Jannah Theme License is not validated, Go to the theme options page to validate the license, You need a single license for each domain name.
USA News

AI could help doctors stay informed about diagnoses: Shots


Dr. Michael Mansour, an infectious disease specialist at Massachusetts General Hospital, is testing an AI-enhanced database that he uses to help make diagnoses.

Craig LeMoult/GBH


hide caption

toggle caption

Craig LeMoult/GBH

img 3934 custom 9bdb00ed78d0e4bec963ddf69ae107fa62137560 s1200

Dr. Michael Mansour, an infectious disease specialist at Massachusetts General Hospital, is testing an AI-enhanced database that he uses to help make diagnoses.

Craig LeMoult/GBH

While artificial intelligence seems to be making its way into all technologies, one area where it is seen as particularly promising is helping doctors make medical diagnoses.

And already, AI is tiptoeing into some medical practices.

Dr. Michael Mansour of Massachusetts General Hospital is an early adopter of a form of AI that could one day change the way doctors access information.

Mansour specializes in invasive fungal infections in transplant patients. “I have a beautiful photo of mushrooms in my office,” Mansour says with a laugh. “I really enjoy helping patients overcome, you know, some pretty devastating mold and yeast infections.”

When a patient arrives with a mysterious infection, Mansour turns to a computer program called UpToDate. It is an incredibly common tool, with over 2 million users in 44,000 healthcare settings in over 190 countries.

Basically, it’s Google for doctors: it searches a huge database of articles by experts in the field, all backed by the latest research.

A visitor from Hawaii brings a mystery

“Here’s an example,” Mansour said, turning to his computer. “If I meet a patient who comes from Hawaii.” The hypothetical patient’s symptoms make Mansour worry about an infection the patient contracted at home, so he types “Hawaii” and “infection” into UpToDate.

“And I get things like dengue virus, jellyfish stings, murine typhus, etc.,” he says, scrolling through a long list of answers on his screen. Mansour says he wishes this list was more specific: “I think Generation AI gives you the opportunity to really refine that.”

Mansour helped test an experimental version of UpToDate that uses generative AI to help doctors access more targeted information from its database.

Wolters Kluwer Health, the company that makes UpToDate, is trying to integrate AI so doctors can have more conversations with the database.

“If you have a question, this can maintain the context of your question,” says Dr. Peter Bonis, chief medical officer of Wolters Kluwer Health. “And saying, ‘Oh, I thought this’ or ‘So what?’ And he knows what you’re talking about and can guide you, the same way you might ask a master clinician to do.”

Software hallucinations are contraindicated

At this point, Wolters Kluwer Health is simply sharing the AI-enhanced program in beta form for testing purposes. Bonis says the company must ensure it is fully reliable before it can be released.

Bonis saw the program make mistakes that people who focus on AI programs with large language models call hallucinations.

He saw him one day citing a newspaper article in his area of ​​expertise that he was unfamiliar with. “And then I looked to see if I could find the study in this journal. It didn’t exist,” says Bonis. “My next query to the big language model was, ‘Did you invent this?’ » He said yes. »

Once these types of problems are solved, AI is seen in the medical world as having enormous potential to help doctors make diagnoses. It is already used as a radiology tool, helping with CT scans and X-rays. Another program called OpenEvidence, led by scientists at Harvard University, the Massachusetts Institute of Technology and Cornell University, uses AI to read the latest medical research studies and synthesize information for users.

AI could do the prep work before a patient’s appointment

Some doctors hope to use AI to go through and summarize a patient’s medical history before an appointment.

“It’s a long and very random process,” says Dr. June-Ho Kim, who directs a program on primary care innovation at Ariadne Labs, a partnership between Brigham and Women’s Hospital and the Harvard TH Chan School of Medicine. Public Health. . “And you might see a big language model that can digest that and produce some kind of incredibly useful natural language summary.”

In some cases, Kim says, AI technology can also help primary care doctors care for their patients without needing help from specialists. “This will free up time for specialists to focus on the more complex cases that they really need to focus on, rather than those that can be answered with a few questions,” he says.

A study published in the Journal of Medical Internet Research in August we tested the diagnostic skills of the popular ChatGPT program. Researchers fed 36 clinical scenarios into ChatGPT and found the AI ​​program was 77% accurate when making final diagnoses. However, with more limited information based on patients’ initial interactions with doctors, ChatGPT’s diagnoses were only 60% accurate.

“This needs to be improved,” says Dr. Marc Succi of Mass General Brigham, who was one of the paper’s authors. “We have delved into some specific parts of the clinical visit where it needs improvement before it is ready for prime time.”

Like a stethoscope, Succi says, AI will ultimately prove to be a reliable medical tool.

“AI will not replace doctors, but doctors who use it will replace those who don’t,” Succi says. “It’s the equivalent of writing an article on a typewriter or on a computer. It’s that level of leap.”

Mansour, a transplant fungal infection specialist at Massachusetts General Hospital, says he hopes AI will allow him to spend more time with patients. “Instead of spending those extra minutes looking for things, you could allow me to go talk to this person about their diagnosis and what to expect from management,” he says. “It restores the patient-doctor relationship.”

That relationship is strained as doctors get busier, Mansour says, and maybe AI can help.


npr

Back to top button