.jpg){kind=link}
- Published
- Updated
Finding hundreds of publications for a clinical evaluation does not explain how the search was performed. Another reviewer should be able to investigate the same question using the recorded method. Reproducibility helps assess scope and execution differences, but does not itself guarantee that all relevant evidence was identified.
Scope of the decision
The research question and database query are separate but connected records. Platform, execution date, full query, searched fields and active filters may affect retrieval. If the planned and executed strategies differ, preserve the change and its rationale. These are practical tools for examining the method.
Worked review
For example, a team reruns a query for a device family and obtains very different results. The earlier record contains only a keyword list. Review reveals that the original search used a title filter. The updated record retains both query and filters and examines which publications the scope difference affected.
Practical checks
- Retain the executed query together with its platform and date.
- Explain deviations from the planned method and their effect on results.
- Test whether another reviewer can understand an example search from the record.
Acceptance and limits
Acceptance means traceable steps and known limitations. A database update may prevent identical counts later, which is not automatically an error. It should still be possible to investigate whether the difference arose from changed methods or updated content.
Frequently asked questions
- Is a screenshot alone sufficient?
- Not if it omits the query or context; necessary method details should remain accessible.
- Do identical counts establish a good search?
- No. Relevance to the question and suitability of the strategy need separate assessment.