Does artificial intelligence have a role in telehealth screening of ear disease in Indigenous children in Australia?

Abstract

Background: Artificial intelligence models have been used to successfully detect ear disease from video-otoscopic images. Artificial intelligence is highly context specific, and to date there are no studies looking specifically at classifying ear disease amongst Aboriginal and Torres Strait Islander children. The aim of this study was to establish proof-of-concept of both the feasibility and effectiveness of using artificial intelligence to detect ear disease in a telehealth ear screening service for Indigenous children. Methods: Labelled video-otoscopic images from the Health-e-Screen4kids service located in Cherbourg, Queensland were used to train an artificial intelligence binary classifier model of ear disease. The model uses video-otoscopic images as input and outputs a binary prediction of whether the ear is normal or abnormal. The accuracy of the resultant model was assessed by comparing the output of the model to the image labelled by an Ear, Nose, and Throat (ENT) specialist. Results: The artificial intelligence model achieved an overall accuracy of 80.99% on the validation set (images used for training the model) and 78.90% on the test set (images not used for training the model), with a sensitivity of 77.46% and a specificity of 80.46%. Conclusions: The application of artificial intelligence models based on machine learning to classify ear disease amongst Indigenous children is feasible and can achieve an accuracy of more than 80%. The model has not been externally validated. The performance of this model is unlikely to exceed that of an ENT specialist and is therefore more likely to be useful for other health disciplines that have a key role in the delivery of primary care in Indigenous communities. The findings from this study are encouraging of further research and development of artificial intelligence models for the detection of ear disease amongst Aboriginal and Torres Strait Islander children.