The problem it solves
Screening takes a large share of a systematic review and is prone to fatigue, disagreement and error. The challenge is not only to decide quickly, but to show why each record was included, excluded or sent back for review.
What Vetra does
- Screens and classifies records with the full systematic review design in mind.
- Applies a first pass on title and abstract, and then on full text when the paper is available.
- Suggests a preliminary decision based on the criteria defined in the protocol.
- Generates a short justification and a confidence score linked to population, intervention, comparator, outcome, design or exclusion reason.
- Allows human reviewers to accept, modify or reject the suggestion with change tracking.
Human control and disagreements
Vetra always shows the confidence score behind each suggested inclusion or exclusion decision, together with the exact rationale based on the review design. When uncertainty is high or the system cannot decide with enough confidence, it leaves the paper flagged for human review, increasing the sensitivity of the process.
What is recorded
- AI-suggested decision, confidence score and final researcher decision.
- Reason for inclusion or exclusion and the affected criterion.
- Change history, responsible reviewer and date.
- Exportable data for PRISMA diagrams and process audit.