(FeaturedNews.com) – Medical professionals employed in the emergency departments could soon be in a position to employ artificial intelligence in order to determine more quickly whether or not a patient should be admitted to the hospital.
On Tuesday, the Journal of American Medical Informatics Association published a study on the topic. Researchers employed OpenAI’s GPT-4 to analyze over 864,000 emergency room medical files in seven Mount Sinai hospitals. Then, they tested how accurate the program was in predicting whether or not those visits would result in admissions to the hospital.
According to the study, around 18 percent of visits, which amounts to close to 160,000 visits, resulted in people being admitted to the hospital. The language model had around 83 percent accuracy in determining whether or not a patient should be submitted. This means that in the future it could be possible for emergency room doctors to use this type of technology in order to determine who should or should not be submitted.
Icahn School of Medicine professor of medicine Girish Nadkarni argued that it would be hard to achieve 100 percent accuracy because of the nature of some of these things. He added that when talking about having someone transferred to a hospital bed it could potentially not even be necessary to have perfect accuracy as the decision could ultimately be reversed.
The program could end up being useful for administrative and logistical tasks in emergency rooms and could help streamline the workflow.
Copyright 2024, FeaturedNews.com
