Mann 2016.

Study characteristics
Methods	Cluster RCT
Participants	Two urban academic primary care practices, New York, USA 49 patients
Interventions	Guided physicians through counselling patients with prediabetes on lifestyle modifications
Outcomes	Clinical endpoint (various outcomes pertaining to weight, body mass index, HbA1C, lipids)
Co‐Interventions	Educational: Distribution of educational material to patients in control group; single educational session for providers in intervention group Beyond Clinician Education: Distribution of pedometers to patients in intervention group; audit and feedback for providers in intervention group
CDSS Features ‐ Acknowledgement of CDSS Required	No
CDSS Features ‐ Other	Conveyed patient‐specific information, decision support was complex, developed by study investigators, developed in consultation with users, makes care recommendation, possible to execute desired action, 'pull' mode of delivery, targeted underuse, user workflow considered in design
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Low risk
Incomplete outcome data (attrition bias) All outcomes	Low risk
Baseline characteristics similar?	Low risk
Unit of analysis error	Low risk