Practical methods for incorporating summary time-to-event data into meta-analysis: updated guidance

Abstract

Abstract Background Our prior guide to estimating hazard ratios (HRs) and related statistics from published summary (aggregate) data (ref) has been cited widely and utilised extensively by systematic reviewers and meta-analysts worldwide, but some still have difficulties applying the methods and/or utilising the accompanying spreadsheet. Therefore, we provide additional guidance and an updated calculations spreadsheet.Methods This update is driven by queries we have received, responses to a survey of Cochrane editors and the emergence of alternative methodologies. It also and benefits from our considerably increased experience in applying these methods across a range of scenarios.Results We provide clarifications of ambiguities, additional scenarios that researchers might face and further advice on data extraction to facilitate estimating HRs and related statistics from published time-to-event data. Also, we compare these methods with proposed alternatives. This is accompanied up by an updated and enhanced calculations spreadsheet that produces all the summary statistics required for pooling HRs in meta-analysis.Conclusions This updated guidance and spreadsheet will provide valuable additional tools for those producing meta-analyses based on published, summary, time-to-event data.