Xuxi Zhang

intervention (Fig.1). Two reviewers (Zhang and Yang) independently examined the full ‐ text of the remaining 77 articles to identify eligible ones. Of the 77 articles, 56 were excluded because basic information and outcomes that could be used to evaluate the effect of glycemic control were missing (Fig.1). Among the remaining 21 articles, 20 13 ‐ 32 were included in our meta ‐ analysis study. One study 33 identified through Pubmed was excluded because the posttest results (mean of HbA 1c ) of control group (CG) and intervention group (IG) were not accurate as the dropouts of IG were more than 20% without statistical tests to assess representativeness of those retained, and the attrition rate of IG was significantly higher than that of CG. Figure 1 Flowchart of Study Selection Data extraction and quality assessment Data from the 20 eligible studies were extracted by one of the two reviewers (Zhang) with a standardized data extraction form and all the data extracted from the eligible studies were checked by another reviewer (Yang). The data extraction included: the first author’s name, the year of publication, number of participants (including details of IG and CG at baseline and follow ‐ up intervals separately, attrition and completion rate), participant characteristics (including types of patients, mean age, gender ratio, insulin usage and other features), study design, study location (country), intervention method, intervention duration, follow ‐ up intervals and data of hemoglobin A1c (HbA 1c ) values. Moreover, we numbered the eligible studies from 1 to 20. 6 131 Peer support among adults with type 2 diabetes