Acute otitis media (AOM) is the most common bacterial infection in young children and a major contributor to healthcare utilization and costs. Diagnosis relies on otoscopy and identification of tympanic membrane (TM) bulging, yet physician-performed otoscopy has variable diagnostic accuracy and depends heavily on clinical expertise.
Digital otoscopy combined with artificial intelligence (AI)-based image analysis has demonstrated high diagnostic accuracy in previous studies. In a recent study by our group, digital otoscopy with an automated neural network classifier reliably ruled out otitis media in children.
The PARENT-EAR randomized controlled pilot trial aims to evaluate whether parent-performed digital otoscopy at home, combined with retrospective AI analysis of captured TM images, can reduce physician visits due to suspicion of AOM. The study also assesses feasibility, diagnostic accuracy of parents after training, and potential effects on antibiotic use and healthcare utilization.
Methods and outcomes The PARENT-EAR trial is a randomized controlled pilot trial conducted at Vaasa Central Hospital and Turku University Hospital. Recruitment will occur between September and December 2026 (extended if necessary), with follow-up ending May 31, 2027 (5-9 months per participant).
Children aged 12 to 47 months with at least one AOM episode in the past 12 months will be randomized 1:1 using permuted block randomization. Exclusion criteria include tympanostomy tubes, tympanic membrane perforation, middle ear effusion at recruitment, immunosuppression, previously diagnosed hearing loss or anatomical abnormalities predisposing to AOM, and otoscopy experience of the parent.
Intervention group:
Parents receive a digital otoscope (The Ear Penguin by Sibbo Medical Devices, MDR EU Class I device) and training on TM image interpretation. They complete a TM image classification quiz before and after the training. During follow-up, parents perform home otoscopy when suspecting AOM, capture TM images/videos, interpret findings, and record symptoms in a diary. If AOM is suspected or the child appears ill, medical evaluation is instructed.
Control group:
Standard care without home otoscopy. Parents complete the same symptom diary and seek care according to usual practice.
Statistical analyses will compare outcomes between groups. AI-based TM classification will be compared against expert panel evaluation.
This is a pilot study with a target sample size of 20 children per arm, limited by device availability.
Ethical aspects, data collection and handling The study will be conducted in accordance with the Declaration of Helsinki. Written informed consent will be obtained from all parents after provision of oral and written information. Participation is voluntary, and families may withdraw at any time. Investigators may discontinue participation if deemed in the child's best interest.
The digital otoscope is a CE-marked Class I medical device used solely for visual inspection of the tympanic membrane. Pneumatic pumps are not provided. Potential risks include minor ear canal abrasions or discomfort; serious adverse events are considered rare. Any adverse events will be managed according to standard clinical practice and reported to relevant authorities when required. The study is investigator-initiated. The device manufacturer provides the otoscopes but has no role in study design, data access, or analysis.
Data will be collected via paper symptom diaries, monthly REDCap questionnaires (University of Turku), and national electronic health records (Kanta). TM images and videos will be securely stored and imported into a research database with relevant metadata.
All data will be handled in accordance with the EU General Data Protection Regulation (GDPR 2016/679). Study data will be pseudonymized and identified only by study ID numbers. The key linking personal identifiers to study IDs will be stored separately and securely. Paper diaries will be destroyed after digitalization. Consent includes permission to access health records and to use collected images and data for scientific purposes. Results will be published in peer-reviewed journals and as part of a doctoral thesis.
