Skip to Main Content

Middle ear infections, which are detected by fluid buildup in the ear canal, are the most common reason for pediatric visits. Some 2 million kids every year have these infections, but beyond signs of discomfort or sometimes a fever, it’s hard to know at home what may be happening. And left unattended, such infections could cause serious complications, including hearing loss and meningitis.

A general physician, either in the emergency room or in a regular office, would only have an otoscope — which can spot fluid buildup in the ear — to make an assessment, but these devices are notoriously unreliable.

“An otoscope has a 51% chance — a coin flip — of telling you whether or not there’s fluid in the child’s ear,” said Justin Chan, a computer scientist at the University of Washington.


More accurate evaluation will likely require a visit with a specialist, but that takes time, which often leaves parents — and children — in the lurch.

This is the scenario that Chan and his colleagues are working to avoid. In a new study published Wednesday in Science Translational Medicine, Chan and his team describe a smartphone app — one that works on iPhone and Android — that can identify the presence of fluid in a child’s ear. The researchers found that physicians — and parents — can use it with relative ease. All the app needs is a small paper funnel, which can be fashioned at home from a piece of paper and with directions in the app.


The funnel is placed on the outside of the ear, at which point the app sends a bird chirp-like sound into the ear. Depending on the sounds that the app picks up in return, a machine learning algorithm built into the app is able to tell whether or not there is liquid in the ear. “It’s like tapping on a wine glass,” Chan said. “Depending on whether it’s empty or not, it’s going to sound different.”

The study did test a small set of patients — 98 in total — although plans for larger studies are underway. Kids also have to sit still for the app to work properly: It takes a little over a second for the app to send the chirps, but for a kid in discomfort, this could feel like a long time.

Chan spoke with STAT about his team’s work and the plans they have for getting the app “into millions of hands.” This interview has been condensed and lightly edited.

How is your app better than what’s currently available?

One advantage is that the app is more accurate than an otoscope. The probability that our app will detect correctly if there is fluid in your child’s ear is 85%, which is comparable to specialist tools. But the difference is you can use our app in the comfort of your home.

How did you find the parents responded to the app?

We approached parents just before their kids were about to have surgery to have ear tubes put in. We did a couple minutes of training, and found that the parents had close to identical accuracy to the physicians who had been testing them. We had 25 parents do the test, and 24 of the cases, the parents had the same result as the physicians who used the app.

How do you imagine this app will integrate with the current health care landscape?

One way our app could be used is with telemedicine. You could share the results with the doctor and book an appointment with the specialist if the doctor thinks you need to see one. It could reduce the amount of waiting time and even reduce costs. These features are not yet available, but we’re working towards them.

Is it already available in app stores? And what does the future look like?

It’s not currently available in app stores, but we’re hoping to have it in there in a year’s time. We’re also hoping to get FDA approval by the end of year. We created a company — Edus Health — and are partnering with physicians in developing countries. One need is for parents, but doctors in developing countries may not have access to specialist equipment.

  • I’m fearful this “technology” may result in unnecessary antibiotic treatment unless it is followed up by an adequate examination.The presence of middle ear fluid does not necessarily mean that a bacterial otitis media is present. As any practicing pediatrician knows a viral serous otitis may not require antibiotics. If the technology accurately detects middle ear fluid then a thorough pneumatic otoscopic examination is indicated.

  • A middle ear effusion (fluid on the ear) is not the same as an ear infection. It is common with colds and resolves on its own. Many children with fluid have no need of an antibiotic or any treatment. While an otoscope may not be the best way to diagnose fluid, I’m sure it’s accuracy in the hands of an experienced physician to diagnose an ear infection (redness, bulging, opaque fluid) is much higher than 50%. This possibly might be more helpful to identify chronic effusions, but used as a telemedicine tool, it could lead to over-prescribing of antibiotics.

  • AS a GP, the addition of such a tool would be excellent but since kids with otitis media are often squirming around, the tool has to be able to be used under routine conditions of the baby moving and screaming.

    • As a pediatrician , I think this tool would be an excellent addition to the armamentarium, but should not thoroughly replace visualization of the tympanic membrane. As mentioned above, limitations would include squirming children and also significant wax buildup that would potentially interfere as well. Also the first line of the article is incorrect – the fluid build up is behind the eardrum, not in the canal. I wonder as well if it differentiates between acute suppurative otitis media ( an actual middle ear infection) and acute otitis media with effusion ( non-infected fluid behind the eardrum). These are among the reasons that although many entities have tried, they still haven’t totally replaced doctors with computers.

Comments are closed.