SUGGESTED WORKFLOW
Screen a review
Prioritize reading while keeping inclusion decisions and stopping rules auditable.
An editorial workflow to adapt, not a benchmark reproduced by this site.
Prepare
A deduplicated export of titles and abstracts, eligibility criteria, and a small set of known relevant records.
Set the protocol
Write inclusion criteria and the stopping policy. Export the search query, database, date, and deduplication method alongside the records.
Pilot with known records
Label a small batch by hand. Compare an ASReview prioritization run with your manual decisions; Rayyan is an alternative interface for collaborative screening. Confirm current plans before choosing.
Review the misses
Inspect disagreements, missed seed papers, and excluded records. Keep the original labels when revising the protocol and explain changes.
Export an audit trail
Save decisions, dates, versions, and the stopping rationale. If using SYNERGY to practice, treat its datasets as training exercises rather than proof of performance in your topic.
Expected output
An inclusion/exclusion log, model and software versions, a documented stopping decision, and a reproducible screening export.
Check your result
Known relevant records are recovered; disagreements and a sample of excluded records are reviewed; the stopping rule is recorded before applying it. Set recall targets for your review protocol rather than borrowing a universal percentage.
Limits
This is a suggested workflow, not a benchmark run by this site. AI prioritization does not replace a defensible search strategy or human eligibility decisions.