Challenges of coordinating large-scale systematic reviews

Session Type
Oral presentation
Capacity building in evidence synthesis
Yogasingam S1, Lepine J2, Sullivan K3, Antony J4, Konnyu K5, Ivers N6, Grimshaw J1
1Ottawa Hospital Research Institute, Canada
2University of Laval, Canada
3Canadian Institute for Health Information, Canada
4Ontario Health, Canada
5Brown University, United States of America
6University of Toronto, Canada

Background: Large scale systematic reviews are becoming more common due the rapid growth in publications on certain topics. They raise unique challenges for managing and coordinating the review process for which there is limited guidance. We will present the challenges we encountered and how we addressed them during two large Cochrane systematic reviews - Quality improvement strategies for diabetes care (n=553 trials) and Audit and feedback (n=290 trials).
Objectives: To address the logistical challenges that are common in the coordination of large systematic reviews.
Methods: We will compare the approaches to coordinating two large systematic reviews. We will focus on common logistical challenges such as recruiting and maintaining a team with enough availability to complete the project in a timely manner; delegating work and ensuring consistency between reviewers on a large team; and managing and housing a very large dataset. We will also compare tools for managing large reviews such as the systematic review programs DistillerSR and Covidence for usability and customisability as well as online repositories for external data storage.
Results: The main challenge in coordinating large systematic reviews is dedicating enough resources to ensure the review can be completed before another update is required. Having a large team can be useful for keeping momentum going in screening and extraction, however it is vital to ensure that reviewers are being consistent by using clear and specific instructions and having a training period for each reviewer. Smaller teams are better for ensuring consistency in reporting, however screening and extraction takes longer. Use of technology such as DistillerSR and Covidence, facilitates the organisation of screening a large set of studies. DistillerSR offers more customisability but may be cost prohibitive. Covidence links nicely with RevMan Web, but may not be as user friendly. The use of an online data repository such as Zenodo is useful for streamlining the main review and preventing any issues with the Cochrane CENTRAL system when it comes time for publication, however there is always a risk of repositories no longer being in service in the future.