Back

Assessment of concordance between related systematic reviews and between related guideline recommendations: protocol for a methodological survey.

Agarwal, A.; Albarqouni, L.; Badran, N.; Brax, N.; Gandhi, P.; Pereira, T.; Roberts, A.; El Zein, O.; Akl, E.

2022-07-12 epidemiology
10.1101/2022.07.11.22277498 medRxiv
Show abstract

Independent systematic reviewers may arrive at different conclusions when analyzing evidence addressing the same clinical questions. Similarly, independent expert panels may arrive at different recommendations addressing the same clinical topics. When faced with a multiplicity of reviews or guidelines on a given topic, users are likely to benefit from a structured approach to evaluate concordance, and to explain discordant findings and recommendations. This protocol proposes a methodological survey to evaluate the prevalence of concordance between reviews addressing similar clinical questions, and between clinical practice guidelines addressing similar topics; and to identify methodological frameworks for the evaluation of concordance between related reviews and between related guidelines.

Matching journals

The top 3 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.