OBJECTIVES: The objective of this systematic review was to evaluate the effectiveness of population health interventions to prevent type II diabetes.
METHODS: Building upon a comprehensive review of the evidence to 2002 compiled by a team from the University of Waterloo, we searched eight electronic databases from 2003]2006 and the reference lists of all retrieved articles for relevant studies. Criteria for inclusion were: 1) article describes an intervention designed for the primary prevention of type II diabetes or its related risk factors; 2) the intervention is implemented 'universally' rather than targeting people at high risk; and 3) the intervention addresses one or more of the social determinants of health (e.g., changes to the physical environment). Screening and quality assessment were done independently by two reviewers. Standardized tools were used to rate studies for internal validity. Discrepancies were resolved through discussion and a third reviewer if necessary.
RESULTS: The University of Waterloo had identified 18 interventions meeting our criteria. Our search of the literature identified a further 29 new interventions. Eight of these have yet to report outcome data, and so have been flagged for future review, making 47 population health interventions in total. Of these 23 took place in schools, 10 in communities and 6 in worksites. All but four studies showed a significant impact on at least one risk factor relevant to diabetes. The factors that are affected vary among studies, and the effects are often small. The results show little consistency or coherency. None report changes in health state or incidence of disease. Many of the studies are short term and prone to other forms of bias.
CONCLUSIONS: The quality of the studies continues to raise concerns. No one is measuring changes in the incidence of diabetes or quality of life. Many are relying on self]reported measures of behaviours, and these measures are prone to bias. Follow up periods are often very short. Reporting of the research and the implementation of the interventions is often lacking, making it difficult to assess internal and external validity. That said, the weight of evidence suggests that population health approaches in schools and communities can reduce risk factors associated with diabetes. However, there is no obvious pattern in the relationship between what is done in each of the interventions and what is achieved. Thus, there is no simple, compelling policy recommendation possible at the moment. A realist review of each of the interventions found to be effective here might shed more light on why some interventions work in some ways for some people, and provide better insight into what might be done to reduce the risk of diabetes.