Estimating data from figures with a Web-based program: Considerations for a systematic review

摘要: Background Systematic reviewers often encounter incomplete or missing data, and the information desired may be difficult to obtain from a study author. Thus, systematic have resort estimating data figures with little no raw in study's corresponding text tables. Methods We discuss case which participants used publically available Web-based program, called webplotdigitizer, estimate 2 figures. We evaluated intraclass coefficient accuracy of estimates true inform considerations when using estimated reviews. Results The for both were consistent, although distribution figure continuous outcome was slightly higher. For outcome, percent difference ranged 0.23% 30.35% while event rate 0.22% 8.92%. figures, excellent (>0.95). Conclusions Systematic should consider transparent cannot obtained authors perform sensitivity analyses pooled results reduce bias.