Xuxi Zhang

Of the 1991 articles, on the basis of the study titles and abstracts, two reviewers (Yusufu and Zhang) excluded 1903 articles that: were not RCTs; were not original research (e.g. reviews, secondhand ‐ data analysis, and design studies); studied type 1 diabetes, gestational diabetes or other specific types of diabetes; studied patients under 18 years old; did not study the effect of interventions targeting modifiable risk factors of DR on the prevention and/or control of DR, adopted medical treatment of DR as interventions (e.g. drugs or medicines, medical examinations, or surgeries); or were not published in English. Two reviewers (Yusufu and Zhang) independently examined the full ‐ text of the remaining 88 articles. Among those, 72 were excluded mainly due to lack of basic data that would be needed to evaluate the effect of intervention on the prevention and control of DR (Figure 1). In case of disagreement, the reviewers discussed with a third researcher (Sun) to reach an agreement and all disagreements were resolved by consensus. Finally, 16 articles 5 ‐ 8, 10 ‐ 21 on 22 studies were included in this meta ‐ analysis. Figure 1 Flowchart of study selection 7 165 Prevention and control of diabetic retinopathy