Dale 2007.

Methods	Setting: Midlands, England. Recruitment: Participants recruited through 43 general practices in the West Midlands. Randomisation: RCT 3 arm. This study explicitly drew on the definition of peer support by Dennis as "the provision of emotional, appraisal and informational assistance by a created social network member who possesses experiential knowledge of a specific behaviour or stressor and similar characteristics as the target population" (Dennis 2003a). Peer training: Peer supporters were trained at a two day training event on motivational interviewing and listening skills.
Participants	231 patients with a recent glycated haemoglobin greater than 7.4% and who had been advised of the benefit of reducing their glycated haemoglobin with or without a change in prescribed tablet based therapy.
Interventions	Intervention:Diabetes Specialist Nurses and peer telephone supporters attended a 2‐day training programme that focused on empowerment, motivational interviewing and active listening skills. For both intervention arms the first call was made 3 to 5 days after recruitment. Caller: Calls were made by peer supporters but timing was negotiated with the recipient and the number of calls were as often as the participant deemed necessary. Control group: No intervention. Participants were encouraged to follow medical advice given to them by their GP or practice nurse.
Outcomes	Physical health outcomes: Glycated haemoglobin (blood specimens), cholesterol, body mass index. Psychological health outcomes: Diabetes distress (Problem Areas in Diabetes Scale ‐ self report (validated)). Behavioural health outcomes: self‐efficacy (Diabetes Management Self‐Efficacy Scale ‐ self report (validated)). Social health outcomes: none reported. Impact on participants: Satisfaction (questionnaire developed for use in study ‐ self report (not validated)). Impact on participants: none reported. Impact on peer supporter: semi‐structured interviews (qualitative data) (not validated). Method of assessing outcome measures: paper questionnaire and clinical records.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Low risk	Table of random numbers.
Allocation concealment?	Low risk	A ‐ Adequate.
Blinding? Participants	High risk	Not done.
Blinding? Providers	High risk	Not done.
Blinding? Outcome assessors	Low risk	Reported by authors.
Blinding? Data analysts	Low risk	Reported by authors.
Incomplete outcome data addressed? All outcomes	Low risk	Intention‐to‐treat analysis.
Free of selective reporting?	Low risk	All outcomes reported.