NIST Partners With University of Rochester, Google to Develop AI Model for Predicting Cardiac Events

The National Institute of Standards and Technology has teamed up with the University of Rochester and Google to explore the potential of artificial intelligence to identify irregular heart rhythms in firefighters.

Electrocardiogram data from over 100 firefighters generated by the university and machine learning were used to train an AI model called Heart Health Monitoring to predict abnormal heart rhythms, which may lead to cardiac events.

The H2M model was also trained on data from non-firefighters; however, it resulted in a 40 percent error rate, the NIST said.

The project’s goal is to have the model incorporated into portable heart monitors that firefighters can wear to identify and warn them of heart trouble signs early on.

A study from the National Fire Protection Association revealed that cardiac events caused the deaths of 36 firefighters on duty in 2022. Deaths due to a cardiovascular cause are more common among on-duty firefighters compared to police officers and emergency responders, the NIST said.

Researchers from the NIST, the University of Rochester and Google published the results of their project in the Fire Safety Journal.