Adam has come to see his family doctor because he has a sore, scratchy throat, and feels generally unwell.
Dr. Doug Kavanagh probes a little, asking how long he’s had the symptoms, if he has a rash, or fever, or gastrointestinal symptoms, working on a differential diagnosis. He takes Adam’s temperature, checks his lymph nodes and, because his tonsils are inflamed and marked with little white spots, does a rapid test to confirm the patient has strep throat.
It’s a standard visit except for one thing: Dr. Kavanagh sits directly across from the patient, not at a desk behind a computer. He doesn’t take notes either.
Yet, the moment Adam leaves the office, his electronic medical record is updated with detailed encounter notes, all produced seamlessly by a medical dictation chatbox, better known as an AI scribe. A prescription for antibiotics is on its way to Adam’s pharmacy.
“This has been utterly transformative to my practice,” Dr. Kavanagh says. “An unbelievable productivity boost.”
The patient in the simulation, Adam, also appreciates the invisible bot (a program on the doc’s laptop, in this case the WELL AI voice) because he feels more seen and heard. “Nobody likes it when the doctor’s staring at a computer instead of you,” he says.
Physicians spend, on average, more than 12 hours a week typing notes into charts. It’s work that’s often done after hours.
Charting is just one part of the paperwork. There are referrals, countless forms for insurance, sick notes, billing, scheduling, and more. (Physicians often have minimal office staff because the costs come out of pocket.)
And it adds up.
According to one study, doctors spend more than 18 million hours a year doing administrative work that is unnecessary or could be done by someone else (including an AI bot, presumably.)
That time suck is the equivalent of 55.6 million patient visits, and a lot of missed family dinners.
Surveys have repeatedly shown that the No. 1 contributor to physician burnout is paperwork.
But we have to be careful – technology can be a source of that burden, too.
Physicians often have a love-hate relationship with electronic medical records (EMRs). They are essential but don’t always connect seamlessly with systems being used by other practitioners and governments. So we still use fax machines extensively.
As AI tools are integrated into practice, we have to be careful to not repeat the errors of EMRs. There are dozens and dozens of types of EMRs in use, and they lack interoperability. The technology was also introduced slowly because of fights over who should pay for them.
AI-based tools like scribes and chatboxes need to be integrated more thoughtfully, for the sake of patients and practitioners alike, especially given the fact that the possibilities are mind-boggling.
AI is already being used widely for scheduling and, to a lesser extent, for charting.
But ChatGPT, which has only been around since November last year, can be used for everything from writing sick notes – which shouldn’t be necessary at all, but that’s another matter – to helping formulate treatment plans and more.
AI bots can already read diagnostic scans as well or better than humans. The next generation of AI scribes will be able to actually converse with patients and do real-time translation, which would be a great help in multicultural Canada.
But we can’t forget that AI is a tool.
Human oversight of bots’ recommendations is essential, especially since there are sometimes glitches, including ChatGPT’s “hallucination” problem. (It occasionally makes things up.)
AI can scrape the web for huge amounts of information, and then summarize and interpret it. But it has a “garbage in, garbage out” limitation. Worryingly, it can perpetuate biases that pre-exist in research, and amplify misinformation.
If we use data smartly and ethically, artificial intelligence and machine learning tools can help revolutionize medicine and make personalized medicine a reality.
Will AI replace doctors? That’s highly unlikely. However, doctors who use AI smartly will replace those who don’t.
The human touch is essential. Done right, AI should free clinicians from the shackles of technologies that have come to dominate their work in recent decades – that begins with getting out from behind the desk, and away from the computer.
Medicine cannot be reduced to a bunch of algorithms. It’s art as much as science. Technologies like AI should allow doctors to get back to the basics of medicine. Listening, observing, touching, interpreting, and healing.
Just faster, and better.