Medical Humanities

Share               

Medical Artificial Intelligence

Melvin Chen


The secret of the care of a patient, as Peabody (1927) well knew with respect to the humanistic ideal of his medical profession, is in caring for the patient. Current medical practice, however, suggests an industry misaligned with this ideal and in disarray. The shift from private practice to hospital practice has given rise to a progressive dehumanization in the practice of medicine. Burnout is on the rise, with more physicians and nurses feeling emotionally exhausted, depersonalized, and ineffective. Can the introduction of medical AI close the quality gap or quality chasm in the medical domain by promoting better disease surveillance, facilitating early detection, allowing for improved diagnoses, and discovering novel treatments (Fogel & Kvedar, 2018)? Might the introduction of medical AI usher in an era of truly personalized medicine, in which healthcare systems are able to match individual patient characteristics with detailed evidence available in real time at the point of care (Abernethy et al, 2010)? At least two fundamental deficits in the design of medical AI systems – the causality deficit and the care deficit – remain to be addressed if this quality gap is to be closed (Chen, 2019).

This Medical AI project, helmed by a philos​opher (Dr. Melvin Chen, PI) and a physicist (Assoc. Prof. Chew Lock Yue, Co-I), is generously funded by the NTU Accelerating Creativity & Excellence grant and counts two other physicists and an NUH gastroenterologist among its collaborators. The research team has already worked on the theoretical foundations of causal reasoning and various computational approaches to causal reasoning and is currently modelling the differential diagnosis procedure with respect to patients who have a clinical presentation of jaundice. This project aims to address the causality deficit by designing an algorithm that can approximate causal reasoning and assist medical doctors in primary care settings with the differential diagnosis of jaundice.