AI tool can ID pediatric patients at risk of blood clots

Editor's Note

An artificial intelligence (AI) tool called CLOT (Children’s Likelihood of Thrombosis) developed at Vanderbilt University Medical Center accurately identified pediatric patients at high risk for blood clots in a clinical trial. The findings were reported in JAMA Network on October 13.

The researchers analyzed the electronic health records of more than 110,000 admissions to the Monroe Carell Jr. Children’s Hospital at Vanderbilt. Using the tool, they were able to review over 100 patients a day.

Some highlights of the study, titled "Model-Guided Decision-Making for Thromboprophylaxis and Hospital-Acquired Thromboembolic Events Among Hospitalized Children and Adolescents," include:

• The researchers identified 11 factors associated with blood-clot risk and developed a predictive model that could scan medical records automatically and calculate a risk score.
• Despite the accuracy of the test, the study found no difference in outcomes compared to a control group.
• Researchers theorized that the similar outcomes may be due to the fact that the AI’s recommendation to begin blood-thinning therapy in these patients were accepted by treating physicians less than 26% of the time. 

Researchers said the study confirmed the safety and effectiveness of the tool in a healthcare setting and offered insights into how to successfully incorporate it into clinical practice.