As many parents of young children know, Acute Otitis Media (AOM), a type of ear infection, can be a problem ranging from minor to serious. While easily treatable, the bigger ongoing problem with AOM among clinicians is it’s often incorrectly diagnosed.
Nationally, nearly 50% of ear infection diagnoses are wrong, and unfortunately, even experienced pediatricians are only accurate in an AOM diagnoses about 70-80% of the time.
Physicians need an easy fix to this problem; simple, reliable decision support to improve the AOM diagnosis, reduce unnecessary antimicrobial therapy, and lower the nearly $4 billion in direct and indirect costs of the 20 million prescriptions written every year to treat AOM.
In response to this need, Alejandro Hoberman, MD, alongside Nader Shaikh, MD, and Jelena Kovačević, PhD, Dean of the NYU Tandon School of Engineering, developed The Pitt-CMU iTM app, a tool used in the automated diagnosis of AOM.
Using just an iPhone and a standard otoscope, physicians can now more accurately diagnose AOM. The handheld app uses voice recognition to capture images and videos that are processed within the app’s algorithm on the Google cloud. The AI then processes those images and videos and gives results in a binary outcome, ear infection vs. no ear infection. Compared to generic classifiers (70% accuracy) and experienced pediatricians (80% accuracy), the app produces 90% accuracy when diagnosing AOM.
The fully developed app has the potential to wholly resolve the decades-long problem of incorrect and over-diagnosis of AOM in pediatric populations, easing the burden on parents, providers, and most of all, the patients.
Find out more about Hoberman’s ongoing research and clinical efforts at the Department of Pediatrics website.