Computer scientists at the University of Washington are engineering a phone app that pregnant patients could someday use to monitor their fetuses' heart rates from home.
The researchers recently tested an iPhone’s speaker and microphone to determine whether it could be used to help estimate fetal heart rates. It is a process clinicians usually do with an ultrasound device, Doppler ultrasound, to check on fetal well-being.
The scientists published their findings in the journal Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies.
“I truly believe that health belongs to everyone,” said lead author Poojita Garg, a second-year doctoral student in computer science and engineering at UW. “I want to use my computing expertise and create accessible health tools, especially for women's health.”
Garg helped test her tool on 23 pregnant patients at UW Medicine. A machine learning model then used the iPhone audio to estimate fetal heart rates. Results showed the technology was within a clinically acceptable range of error.
“I hope to bring it to the market so that people can access it. But before that, we need bigger clinical testing,” Garg said. She wants testing to include thousands of people, “because if we get more data, we can actually train more sophisticated artificial intelligence.”
Garg also wants to test the tool on patients displaying irregular fetal heart rates.
Pediatric cardiologist Bhawna Arya said the technology shows promise, but should ultimately be used in concert with a medical team. Arya is the director of fetal cardiology at Seattle Children’s Hospital, where her team has already studied the benefits of giving patients fetal heart rate monitors that have been approved by the U.S. Food and Drug Administration. These patients were instructed to call their clinician immediately if they detected an irregular fetal heart rhythm at home.
Still, Arya saw how a phone app might help in certain instances.
“For the families who can't afford it or can't get to us, wouldn't it be nice if you could have a phone-based app,” Arya said. Still, "in an ideal world, people wouldn't use those monitors without being plugged into a medical provider.”
Before the AI-based tool comes to market, Arya said, there should be larger clinical studies testing the app on abnormal heart rates. For those studies, she hopes to collaborate with the UW researchers.
“When you have an irregular heart rate or a slow heart rate, or skipped beats or fast beats — what will AI do?” Arya said. “What do we need to teach the machine to know not to take those pieces away, but make sure that those are an important part of the evaluation?”
Scientists will also need to test the feasibility of pregnant people using the tool on their own at home. Research shows at-home monitoring can comfort and empower pregnant people. Similarly, Arya said, it will be important to test whether this phone app would give patients peace of mind or increase anxiety about the health of their fetus.
“That third trimester period… you may find some normal variability that could cause the parents anxiety,” Arya said. “As babies get bigger, they run out of space, and they can sort of lean on their umbilical cord, and that can lead to these slow rhythms that can be very normal.”
As Garg continues to develop and study the app for her doctorate, she hopes to make fetal heart rate monitoring more accessible globally as well as locally, in maternity care deserts like those in some Washington counties.
“Such an app would really benefit front-line health workers to quantitatively assess the fetal heart rate when they go door to door, checking on women in their communities,” Garg said.
