Software Reads Facial Expressions To Monitor Kids’ Pain While In Hospital


If you've ever been to a hospital emergency room, you've probably rated your pain using one of those charts showing a series of emoji-esque faces that look progressively more miserable. During future hospital visits, however, clinicians may be able to use images of your actual face to measure your pain level.

Monitoring pain in children can be particularly challenging, as they often lack the ability to communicate it. But running video of children's faces through specially designed software allowed researchers to rate kids' pain about as accurately as parents and even more accurately than nurses were able to, according to a study in the journal Pediatrics that was released online June 1.

"Clinicians - and human beings in general - usually look to the face when looking for signs of pain, particularly in patients. We are taking advantage of that fact and also prior research that shows that if you break down the actions of the face you can actually get a lot of information about pain," senior author Jeannie Huang of the UC San Diego School of Medicine told Tech Times.

Even when adult patients match their pain level to that spectrum of cartoon faces, the results can be imprecise. Since many children are unable to report their own pain, either because they are very young or are developmentally delayed, nurses or parents often assess pain levels for them. This extra step in the already imprecise process of translating pain into words can be problematic. Furthermore, visits from nurses to check on pain may not always coincide with when children are actually experiencing pain.

"We don't treat pain very well in many patients, but we particularly don't do it very well in children," says Huang.

With this software, kids' pain levels can be monitored continuously with video. It uses 46 distinct facial movements to determine a child's level of pain. Based on these movements, the software calculates a pain score. In the study, which included children ages five through 18, the software was able to use video of the patients' faces generate a a pain score similar to their self-reported pain scores.

This is not a step toward robo-nurses, however, according to Huang.

"We're not going to replaces nurses," she says. "But frankly, in this day and age, nurses are busy."

Further research is needed before this technology can be implemented in a clinical setting, but its applications extend beyond just children.

"We would love to extend this to many other patient populations who often have difficulty expressing their pain," Huang says.

Photo: Suzette | Flickr

ⓒ 2018 All rights reserved. Do not reproduce without permission.
Real Time Analytics