This smartphone app uses AI to detect ear infections

A promising new tool has recently emerged in the medical field, targeting ear infections in young children. It's a phone app that uses artificial intelligence to accurately diagnose ear infections, or acute otitis media (AOM). This technological breakthrough could not only improve the diagnosis of this common childhood infection but also help to reduce unnecessary antibiotic prescriptions, a significant concern in the current context of increasing antimicrobial resistance.

Indeed, acute otitis media is common among young children, affecting nearly 70% of them before their first birthday. However, its accurate diagnosis can often be challenging due to similarities with other ear conditions, such as otitis media with effusion, which does not require antibiotic treatment. This confusion can lead to unnecessary antibiotic prescriptions, thus compromising the effectiveness of these drugs.

Faced with the challenge of differentiating between the two types of ear infections, physician-researchers from the University of Pittsburgh and UPMC have developed a smartphone app that leverages artificial intelligence. This app assesses a short video of the eardrum captured by an otoscope attached to the phone's camera, providing a simple and effective way to diagnose AOM. This method could prove to be more accurate than traditional diagnosis by qualified clinicians.

Dr. Alejandro Hoberman, the study's lead author, emphasizes the importance of this tool in combating unnecessary antibiotic prescriptions. Indeed, incorrect identification of AOM can lead to inadequate care or unnecessary antibiotic treatment. Thanks to this app, it is possible to achieve an accurate diagnosis and direct treatment appropriately.

To develop this tool, the researchers compiled a library of over 1,000 videos of the tympanic membrane from children who visited pediatric clinics. These videos were used to train two artificial intelligence models to recognize the distinctive features of AOM. The results were promising, with the models' performances significantly exceeding those of traditional clinicians.

Beyond its diagnostic utility, this app also allows for storing the videos in the patient's medical record, thus facilitating information sharing between healthcare professionals and parents. Additionally, it serves as a valuable educational tool for medical students and residents, helping them to understand and interpret diagnoses.