Recent television news reports are covering the use of AI in medicine, real estate, and other fields. One evening news segment highlighted a new Mayo Clinic study showing that AI can help specialists detect pancreatic cancer up to three years earlier than usual diagnosis. Headlines like this are becoming more common.
A recent NPR article covered a study published in the journal Science that found AI models can diagnose emergency room cases more accurately than ER physicians. In the story, NPR reporter Will Stone describes a patient with a pulmonary embolism, a blood clot that travels to the lungs. The case was analyzed by AI using the patient’s medical records. The AI suggested lupus as a possible cause of the patient’s worsening symptoms.
The study by Harvard and Beth Israel found that an OpenAI AI model performed very well at diagnosing patients. It outperformed both doctors and the earlier GPT-4 model. The model was tested on real cases and did better than two experienced physicians using only electronic health records.
This is the big conclusion for me — it works with the messy real-world data of the emergency department. It works for making diagnoses in the real world.
– Dr. Adam Rodman, a clinical researcher at Beth Israel and one of the study authors (source: NPR)
Reporter Will Stone also highlights another study published in the New England Journal of Medicine. It found that an AI model outperformed a large group of physicians in diagnostic tasks, showing how quickly this technology is improving. The model reportedly relied only on text, which shows there is still a need to include other types of diagnostic data.
Dr. David Reich, chief clinical officer for Mount Sinai Health System in New York, who was not involved in the work, told NPR, “This paper is a beautiful summary of just how much things have improved. You have something which is quite accurate, possibly ready for prime time. Now the open question is how the heck do you introduce it into clinical workflows in ways that actually improve care?”
Stone add that although AI models are excelling at diagnosing complex cases, but their effectiveness in real-world clinical medicine, where outcomes are more subtle and diverse, remains uncertain. While AI won’t replace doctors, rigorous testing through forward-looking trials is crucial to understand its impact on clinical practice.
“It’s a very challenging process to design these trials, but this study is a perfect call to action,” Dr. Rich told NPR.
Acknowledgement: This summary was generated using AI writing tools as part of my AI Explorations process. For more information, please refer to the original articles listed below.
Related Resource Links
- Performance of a large language model on the reasoning tasks of a physician | Science
- In real-world test, an AI model did better than ER doctors at diagnosing patients | NPR
- Next-generation AI for visually occult pancreatic cancer detection in a low-prevalence setting with longitudinal stability and multi-institutional generalisability | BMJ
- Mayo Clinic AI helps specialists detect pancreatic cancer up to 3 years before diagnosis in landmark validation study | Mayo clinic Network