This could help reduce the unnecessary use of antibiotics in young children
Scientists developed a application for mobile phones which can accurately diagnose its infections earsuch as or acute otitis media (OA). This app could help reduce the unnecessary use of antibiotics in young childrenaccording to new research published in the scientific journal “JAMA Pediatrics”.
As it says medicalexpress acute otitis media is one of the most common childhood infections for which antibiotics are prescribed. The new app, which uses artificial intelligence, makes the diagnosis by evaluating a short video of the eardrum recorded by an otoscope connected to the mobile phone’s camera.
“Acute otitis media is often misdiagnosed”said senior author Alexandro Hoberman, professor of pediatrics and director of the Division of Pediatrics at the University of Pittsburgh School of Medicine (UPMC).
“Underdiagnosis leads to inadequate care and overdiagnosis to unnecessary antibiotic treatment, which can compromise the effectiveness of antibiotics. The tool helps us get the right diagnosis and guides the right treatment,” noted the researcher.
According to Hoberman, about 70% of children develop otitis before their first birthday. Although this condition is common, the accurate diagnosis of OM is more difficult as it is often confused with otitis externa or otitis media with fluid, a condition that is generally not caused by bacteria and does not benefit from antimicrobial therapy.
To develop a practical tool to improve the accuracy of diagnosing OMMO, researchers began by creating of an educational library of 1,151 tympanic membrane videos from 635 children who visited UPMC pediatric outpatient clinics between 2018 and 2023. Two trained experts reviewed the videos and rated the cases.
“The ear drum is a thin, flat piece of tissue that runs along the ear canal,” Hoberman explained. “In OM, the eardrum swells like a bagel, leaving a central area that looks like a hole. “In contrast, in children with otitis excretory, we don’t have swelling of the tympanic membrane,” he added.
The researchers selected 921 videos to train two different artificial intelligence models to be able to detect OM by looking at the characteristics of the tympanic membrane, including shape, position, color and transparency. They then used the remaining 230 videos to test the models’ performance.
It was found that both models showed high accuracy and sensitivity over 93%, meaning they had low false negative and false positive rates. According to Hoberman, previous studies of clinicians have reported diagnostic accuracy of OMMO ranging from 30% to 84%, depending on the type of health care provider, level of education and age of the children being tested.
“These findings suggest that our tool is more accurate than many clinicians’,” Hoberman said.
“Another advantage of our tool is that the videos we capture can be stored in a patient’s medical record and sent to other providers”added the researcher.
The scientists hope their technology could soon be widely used to improve accurate diagnosis of the condition and support treatment decisions.
