Artificial Intelligence: an opportunity for a more humane medicine?

Jordi Varela
Editor

Changing doctors for robots is a recurring theme in science fiction, but the appearance of devices with digital assistance, such as Alexa or Siri, and the promotion of self-care through smartphones, admitted The Lancet in an editorial, now seems a plausible reality. On the other hand, Harvard Medical School Professor Warner Slack said that if a doctor can be replaced by a computer, then this doctor deserves to be replaced by a computer. In this context, Eric Topol in "Deep Medicine" brings light to this debate with the defence of an open thesis: "artificial intelligence (AI), instead of replacing doctors, should become an aid for them to recover lost humanism."

Superficial medicine versus deep medicine

People who suffer from diseases, or who are afraid of suffering them, depend on health care systems that offer them access to highly trained professionals who work in rigid and fragmented organizations, with insufficient data, and who generally cannot dedicate the necessary time. Eric Topol describes the current environment as superficial, expensive and ineffective medicine. On the contrary, the deep medicine recommended by the author would be one that would have quality clinical data, managed by the patients themselves with the interpretive support of methodologies from AI. This would allow a more personalized clinical practice, which would promote decisions shared, it would reduce unnecessary procedures, refine the interpretation of the results and help to choose the treatments that have shown more clinical effectiveness in each specific circumstance.

As regular readers of this blog know, we regularly post information about the waste and ineffectiveness of medicine, now called superficial (see the label "right care"). Let me, however, add to the list some data published in "Deep Medicine" on the consumption of tests financed by Medicare: for every 100 elderly people insured by this public insurance company, each year 50 CT scans, 50 ultrasounds, 15 resonances and 10 PET are paid. The author adds to this that it's known that half of these tests are unnecessarily prescribed.

The impact of deep learning

Software that learns opens the door to the mechanization of many routine medical tasks, such as the readings of radiological images, tissue plates, cytologies, dermatological lesions, fundus, electrocardiograms, ultrasound or colon polyps (see post "Machine learning and clinical practice"). In many of these fields, digital efficiency is already superior to human efficiency, without forgetting that machines work with performances unthinkable for professionals, and also don't make mistakes. Nick Bryan anticipates that, shortly, no radiologist will judge any X-ray without having previously been read by a machine.

No one is aware, least of all those affected, that the new panorama forces us to rethink many specialities. Eric Topol and Saurabh Jha propose the birth of a new speciality that arises from the rethinking that radiologists and pathologists must-do of theirs. According to them, it would be a question of training specialists in the application of AI in the clinic, consultants who, in short, they say, will be essential for clinicians.

AI and clinical practice

Reading diagnostic patterns is one of the strengths of AI, but digital advancement is still far from replacing doctors, even in the most vulnerable specialities. AI will increasingly impact clinical practice, with specific services, such as support for the interpretation of research, clinical reasoning, the search for the most recommended treatments or predictive models of survival of patients or advanced diseases. We must not forget the contributions of robotics in surgical procedures and in supporting the design of orthoses and prostheses.

Everything suggests that new doctors will have a more technological practice of the profession and, according to Eric Topol, this should allow them more time for higher quality and more people-centred clinical practice. Haider Javed Warraich, a fellow in cardiovascular medicine from Duke University Medical Centre says young doctors are predisposed to it but are their teachers also ready?