Grading evidence and recommendations for public health interventions: developing and piloting a framework

The objective of this work was to develop a practical scale of grades of recommendation for public health interventions, adapted from the current National Institute for Clinical Excellence (NICE) methodology. A literature review was carried out on the subject of incorporating research evidence into grades of recommendation for public health interventions. The literature search looked at publications from January 2000– May 2004 retrieved from 16 databases. The views of a range of public health experts were also sought for suggestions of other publications to be included in the literature review, and for their comments at various stages of the developing methodology. The principles for development of the framework were that it should be: • Adapted from, and clearly linked to, the current NICE methodology • Based on detailed and transparent reporting and synthesis of all relevant supporting evidence (intervention and observation; quantitative and qualitative). The literature review indicated general agreement that the randomised controlled trial (RCT) has the highest internal validity and, where feasible, is the research design of choice when evaluating effectiveness. However, many commentators felt the RCT may be too restrictive for some public health interventions, particularly community based programmes. In addition, supplementing data from quantitative studies with the results of qualitative research is regarded as key to the successful replication and ultimate effectiveness of interventions. Based on the literature review and consultation with experts, a framework was developed that derives grades of recommendation, incorporating: • Strength of evidence of effi cacy based on the research design and the quality and quantity of evidence (the current NICE system) • Corroborative evidence (from observational and qualitative studies) for the feasibility and likelihood of success of an intervention if implemented in the UK. The precise methods for combining the results from different types of corroborative evidence and for incorporating the size of effects, including (cost–)benefi ts and harms for the different outcomes measured, are still in development. This provisional framework provides a practical and transparent method for deriving grades of recommendation for public health interventions, based on a synthesis of all relevant supporting evidence from research. The methodology is being piloted, alongside the current NICE methodology, within the development of the public health/ prevention aspects of the HDA/NICE guidance on overweight and obesity