Parents as diagnostic assistants: App help us!

With a homemade paper funnel and a mobile phone app, parents are able to detect fluid accumulation in the middle ear of their children.

Sensitivity and specificity featured, as in classical audiology

With a homemade paper funnel and a mobile phone app, parents are able to detect fluid accumulation in the middle ear of their children. What only MacGyver could have done in the past, modern technology now makes it possible for everyone.

Acute and serous otitis media are associated with fluid accumulation in the middle ear. If this is not discovered or is discovered too late, there is a risk of hearing damage as well as restrictions on child development and speech ability.

Scientists have now succeeded in developing a smartphone-based system with which nursing staff and parents can follow the course of therapy at home themselves after an initial medical consultation.

Movement of the eardrum measured

The system, which runs on all common smartphones, generates a short acoustic pulse at the ear with alternating frequency and thus causes the eardrum to vibrate. The vibrations are then recorded as a reflection of the sound on the device’s microphone and evaluated in the app using an algorithm.

The app detects vibration deviations due to a change in the movement of the eardrum and then informs the user of a "suspected fluid in the ear". If everything is normal and the eardrum vibrates unaffected, the message "No suspicion of fluid in the ear" is displayed.

The algorithm works with precision

The algorithm was initially tested on 98 patients between 18 months and 17 years of age. The measurement results were compared with reflectometry instruments used in parallel. The fluid status in the middle ear was documented.

It is interesting to note that the mobile phone-based system, used by a pediatrician, had a sensitivity of 84.6% and a specificity of 81.9%. In comparison, pediatric patients with classical reflectometry achieved sensitivities of 76.9% and specificities of 77.8%. While parents applied the app-based test themselves after a short briefing, in 24 of 25 cases the accuracy of the measurements was comparable to that of clinicians. 

In addition, the app is able to distinguish between different liquid types. She correctly detected 6 of 7 serous diseases, 10 of 11 mucous diseases and 4 of 4 purulent manifestations.

Relevance in practice?

Machine learning in combination with smartphones seems to be increasingly suitable to support doctors in their daily work. The app was able to detect the accumulation of fluid in the middle ear of children as well as with the more complex classic audiometry techniques.

Source:
Chan J et al., Sci Transl Med. 2019; doi: 10.1126/scitranslmed.aav1102