The Economist
March 23, 2019.
Deep Medicine: How Artificial Intelligence Can Make Healthcare Human Again. By Eric Topol. Basic Books; 400 pages
FOR ALL the technological wonders of modern medicine, from gene-editing to fetal surgery, health care-with its fax machines and clipboards-is often stubbornly antiquated.
This outdated era is slowly drawing to a close as, belatedly, the industry catches up with the artificial-intelligence ( AI) revolution.
And none too soon, argues Eric Topol, a cardiologist and enthusiast for digital medicine.
Dr Topol’s vision of medicine’s future is optimistic.
He thinks AI will be particularly useful for repetitive, error-prone tasks, such as sifting images, scrutinising heart traces for abnormalities or transcribing doctors’ words into patient records.
It will be able to harness masses of data to work out optimal treatments (for both conditions and individuals), and improve workflows in hospitals. In short, AI is set to save time, lives and money.
He thinks AI will be particularly useful for repetitive, error-prone tasks, such as sifting images, scrutinising heart traces for abnormalities or transcribing doctors’ words into patient records.
Much of this is hypothetical-but AI is already outperforming people in a variety of narrow jobs for which it has been trained.
Eventually it may be able to diagnose and treat a wider range of diseases. Even then, Dr Topol thinks, humans would oversee the algorithms, rather than being replaced by them.
The fear the author harbours is that AI will be used to deepen the assembly-line culture of modern medicine.
If it confers a “gift of time” on doctors, he argues that this bonus should be used to prolong consultations, rather than simply speeding through them more efficiently.
That is a fine idea, but as health swallows an ever-bigger share of national wealth, greater efficiency is exactly what is needed, at least so far as governments and insurers are concerned.
Otherwise, rich societies may fail to cope with the needs of ageing and growing populations.
An extra five minutes spent chatting with a patient is costly as well as valuable.
The AI revolution will also empower managerial bean-counters, who will increasingly be able to calibrate and appraise every aspect of treatment.
The autonomy of the doctor will inevitably be undermined, especially, perhaps, in public-health systems which are duty-bound to trim inessential costs.
The Hippocratic Oath holds that there is an art to medicine as well as a science, and that “warmth, sympathy and understanding may outweigh the surgeon’s knife or the chemist’s drug”.
That is not just a platitude: the patients of sympathetic physicians have been shown to fare better.
As Dr Topol says, it is hard to imagine that a robot could really replace a human doctor.
As Dr Topol says, it is hard to imagine that a robot could really replace a human doctor.
Yet as demand for health care outstrips the supply of human carers, the future may involve consultations on smartphones and measurements monitored by chatbots.
The considerately warmed stethoscope, placed gently on a patient’s back, may become a relic of the past.
In the end technology may even be able to solve the empathy deficit.
Japanese engineers are working on robots that simulate human presence, or sonzai-kan.
A machine could never truly develop the shared humanity that helps patients heal. That doesn’t mean it cannot be faked.
Originally published at https://www.economist.com on March 23, 2019.
