Smartphone app may allow parents to screen newborns for jaundice

Last Updated: 2014-09-22

By Rob Goodier

NEW YORK (Reuters Health) - A new smartphone application still in development may someday help parents and health care providers screen for jaundice in newborns.

Early testing found that the application has greater accuracy than visual exams by doctors, according to research presented September 16 at the International Joint Conference on Pervasive and Ubiquitous Computing (UbiComp 2014) in Seattle, Washington.

Dr. James Taylor, a pediatrician at the University of Seattle who is helping to develop the technology, told Reuters Health by email that when it's ready for release, the BiliCam's performance might compare to that of the much more expensive transcutaneous bilirubinometer.

"We think that BiliCam will be useful by relieving anxiety of parents about jaundice," Dr. Taylor said. It will also provide an inexpensive, reliable tool to help doctors triage newborns with jaundice, he added.

His team believes the BiliCam will reduce healthcare costs by earlier identification of babies who need treatment, when it's easier and cheaper to deliver it.

The BiliCam could also allow parents to monitor their newborns' health at home simply by taking pictures of the baby with their phone camera's flash turned on. The parent or clinician first places a "color calibrator," a sheet the size of a business card printed with eight blocks of different colors, on the baby's chest. The software processes the images, adjusts color and white balance, accounts for different lighting conditions and skin tones, and then gives a report.

Tests with data from 100 newborns show that the BiliCam correlates closely to blood tests of total serum bilirubin, at a rank order correlation coefficient of 0.85. At that degree of correlation it bests a visual examination, which correlates to the blood test in a coefficient range of 0.35 to 0.75.

Past research shows that doctors using the visual method may also tend to underestimate the degree of jaundice, which may be of particular concern, the researchers say.

The BiliCam's performance may also put it in the same league as the transcutaneous bilirubinometer, which correlates well with the blood test, said Lilian de Greef, a doctoral student in computer science and engineering at the University of Washington who is involved in developing the technology.

An at-home test like the BiliCam may be important for parents, given the high rate of jaundice in newborns.

"Because the majority of all newborns have some level of jaundice . . . and because jaundice doesn't tend to peak until babies are at home with their families (out of the nursery) at day of life 3 or 4, having access to this app will provide a chance for families to access data without a trip to the clinic, hospital, or birthing center," Dr. Wendy Sue Swanson said.

"This really puts the power of diagnostics into the patient and family's hands and that's earnest opportunity for improved health care driven by patients," said Swanson, a pediatrician and executive director of Digital Health at Seattle Children's Hospital, who is familiar with BiliCam but is not involved in developing the application.

The BiliCam is not quite ready to roll out, however. The researchers are tackling issues like how to display the results of the test and how to tweak the color calibration card to make it more accurate and user friendly.

"Currently, everything we developed was mostly to collect data and process them for our evaluation. There's still lots of work to be done on how to present everything for users like parents," de Greef said.

When they think it is ready, the researchers plan a phased release, first to clinicians and then to parents.

"We think that parents will use BiliCam like they use thermometers now. Instead of just knowing that their baby looks jaundiced they will have an estimate of the bilirubin level that can be (reported) to the baby's healthcare provider," Taylor said.

The BiliCam research was funded in part by Coulter Foundation and a National Science Foundation Graduate Research Fellowship.

A full copy of the paper presented at the meeting is available from the Association for Computing Machinery, at http://bit.ly/1shPIML.