Any parent with a young child in La Verne has no doubt made a trip to the doctor’s office in order to relieve their child’s pain and discomfort. Fluid buildup in the middle ear is common in kids due to their still-developing anatomy. But more often than we’d like to admit, the go-to tool of the trade isn’t able to diagnose an ear infection accurately. Help may be available soon in the unlikely form of a smartphone app.
How an Otoscope Works (and Doesn’t Work)
Pretty much everybody is familiar with the otoscope. This medical device contains a handle, a head and a light source, and is used to look into the ear canals, helping an audiologist spot excess fluid or earwax buildup and other signs of ear infection. But its accuracy isn’t nearly as great as you might think. For the two million U.S. children who experience ear infections every year, that can lead to a longer, more painful recovery time.
Justin Chan, a computer scientist at the University of Washington, says an otoscope is about as effective as flipping a coin in telling the audiologist whether there is fluid in a child’s ear, pegging the odds at about 51 percent. More advanced diagnostic tools are often required in order to make this diagnosis, but many parents don’t bother following up with a specialist. A new smartphone app, currently in the testing stages, may help both parents and doctors.
Chan assembled a research team at UW in order to test out this smartphone app. Their results were published in a recent issue of Science Translational Medicine, and were rather impressive. The app was able to accurately detect fluid in the middle ear nearly 85 percent of the time. That’s comparable to the advanced tools an ear, nose and throat specialist would use to make a diagnosis.
Using the app is easy. A parent would simply need a paper funnel (instructions are included for assembling this at home) and a child who will sit still for a few seconds. That last requirement may be the most difficult! The funnel is inserted into the outer part of the ear and emits a chirp; from there AI does the rest, relying on a built-in-in machine learning algorithm to detect whether a sound is produced in response – an indication of an ear infection. Chan compares it to tapping on a wine glass. “Depending on whether it’s empty or not,” he says, “it’s going to sound different.”
The app is still in the testing stages and will require a larger group of test subjects for more accurate results. But with such encouraging early results, Chan and his group believe the app will be available to Android and iPhone users in about a year. They are excited for its potential, which could extend beyond detecting the presence of fluid in a child’s ear to include a telemedicine function – data sharing with the child’s pediatrician, who could then refer them to a specialist without the need for an office visit.