Augmentation of a multidisciplinary team meeting with a clinical decision support system to triage breast cancer patients in the United Kingdom

Abstract

Aim: Multidisciplinary team (MDT) meetings struggle with increasing caseloads. Recent National Health Service (NHS) guidance proposes that patients are triaged for ‘no discussion at MDT’. We examine whether an artificial intelligence (AI)-based clinical decision-support system (CDSS) can support human triage. Methods: Local best practice breast cancer MDT treatment decisions were compared with treatment decisions made by: two, two-person MDT triage teams with and without the CDSS; the CDSS acting ‘alone’; and the historical MDT. A decision tree on whether to triage patients to the CDSS or the MDT was created using supervised learning algorithms. Results: When localized, the CDSS achieved high concordance with local best practice (treatment plan decisions: 92% CDSS vs 96% team 1 vs 92% team 2, not significant [NS]; treatment type decisions: 89% CDSS vs 93% team 1 vs 82% team 2, NS). Using a decision tree 40.2% of cases can be correctly triaged to the CDSS for a treatment plan, and 34.6% for treatment type recommendations. Conclusion: AI-enabled CDSSs can potentially reduce the clinical workload for a breast cancer MDT by up to 40%. Before routine deployment they need to be appropriately localized and validated in prospective studies to evaluate clinical effectiveness and economic impact.

Ähnliche Arbeiten