Your Hospital Is Like a Modern Airport. That’s Not Good

Derick Alison
Derick Alison
6 Min Read

SAN DIEGO — Avi Goldfarb, PhD, wants you as a healthcare professional to consider how much your institution is like a modern big-city airport.

That is, “a monument to uncertainty” largely built as “compensation for failures” to deliver on its basic mission.

Goldfarb, an economist at the University of Toronto, is a leading thinker about how artificial intelligence (AI) will transform the world’s economies. He delivered the keynote address opening the American College of Rheumatology annual meeting.

The air travel industry’s fundamental mission, he explained, is to get its customers to their destinations smoothly, quickly, and efficiently. If that worked all the time, there would be no need for the resort-like amenities that are now de rigueur in new airports. The fact is that “people don’t know how long they’re going to be there,” Goldfarb said. The restaurants, boutiques, spas, and lounges are all there to occupy people forced to wait… and potentially wait and wait some more.

He contrasted that with the bare bones facilities that serve owners’ of private jets at those same airports — at Boston Logan, it’s a “dingy shed” with a few tables and chairs. Billionaires rarely have to wait, they just need a place to stay out of the rain as their planes are pulling up. So it doesn’t have to be equipped like a resort.

He did not expand on how much of a modern healthcare system’s operating principles may built be along similar lines, inviting his listeners to think upon it themselves. Healthcare ideally should help people avoid getting sick and to cure them when they do. How often is that not the outcome?

Goldfarb said that what AI does is to make prediction cheap and thus widely accessible. And medical diagnosis is a form of prediction. A patient has a set of symptoms and characteristics, and the doctor must predict what is causing it. This has traditionally required much training, skill, and judgment and thus has been expensive. Now it’s becoming much less so.

What that means, Goldfarb suggested, is that more healthcare workers will be able to apply diagnostic tools, freeing physicians and other advanced professionals to focus on those areas where their skills can’t be replaced by machines.

He downplayed many of the concerns around AI: that highly paid professionals will find themselves replaced by robots or that humans will be ruled by them. Goldfarb reminded the audience that human judgment is embedded in much of what passes for AI. It’s still a human who provides the problem for the AI algorithm to solve, and it’s still a human who decides what to do with that solution.

He gave a simple example: a letter to be composed by ChatGPT. The human inputs the prompts for the letter’s topic and other parameters and, presumably, examines the output and perhaps gets rewrites with new prompts before actually sending it.

But healthcare and many other industries are virtually certain to undergo massive transformation as innovators figure out how to make the best use of cheap prediction.

He likened the current situation to that at the turn of the 20th century, when electricity and the devices making use of it (lights and motors) became ever cheaper, and quickly.

Up to that point, factories depended on steam engines for power. Most had a single large engine to power everything in a factory, and the machines needing the most power had to be closest to it because it wasn’t possible to transmit the power over significant distances.

At first, factory owners wanting to use the new technology simply pulled out their big steam engines and replaced them with big electric motors, otherwise leaving the factories and their production methods the same as before. Goldfarb called this a “point solution.”

But then, Henry Ford and others began to consider how to reorganize production from the ground up around the new technology. Electricity made it possible to put multiple smaller motors almost anywhere in a factory. Then it could be organized around the logic of the finished product. Thus the assembly line was born.

Something analogous is coming for healthcare, he suggested. What that will look like, however, isn’t clear. Of course, technologies falling under the AI rubric are now being used throughout medicine. But Goldfarb said these are all point solutions — the basic organization of healthcare remains exactly the same, except that physicians consult an AI-equipped tool in making diagnoses rather than managing it entirely on their own.

If nurses, pharmacists, and others can do as well at making diagnoses as the best physician, it’s inevitable that they will take over that function, Goldfarb suggested.

Yet there will still be demand for the other skills, the human judgment that physicians are uniquely capable of providing. How to maintain and boost that, as other workers take over the diagnostic aspect, is the challenge for reorganizing healthcare.

  • author['full_name']

    John Gever was Managing Editor from 2014 to 2021; he is now a regular contributor.

Source link

Share this Article
Leave a comment
adbanner