When journalists write about the disruptive power of artificial intelligence in healthcare they tend to zero in on radiology and pathology and for good reason. Both trades involve the interpretation of patterns from quantifiable image data – a thing that AI has proven highly capable of in several studies and commercial applications from facial recognition to the classification of hotdogs.
In each of these articles the elephant in the room is what will come of the radiologists and pathologists?
I learned today that this question was addressed head on last year in an article in JAMA titled, “Adapting to Artificial Intelligence, Radiologists and Pathologists as Information Specialists.”
The title is interesting enough. Then I saw that the two authors were Drs. Saurabh Jha (a radiologist) and Eric Topol and I knew I had to sneak a copy out from behind the firewall. Most people that follow healthcare know Topol. He’s a book author, MD, insightful, opinionated but professionally so. I’ve read Jha’s posts on The Healthcare Blog and follow him on Twitter (worth it). He’s intelligent, outspoken, and contrarian. Half the fun is watching him walk the line in the traditionally deferent culture of care and he does it well. I imagine these guys had a hell of a time threading the needle on this one. Here’s what they had to say:
AI will replace much of what radiologists and pathologists do today.
Drs. Jha and Topol open by shredding any notion that AI won’t be capable of doing most of what radiologists and pathologists do today. They cite examples of how startups and big companies alike have already proven to be even more effective than humans. They dedicate half of the 2 page article to making this unmistakable and convincing point – this is happening whether you like it or not. Period. The time to adapt to this new reality is now.
“To avoid being displaced by computers, radiologists must allow themselves to be displaced by computers.”
The stage is set for the real test in page 2. How will Drs. Jha and Topol handle the “what’s next for humans” part?
What ensues is riveting. I’ve re-read it three times. Here’s the play by play:
What are radiologists and pathologists really?
In the words of Drs. Jha and Topol:
“The primary purpose of radiologists is the provision of medical information; the image is only a means to information. Radiologists are more aptly considered “information specialists” specializing in medical imaging.
This is similar to pathologists, who are also information specialists. Pathologists and radiologists are fundamentally similar because both extract medical information from images.”
It’s worth recapping:
- Radiologists and pathologists are the same in that their primary job is the interpretation of images.
- What’s image interpretation and results delivery if not an information specialty?
- The images are a means to an end – you’re really information specialists.
That’s some meta stuff. But how will radiologists and pathologists feel about redefining their value proposition? Do they consider themselves information specialists after years of learning to read very specific images for very specific patterns? For that matter, is the diagnostic exam the means to an end for the primary care physician who also is really just an information specialist?
For now, who cares. It’s a bold opening argument. So bold that the authors follow up by doubling down on on the need for bold strokes.
Your two fields have long embraced technology. AI is better than you at many of the more complex tasks you do. Trained pigeons can do your job.
Before you accuse me of making this up, here’s the whole paragraph:
“Pathologists have embraced machines and technologies. Some tasks once performed manually by pathologists have been automated, such as cell counts, typing and screening of blood, and Papanicolaou tests, leaving pathologists with more complex tasks. Artificial intelligence can perform the more complex tasks of pathologists and, in some instances, with superior accuracy…Even though such studies need larger-scale validation with more diverse tissue types, it is clear in both radiology and pathology that many tasks can be handled by artificial intelligence. To underscore the commonality between radiology and pathology, researchers using operant conditioning trained pigeons to spot abnormal calcifications on mammograms and detect breast cancer on histology.”
The stage having been set (complete with my pigeon is smarter than your honor student example), Drs. Jha and Topol propose what I’m sure will be considered heresy by many:
Your two disciplines should be merged into a single entity known as the information specialist.
That’s huge. Neither of your positions need exist any longer. They should be merged. Forget the images, stick with the value-adds: interpreting the tougher images (though the previous paragraph argues that time’s running short on this task), suggesting further tests, and guiding clinicians. You will still be the “physician’s physician.” You just won’t have your title, nor do what you were trained to do.
I knew these guys wouldn’t disappoint.
AI + humans = greater productivity and influence for those that straddle the man v. machine boundaries
Jha and Topol make the case, “A single information specialist, with the help of artificial intelligence, could potentially manage screening for an entire town in Africa.”
Of course, it’s true. Those that learn to straddle the man / machine worlds are presented with new opportunity and influence. However, I’d prefer that anyone using this argument acknowledge the downside of this new supply versus demand reality.
I imagine this was the pitch to the knitters when the steam powered knitting machines came to replace them. Horse farmers were probably assured that Henry Ford’s efforts would somehow improve their lives. Generations later his managers likely sold the same fiction to their factory workers shortly before the demise of Detroit’s manufacturing economy.
Tell me again how many “information specialist” will we really need when one can do the work of thousands of what we once archaically referred to as “radiologists?”
In this case, let’s cut them some slack. They did, after all, just propose torching two disciplines.
Information specialists should train in different things.
“The trainee need only master enough medical physics to improve suboptimal quality of medical images. Information specialists should be taught Bayesian logic, statistics, and data science and be aware of other sources of information such as genomics and biometrics, insofar as they can integrate data from disparate sources with a patient’s clinical condition.”
I don’t know how to interpret this in any other way than, “you should no longer be trained, in any recognizable way, as a radiologists or pathologist.” If my interpretation is reasonable, it begs the question: is it really a radiology and pathology specific evolution that’s required? Or is this the case for a new and needed sub-discipline that radiologists and pathologists should consider applying to? And what about the existing AMA-accredited sub-discipline of medicine known as medical informatics? For another time…
“There may be resistance to merging 2 distinct medical specialties, each of which has unique pedagogy, tradition, accreditation, and reimbursement.”
I’ll bring the popcorn.
“United, radiologists and pathologists can thrive with the rise of artificial intelligence.”
Ummm ok. Just as long as we’re no longer radiologists and pathologists, right…?
“Jobs are not lost; rather, roles are redefined; humans are displaced to tasks needing a human element.”
Tell that to the displaced radiologists.
Their closing thoughts are surprisingly optimistic in light of the severity (and honesty) of their message to this point. They’re also more than a bit contradictory. My guess is that they had an animated back and forth on this one – maybe with the JAMA editors, maybe with each other.
In summary, the piece is an incredibly interesting read. It is as bold as anything I’ve come across in a peer reviewed medical journal and takes on a controversial and politically charged topic without apology. The evidence supporting the need for change is compelling. The call for a new specialty is, in my opinion, correct. I’m not sure that this need represents a natural evolution reserved for radiologists and pathologists versus an off ramp for them. Either way, radiologists, pathologists, clinicians, and administrators will have plenty to consider after reading this piece. Hats off to Drs. Jha and Topol for taking it on.