News
Article
Medical Economics Journal
Author(s):
How Owais Durrani, DO, emergency medicine at UT Health, uses artificial intelligence.
Owais Durrani, DO
Dr. Owais Durrani started using an AI scribe less than a year ago. He opens the app, tells the patient it is helping him take notes, and asks for their permission to use it — and 99% of the time, the answer is yes. To help get a more complete note, he will verbalize the physical exam.
“If I’m pushing on their stomach, I’ll say something like, ‘So it seems like it hurts more in the right lower quadrant,’ and what it does is it translates that physical exam as well,” Durrani says, adding that the AI-generated note is about 80% to 85% complete as soon as the patient encounter is over. “It allows me to not feel rushed, especially in the ER [emergency room] setting, where we have so many patients. I can actually sit down, spend a little more time on the exam, explain why we’re ordering certain tests, and know that my note is going to be majority done, rather than me then going back and starting at zero.”
He’s also seen AI used in training on medical devices, like ultrasound, where the AI might instruct that a probe be turned a certain way or label things on screen when the optimal view is achieved. He said this helps when someone is instructing a larger class and can’t be with every student at once.
The other area of AI use is assistance with managing the EHR portal’s inbox. “There are hundreds, sometimes thousands, of in-basket messages from patients that may be anything from how frequently they should take this medication to what pharmacy was this sent to or how soon should I follow-up. The AI will prepopulate a message that you can review. It doesn’t involve any medical advice, so if it’s a very medical thing, it’s not going to put in advice or recommendations.”
Despite the rapid advances, Durrani says the broad consensus of his medical school friends and work colleagues is that AI may provide some helpful tools but won’t fundamentally change anything in the near future.
“We’ve been hearing about how AI is going to replace radiologists for the last decade, right?” Durrani says. “Have radiologists been replaced? No. Radiologists still exist, so it’s a lot of talk but not a lot of change that they’ve seen personally. And so, I think they take a cue from that.”
However, he does see in the future that some of these tools may become the standard of care, regardless of what physicians may personally think about them.
“Eventually, there’s going to be some AI technology that is going to be so useful or so good, where if you’re not using it, you’re going to be behind your colleagues, and then you may be the one that is the problem or liable,” he says.
Durrani sees AI as more of an augmentation of the profession, not a replacement for it. Just like AI has augmented radiologists but hasn’t taken their jobs away, AI tools can help other areas of need. But he says it is important for physicians to take leadership roles in how AI technology is developed and implemented, lest the profession end up with another EHR debacle, where something that was supposed to be positive ended up being a poorly designed workplace burden.
“Eventually, the FDA is going to approve a variety of these algorithms and technologies for diagnostic and clinical work flow items, and so we need to make sure that when that does happen, it’s not a physician trying to approve 10,000 things and we’re just like order assistance. We need to be in the lead, and we need to make sure that these things are actually augmenting us rather than potentially make it where we’re just following them. Those conversations are the most important now because once something gets implemented, it’s hard to go back.”