How can we tell when a newborn is in pain?
Tina Ho, associate professor of pediatrics at the University of South Florida, helps determine this.
Tina Ho joined USF in 2016 as a neonatologist in the Morsani College of Medicine Pediatrics Department, Neonatology Division and the Jennifer Leigh Muma NICU at Muma Children’s Hospital at Tampa General Hospital. Her research interest is aimed at examining the effects of anemia and oral iron supplementation on intestinal microbiome, inflammation and barriers in Very Low Birth Weight (VLBW) infants. Her findings will improve clinical management of anemia and optimize oral iron supplementation to give VLBW infants the most benefits and least harms.
Using AI to detect silent pain in newborns
When a newborn is in pain — especially one born too early or too weak to cry — how do we know?
That’s a question our research team has been working to answer with the help of artificial intelligence. Right now, Neonatal Intensive Care Unit nurses assess pain based on what they see. But those signs can be subtle, and the process is often subjective.
Our goal is to build a tool that can continuously detect pain in real time — not based on just one signal, but by combining multiple signs like facial cues, body movements, heart rate and breathing patterns. That way, nurses get an alert when an infant is in pain and can respond quickly.
To detect these signals, we’re using non-invasive cameras in addition to the routine vital sign monitors that are currently used in all hospitals. It’s designed to be practical, scalable, and most of all, gentle for the baby.
This work is truly interdisciplinary. I’ve been fortunate to collaborate with Dr. Yu Sun, a computer scientist specializing in AI and health care, as well as Dr. Dmitry Goldgof and Marcia Kneusel, who bring deep expertise in machine learning and nursing.
Together, we’re collecting and analyzing data from three NICUs across the country to train AI that can recognize signs of distress in newborns more accurately and consistently. It’s still early, but we’re hopeful this system can enable more effective pain management and help avoid unnecessary medication.
As a neonatologist, I’ve seen firsthand how hard it can be to tell when a baby is in pain — and how much better they do when we treat the pain promptly. This kind of innovation has the potential to change how we care for our tiniest patients — and give families greater peace of mind during a very difficult time.
Read More:
[USF] – Crying isn’t the only clue: USF researchers using AI to detect silent pain in newborns
