[Skip to Navigation]
Comment & Response
January 31, 2022

The Importance of Using the Appropriate Model for Systematic Reviews and Meta-analyses—Reply

Author Affiliations
  • 1Cardiovascular Division, John T. Milliken Department of Internal Medicine, Washington University School of Medicine, St Louis, Missouri
  • 2Center for Research in Health and Social Care Management, SDA Bocconi School of Management, Milan, Italy
  • 3College of Medicine and Health, University of Exeter, Exeter, UK
JAMA Intern Med. 2022;182(3):357-358. doi:10.1001/jamainternmed.2021.8132

In Reply We appreciate the comments by Dr Sanz-Sánchez and colleagues on our recent meta-analysis.1 Although most contemporary randomized clinical trials analyze the primary composite outcome using time-to-event methods, many of those trials also report the raw event rates of the individual components of the composite end points, including nonfatal myocardial infarction (MI) and all-cause mortality. In our analysis, we included only studies that either reported raw event rates of individual outcomes or provided those event rates to us by request. Thus, we respectfully disagree that our methods of end point ascertainment were a source of bias. We do agree that an individual patient-level meta-analysis would allow a more granular analysis of the data but acquiring that level of data for 144 studies spanning almost 40 years is neither feasible nor realistic.

Add or change institution