Predicting blood clots before they happen in pediatric patients

The Monroe Carell Jr. Children’s Hospital in Vanderbilt has initiated a study to determine the impact of a predictive model for identifying pediatric patients at risk of developing blood clots or venous thromboembolisms (VTEs).

The study uses advanced predictive analytics to educate medical teams about patients who are at risk for blood clots before they occur.

“Hospital-related blood clots are an increasing cause of morbidity in pediatrics,” said the study’s lead investigator, Shannon Walker, MD, clinical fellow of Pediatric Hematology-Oncology at Children’s Hospital.

Although these events are rarer in children than in adults, Walker noted an increase in blood clot development.

“The reason kids get blood clots is very different from adults,” said Walker, who worked with mentors Allison Wheeler, MD, MSCI, assistant professor of Pediatrics and Pathology, Microbiology and Immunology, and C. Buddy Creech, MD, MPH , director of the Vanderbilt Vaccine Research Program and associate professor of Pediatric Infectious Diseases.

“There was no standardized protocol for preventing blood clots in pediatric patients. Because we noticed an increase in blood clots and we recognized that the adult strategy would not work for our patients, we wanted to look at each patient’s individual risk factors and factors. look at how we can focus on targeted prevention of blood clots. “

The study, which will be published in Pediatrics, describes how the team built and validated a predictive model that can be automated to run within the electronic patient record of each hospitalized patient.

The model includes 11 risk factors and was based on an analysis of more than 110,000 admissions in children’s hospital and has been validated on more than 44,000 individual admissions.

The team is currently studying the use of this model along with targeted intervention in the clinical setting in a study called “Children’s Likelihood of Thrombosis,” or CLOT.

The prediction model is used in this way: every child admitted to hospital is assigned a risk score. The patients are randomized, so in half of the patients, elevated scores are reviewed by a haematologist and then discussed with each patient’s medical team and family to determine a personalized prevention plan. All patients, regardless of randomization, will continue to receive current standard of care.

“We’re not using a one-size-fits-all plan,” said Walker. “This is an additional level of assessment that allows for a highly personalized recommendation for any patient with an elevated score. The score is updated every day, so as risk factors change, the scores change accordingly.

“We are assessing in real time the use of this model as a clinical support tool. We saw a clinical possibility of something that we could improve and have moved on to build the model – to identify high-risk patients.” to run the CLOT trial, which will run until the end of the year. ”

Walker’s study was possible with the help of the Advanced Vanderbilt Artificial Intelligence Laboratory, or AVAIL. Only in its second year does the program lead the way in supporting artificial intelligence tools at VUMC through project incubation and curation, including facilitating clinical trials to assess their effectiveness.

“AVAIL served as a catalyst, in this case bringing experts in a complex pilot development close to it so that a great synthesis could take place,” said Warren Sandberg, MD, PhD, executive sponsor of AVAIL, along with Kevin Johnson, MD.

“The unique thing about this particular project is that we were not only able to predict complications, but also test the model in a rigorous, pragmatic, randomized, controlled trial to see if it would benefit patients,” said Dan Byrne, senior biostatistician for the project and director of research on artificial intelligence for AVAIL.

“The future of this type of work is limitless,” he said. “We can hopefully use this approach to predict and prevent pressure ulcers, sepsis, falls, readmissions or most complications before they occur. At Vanderbilt, we are raising the bar when it comes to the science of personalized medicine and the application of artificial intelligence in medicine in a way that is both ethical and safe. “


The paper’s authors include Walker, Wheeler, Creech, Byrne, Henry Domenico, MS, and Benjamin French, PhD. Ryan Moore, MS, is the biostatistician for the CLOT study.

Disclaimer: AAAS and EurekAlert! are not responsible for the accuracy of any press release posted on EurekAlert! by contributing institutions or for the use of information through the EurekAlert system.

Comments are closed.