In Memory of John Fox – UK pioneer in cognitive approaches to clinical Artificial Intelligence
It was with great sadness and shock that all good friends and colleagues of John Fox heard the news of his unexpected death late this past August. John was a pioneer in developing computational methods for modelling and representing the clinical knowledge in medical guidelines through his PROforma software system, which was designed to formalize high level concepts like decisions, plans, and their logical argumentation under uncertainty, and received the EFMI 20th Anniversary Gold Medal in 1997.
John received a PhD from Cambridge University, having studied under the cognitive psychologist John Morton, which led to a post-doctoral year at Cornell University with visits to Allen Newell and Herb Simon at CMU in Pittsburgh where he began to look into the opportunities that cognitive AI approaches could provide for the practice of medicine. After returning to the UK and trying out Bayesian methods in aid of medical education with clinical simulations (the FIRST AId system) he joined the Imperial Cancer Research Fund (ICRF — now Cancer-UK) in London to establish an AI program, developing the rule-based software framework Xi. The PRforma project evolved from his later work on the Oxford System of Medicine (OSM) initially funded by Oxford University Press as a computer-based reference tool for practitioners, but later by a number of EC grants and funding for the start-up InferMed Ltd from ICRF. John focused his work on techniques from computer science and software engineering which could help build systems for advising humans on safety-critical matters such as those arising in medical, nursing and other healthcare professional situations. With co-author S. Das he wrote the book Safe and Sound: Artificial intelligence in Hazardous Applications (2000).
He moved his research team from London to Oxford University shortly thereafter to work on the CREDO project, a collaboration with Edinburgh University and the Royal Free Hospital in London, which was aimed at showing the benefits of cognitive systems in clinical decision-making and management. It was focused on cancer especially, with both a commercial company (Deontics Ltd) and a not-for-profit company (OpenClinical) to help develop software for publishing models of best clinical practice as executable knowledge to “empower clinicians and the informatics community to share their expertise”. This John states in his contributed short personal narrative in the soon-to-be published IMIA eBook entitled International Medical Informatics and the Transformation of Healthcare (Kulikowski, Mihalas, Greenes, Park and Yacubsohn, Eds.), where he also emphasizes that informaticians owe a “duty of care” to their users, clinicians and their patients by writing software that address quality and safety issues in the designs and implementations that are applied clinically, especially when they deal with safety-critical matters – even as he noted that they still often tend to be ignored.
John Fox will be deeply missed by his many friends and collaborators in the medical and health informatics communities among whom he was known for his insights into clinical cognition, his personal qualities as a committed mentor and inspirer of researchers and collaborators, and for the enterprising energy with which he pursued the principled development of systems that would prove both safe and useful in clinical practice.