CREST_Triage is a tool which helps editors decide whether to advance a systematic review submission to peer-review.
CREST_Triage is intended to help make decisions about advancing a SR manuscript to peer-review more transparent, and help the handling editor apply a consistent standard to each submission they receive. It does this by asking the handling editor to answer the same basic-but-comprehensive set of questions for each submission.
The answers to the questions are recorded and reproduced in a triage report, which the handling editor can share with the authors to show the reasoning behind their decision.
CREST_Triage is largely agnostic about the standard which the user applies to the submission - it is up to the editor to decide whether, for each question, the submission has done well enough to be advanced to peer-review.
This agnosticism is for three reasons: firstly, the standard is for the editor to decide; secondly, practices can be expected to shift over time, so what is good practice now may not be good practice in future; thirdly, what is appropriate for a given context or research question may be different to another.
Therefore, since a one-size-fits-all approach could stifle innovation in methods or force inappropriate methods on unanticipated review contexts, CREST_Triage only asks about the extent to which the user thinks the methods are acceptable. It prompts the handling editor to consider whether the methods measure up to whatever standard they consider appropriate for their journal.
Although different journals will likely have different requirements (and may even have different requirements for different types of SR submission), behind CREST_Triage there is an assumption that SRs should always do each of the following:
- Be based on a valuable research objective
- Run a thorough search for evidence
- Select all the relevant studies for analysis
- Critically appraise the included studies
- Synthesise the included studies into summary results
- Describe their confidence or certainty in the summary results
The tool assumes a one-stage peer-review process where a final SR manuscript has been submitted without peer-review of planned methods by the receiving journal. If a two-stage review process applies, then a stage-one or stage-two version of this tool should be used.
While a two-stage review version of the tool does not yet exist, a protocol can be assessed for the plan to conduct each of the above six criteria, and a stage-two submission can be assessed for compliance with planned methods, and validity of results and interpretation, on the above six criteria.
CREST_Triage is the work of Paul Whaley (words) and Marek Isalski (code), supported by a grant from Lancaster University. We were helped by many patient test users, reviewers and grad students — too many to acknowledge here, but who will be named in a forthcoming manuscript about the tool.