The tool is designed to help clinicians identify a common type of ear infection that's often overlooked or misdiagnosed
Clinicians at UPMC and the University of Pittsburgh have developed an AI algorithm that can identify acute otitis media (AOM), one of the most common childhood infections.
While some 70% of children have an ear infection before their first birthday, those infections are hard to spot and are often misdiagnosed as fluid buildup. To identify AOM, clinicians need to peer into the eardrum and identify subtle signs of an infection, often a difficult task when dealing with an infant or small child.
To help clinicians make a better diagnosis, researchers created an AI tool that can analyze a video of a patient’s eardrum, taken by an otoscope connected to a camera.
“Acute otitis media is often incorrectly diagnosed,” Alejandro Hoberman, MD, a professor of pediatrics and director of the Division of General Academic Pediatrics at Pitt’s School of Medicine and president of UPMC Children’s Community Pediatrics, said in a press release. “Underdiagnosis results in inadequate care and overdiagnosis results in unnecessary antibiotic treatment, which can compromise the effectiveness of currently available antibiotics. Our tool helps get the correct diagnosis and guide the right treatment.”
“The eardrum, or tympanic membrane, is a thin, flat piece of tissue that stretches across the ear canal,” he added. “In AOM, the eardrum bulges like a bagel, leaving a central area of depression that resembles a bagel hole. In contrast, in children with otitis media with effusion, no bulging of the tympanic membrane is present.”
Hoberman and his team created the tool by studying more than 1,100 videos of the tympanic membrane in children who had visited a doctor for treatment between 20-18 and 2023. They used the videos to develop two AI models that can detect AOM by studying the features of the tympanic membrane, including shape, position, color, and transparency.
According to Hoberman, the AI tool has a 93% success rate in identifying AOM. That’s better than various studies that have put the success rate of physicians studying a patient’s ear at between 30% and 84%.
“These findings suggest that our tool is more accurate than many clinicians,” he said in the press release. “It could be a gamechanger in primary healthcare settings to support clinicians in stringently diagnosing AOM and guiding treatment decisions.”
Healthcare organizations across the country are developing hundreds of AI tools to aid clinicians, drawing on technology that can often analyze data more efficiently than the human eye. These tools can also store the data in the EHR, enabling clinicians to review their work, show the results to parents, and use the data for training.
Eric Wicklund is the associate content manager and senior editor for Innovation at HealthLeaders.
KEY TAKEAWAYS
AOM, or acute otitis media, is a common ear infection that occurs in roughly 70% of infants before their first birthday, but is often overlooked or misdiagnosed.
Researchers at UPMC and the University of Pittsburgh have developed an AI tool that can analyze videos of a patient’s eardrum and identify AOM with more accuracy than most clinicians.
The data can also be stored in the EHR for review, to show parents how their children are being treated, and to teach new clinicians how to identify and treat AOM.