DrumBeat.ai: Addressing Paediatric Indigenous Ear Disease in Rural and Remote Australia Using Artificial Intelligence

Abstract

Ear disease is an important public health issue for Aboriginal and Torres Strait Islander children. Chronic ear disease can lead to long-term hearing loss. The consequences of long-term hearing loss include impacts on speech and language development, academic performance, behaviour, social skills, and future employment opportunities. Artificial intelligence (AI) is an emerging technology that has the potential to transform ear disease screening and triage.

This thesis is comprised of an introductory chapter (Chapter 1), a systematic review and meta-analysis of machine learning techniques for ear disease classification (Chapter 2), a comprehensive overview of a telehealth service evaluating ear disease in rural and remote areas (Chapter 3), an assessment of inter-rater agreement in ear disease diagnoses using a telehealth approach (Chapter 4), an exploration of the performance and generalisability of image classification algorithms for otoscopy (Chapter 5), an evaluation of a supervised image classification system to triage otoscopic images from Aboriginal and Torres Strait Islander people in rural and remote areas (Chapter 6), and a discussion regarding the significance of findings, clinical and public health implications, strengths and limitations, and future directions (Chapter 7).

This thesis provides a framework illustrating an application of AI to address an important clinical need. The implications of this research extend to a diverse audience, including Aboriginal and Torres Strait Islander communities, healthcare professionals involved in ear disease screening and triage, otolaryngologists engaged in telehealth programs and outreach services, data scientists exploring applications of AI in healthcare, and policymakers evaluating innovative solutions to enhance healthcare delivery in resource-constrained settings.