people versus older people)
primary prevention versus clinical treatment for behaviour-related disease.

More detail on the evidence gaps identified during the development of this guidance is provided in appendix D.

Behaviour change: general approaches