Automated Evidence Surveillance with AI-Enabled Ranking in Living Evidence-Based Guideline Maintenance: A Simulation Study (Protocol)

This study aims to simulate and evaluate the performance and efficiency of the following evidence surveillance workflows in the context of living EBG development and maintenance:

i) Single database, topic-specific searches only

ii) Single database, overarching search only

iii) Single database overarching search with an overarching vector search

iv) Single database overarching search with topic-specific vector searches

The study will be conducted on a dataset constructed from the technical report of the International 2023 PCOS Guidelines, and prior work investigating the coverage of Medline / PubMed, Embase, OpenAlex and Semantic Scholar in indexing articles that were included in the 2023 edition of the International PCOS Guidelines.

Results of this study will be disseminated as a peer reviewed journal article, with code and data required to replicate analyses available as a publicly accessible Github repository.