Computerised Decision Support in Cancer Multidisciplinary Meetings: Principles and Practice
30th March @ 4:00 pm - 5:00 pm
I will describe an AI argumentation-based decision support system for breast cancer multidisciplinary meeting (MDM) which was successfully trialled at the Royal Free Hospital, Breast Unit. The system models the patient care pathway, automatically aggregating 16 best practice guidelines, and using routinely collected data recommends individualised diagnostic and treatment actions. The system was evaluated in a real setting and demonstrated excellent agreement with the MDM team decisions. One of the system’s most important strengths is the ability to provide real-time audit. I will explain the decision-making modelling approach and I will discuss a range of strengths and weaknesses of the system, from knowledge representation to clinical service delivery. I will end the presentation discussing how to build on this research to learn clinical pathways from EHR data and perform automatic computation of individualised care plans using AI decision-theoretic planning.