Centre for Reviews and Dissemination
- Centre for Reviews and Dissemination
- News
- 2024
- SWAR machine learning
Prize winning Study Within a Review launched: Benefits and challenges of using machine learning technologies to prioritise screening in systematic reviews
Posted on 22 July 2024
Inspired by the unexpected benefits of using prioritised screening in a York Evidence Synthesis (YES) Group systematic review of ‘Multi-cancer early detection tests (MCED tests) for general population screening’, the SWAR aims to identify advantages and implications for the review process, including project management, and any challenges associated with the use of machine learning technologies in screening. The findings of the SWAR will inform which types of review question might benefit most from prioritised screening, and where prioritised screening might not provide tangible benefits over traditional manual screening methods.
SWAR lead CRD researcher, Sarah Nevitt, presented a poster and lightning talk describing the hidden benefits of prioritised screening observed by the YES group review team during the MCED tests review and the aims of the SWAR at the Research on Research online festival: AI and research: a promising relationship? The lightning talk was awarded the prize of the "Most important improvement in research practice," voted for by the festival's attendees.
A number of ongoing reviews across CRD—a review of clinical and cost-effectiveness of enzyme replacement therapies for late onset Pompe disease, scoping reviews relating to health inequalities and ageing, and a scoping review of the health of care experienced people—are using machine learning technologies to prioritise screening. Systematic review teams are recording their experiences, including any advantages, challenges and implications for the review process to inform the SWAR. Going forward, we hope to further expand the range of systematic review types, review topics and review teams included within the SWAR to better understand the role of machine learning technologies to prioritise screening in systematic reviews.