Boston researchers are testing whether artificial intelligence could one day help paramedics treat children during life-threatening emergencies.
Inside a Pro EMS ambulance bay in Cambridge, first responders recently worked through a high-stakes emergency drill. The patient was not real, but the scenario was designed to feel urgent: a child with no pulse who wasn’t breathing.
The simulation is part of research led by Dr. Tehnaz Boyle, a pediatric emergency physician and associate professor of pediatrics and emergency medicine at Boston University. Her team is studying whether connecting paramedics with doctors over video can improve patient care in the field.
All of the simulations are recorded to observe how paramedics respond during pediatric emergencies.
“It’s an opportunity to capture it on video so we can actually assess the processes of care,” Boyle said.
NBC10 BostonNBC10 Boston
Paramedics simulating emergency treatment on a mannequin-shaped toddler.
From the team’s initial study, another goal has emerged – using those recordings to train AI models that could eventually support first responders during real calls.
“I think in the early stages, success would look like being able to develop a model that could identify actions reliably, so flag that an action actually occurred correctly and not mistakenly flag an action that could cause harm,” Boyle said.
For example, during CPR, AI could analyze technique and tell responders how to adjust. There are certain details researchers want an AI model to be able to recognize.
“How many CPR compressions were needed, what was the time gap between them, what was the hand position, so all of these are questions that we’re very interested in,” said BU computer scientist Deepti Ghadiyaram.
A new study has found that teens are forming deeper connections with generative AI, raising fresh questions about what it could mean about their development.
Pediatric emergencies are rare, but the stakes are high. Researchers say that makes real-time support especially important.
The work is still in its early stages. If successful, the technology could eventually help not only paramedics, but also bystanders during 911 calls.
“This is an example of how we can use this technology to benefit people and ultimately, hopefully, save someone’s life,” Boyle said.