A proposed framework to guide evidence synthesis practice for meta-analysis with zero-events studies

ObjectiveIn evidence synthesis practice, researchers often face the problem of how to deal with zero-events. Inappropriately dealing zero-events studies may lead research waste and mislead healthcare practice. We propose a framework guide better in meta-analysis.Study design settingWe used two dimensions, one respect total events count across all comparative arms meta-analysis, second whether included have single or both zero-events, establish for classification meta-analysis studies. A dataset from Cochrane systematic reviews was evaluate classification.ResultsThe proposed classifies into six subtypes. The matched well large real-world dataset. applicability existing methods were then presented under each subtype based on this framework, 5-step principle help practice.ConclusionsThe should be considered by when making decisions selection meta-analysis. It also provides reasonable basis development methodological guidelines