OTITIS MEDIA VOCABULARY AND GRAMMAR

Area of Science:

• Medical Informatics
• Otolaryngology
• Artificial Intelligence

Background:

• Accurate diagnosis of otitis media in children is challenging, often leading to antibiotic overprescription.
• Distinguishing acute otitis media from otitis media with effusion is critical for appropriate treatment.

Purpose of the Study:

• To develop an automated algorithm for classifying diagnostic categories of otitis media.
• To improve diagnostic accuracy and guide appropriate antibiotic use in pediatric otitis media cases.

Main Methods:

• Developed an "otitis media vocabulary" based on visual cues used by otoscopists.
• Designed an "otitis media grammar" to combine vocabulary terms mirroring clinical decision-making.
• Implemented an automated classification algorithm using this framework.

Main Results:

• The algorithm achieved 84% classification accuracy for otitis media diagnostic categories.
• Performance was comparable to or exceeded that of untrained clinicians and existing state-of-the-art classifiers.

Conclusions:

• The proposed automated algorithm offers a reliable method for classifying otitis media.
• This technology has the potential to reduce unnecessary antibiotic prescriptions by improving diagnostic precision.