Potential mechanisms for change in diabetes prevention programs: A systematic review

To investigate potential mechanisms for change in diabetes prevention programs (DPPs), and assess the strength of associations.

Weight loss would be less strongly associated with improved health than other mechanisms. This unfunded study was pre-registered (CRD42020154362; https://www.crd.york.ac.uk/PROSPERO/view/CRD42020154362).

Empirical articles investigating psychosocial DPPs that investigated mechanisms, mediators or moderators (MMMs) of the relationship between the intervention and health were included. Retracted, qualitative or observational studies were excluded. Five biomedical databases were searched from inception until 9/2019 on 11/26/2019. The PEDro checklist assessed study quality. Outcome reporting bias was evaluated via comparing trial pre-registration to reports. We provide a narrative synthesis of the effects of MMMs.

Included were 108 studies (502,257 participants). Studies were heterogeneous. More studies did not support than demonstrated support for all MMMs. Because there was little evidence to support any MMM, our hypothesis was not tested; 30% of studies tested weight/adiposity, weight loss or weight loss intervention as MMMs. Of those, 79% either found no support or found that interventions were less effective/harmful for fatter people.

Few DPP studies investigate why these interventions work. There is no compelling evidence to support any specific MMM in DPPs, and no evidence to suggest that weight loss is what causes benefits. The included studies were of good methodological quality, but tests of mechanisms were rarely pre-registered; 46% of the included studies were drawn from three datasets.