Researchers at Nashville, Tenn.-based Vanderbilt University Medical Center developed an artificial intelligence model to identify pediatric patients who are at high risk of developing blood clots. While the model was shown to be accurate in identifying those at highest risk, physicians were reluctant to follow the accompanying recommendation, according to findings published Oct. 13 in JAMA Network Open.
Researchers analyzed the electronic medical records of more than 110,000 admissions to Vanderbilt's Monroe Carell Jr. Children's hospital to identify 11 factors tied to blood-clot risk. They then created a predictive model using this information and calculated a daily risk score for every pediatric admission.
A clinical trial ran from November 2020 to January 2022 and included more than 17,000 pediatric hospitalizations. In the study, scores for high-risk patients were accompanied by a recommendation to start blood thinning therapy to prevent the development of blood clots.
Researchers found no difference in the rate of blood clots between the study and control groups at the end of the trial, which they believe is because physicians were reluctant to accept the recommendation. They found the recommendation to begin blood thinners in high-risk patients identified by the model was followed only 25.8% of the time.
The researchers are planning another study to evaluate the providers' reluctance to accept the recommendation and to overcome the concerns.