Table 1

Overview of recommendations for RR conduct

General recommendations on team characteristics and organisationEnsure that the team has sufficient SR experience and that teams are a manageable size. Use supportive software and plan review steps to ensure an efficient workflow.
Employ piloting exercises to allow team members involved in a certain task (eg, study selection) to test the tools and processes of this task on a small proportion of records to ensure that all team members perform the task consistently and correctly.
Recommendations on study selection (title/abstract and full-text level)

  • Reduce the number of human judgements involved:

    • conduct dual and independent screening of a proportion of records (eg, 20%) and assess reviewer agreement. If agreement is good (eg, 80%), proceed with single-reviewer screening.

    • Enhance validity of single-screening based on types and numbers of exclusion reasons.

  • Use supportive software.

  • Consider semiautomation in the form of crowdsourcing and/or machine learning.

Recommendations on data extraction

  • Have one person extract the data, with a second person verifying the data for accuracy and completeness.

  • Limit data extraction to only the most important data fields relevant to address the RR question (as agreed on in the protocol).

  • Where available, extract primary study data directly from existing SRs.

  • Use supportive software.

Recommendations on RoB assessment

  • Use validated and study design–specific tools to assess the RoB of the included studies.

  • Limit the RoB assessment to only the most important outcomes (for RoB tools with outcome-specific questions).

  • Have one person perform the RoB assessment and have a second reviewer verify the judgements.

  • Complete omission of RoB assessment is discouraged, as this information informs the interpretation of the evidence and review implications.

  • RoB, risk of bias; RR, rapid review; SR, systematic review.