Anyone who has been to an emergency department for a non-life-threatening event knows the pain of waiting for hours before a bed becomes available for admission to the hospital for further care. Now, two studies show that using GPT-4, a large language model created by OpenAI, has the potential to help emergency department staff determine which patients need the most urgent treatment and which patients last will require hospital admission.
Journalists could find interesting stories by following these and other studies of AI technology in the emergency department setting and interviewing emergency medicine experts who are trying out such technologies.
In the first study, published May 7 in JAMA Network Open, researchers at the University of California, San Francisco, entered 10,000 pairs of patient information records from recent emergency department visits (minus patient names or other information identifier) in GPT-4 to see if the AI tool could identify which patient had the most severe condition.
The pairs featured one patient with a serious condition such as a stroke and another with a less urgent need, such as a broken wrist. AI correctly selected the patient with the most severe condition in 89% of cases. A subset of 500 pairs of patient information was then rated by both GPT-4 and clinicians. Outcome? GPT-4 was correct 88% of the time, a slight advantage over doctors at 86%.
Having AI assist in the triage process can help doctors allocate their time efficiently and serve as a back-up for decision-making, study author Christopher Williams, MD, said in a UCSF story about the study.
“Imagine two patients needing to be transported to the hospital, but there’s only one ambulance,” he said in the article, “or a doctor on call and there are three people running it at the same time and she has to determine who to respond to.” First.”
However, Williams noted that it’s not quite ready to be used responsibly in an emergency department setting without further validation and clinical trials along with efforts to erase racial and gender biases.
In a second study, published May 21 in the Journal of the American Medical Informatics Association, investigators at the Icahn School of Medicine at Mount Sinai found that GPT-4 also has the potential to predict which emergency department patients will be admitted. hospital.
The researchers entered data including triage notes from more than 864,000 emergency department visits by patients at seven Mount Sinai hospitals; Of them, 159,857 (18.5%) patients were hospitalized. The visits were for a variety of clinical conditions. In the first trial, the program was 77.5% accurate in predicting admissions; this accuracy improved to 83% when the program was given additional data to learn from.
Doctors and hospital staff could theoretically use the technology to shorten patient wait times and determine more quickly how many beds are needed in a hospital, which emergency department patients should be transferred to hospital floors, and which should be discharged , Alejandra O’Connell-Domenech said in May.
However, the researchers concluded that while the technology shows promise, it requires improvements.
“Moving forward,” write the study’s authors, “attention must be paid to the design of these systems to ensure that they enhance rather than complicate the decision-making process.”
Even if such technology is implemented in emergency rooms, doctors will always need to perform their own independent assessment to determine a patient’s treatment, Ajeet Singh, MD, a hospitalist and informatics specialist at the Medical Center, told The Hill. of Rush University. These programs don’t understand the words they’re trained on, but simply imitate reasoning by predicting the relationships of the words to each other, he said.
SOuRCES
#control #emergency #department #care
Image Source : healthjournalism.org