DR AL-RAHIM HABIB has helped train an AI tool to detect ear conditions, using more than 10,000 ear images from over 4,000 children.
Aboriginal and Torres Strait Islander children in Australia have some of the highest rates of ear infections in the world. Indigenous children are three times more likely to develop chronic otitis media and otitis media with effusion, also known as glue ear, compared to non-Indigenous children.
This increased risk is linked to factors such as overcrowded living conditions, limited access to healthcare, and socio-economic disparities. Hearing loss during childhood, a critical stage for development, can affect speech, learning, behaviour, social skills, and future job opportunities.
In rural and remote parts of Australia, healthcare workers and nurses play a critical role as the primary point of contact for ear disease screening, triage, and prevention programs. However, access to otolaryngology services in these areas can be challenging. Telehealth services are helping to bridge this gap by providing more accessible options for screening and specialist consultations.
One effective approach is the “store-and-forward” tele-otology model. This model relies on mobile screening units equipped with advanced tools such as digital video otoscopes and audiometric testing equipment. These tools allow healthcare workers to perform detailed examinations on-site and send the results to specialists in metropolitan areas for review. This model reduces the need for patients to travel long distances and enables timely remote consultations, significantly improving access to care for children and families in underserved areas.
While telehealth models like this provide access to ear exams, they also face challenges. Delays can occur between the time data is collected and when specialists provide treatment recommendations. Managing these delays and ensuring smooth communication between healthcare providers in different locations can be difficult. Helping frontline healthcare workers and nurses to better assess and prioritise cases during initial checks is a key step toward solving these problems, especially in rural areas.
Artificial intelligence (AI) is a powerful tool capable of performing tasks traditionally requiring human intelligence, such as analysing data and identifying patterns. It is increasingly becoming an important part of modern healthcare systems. It uses methods such as machine learning, which relies on pre-determined features to train systems for making predictions, and deep learning, which employs multiple layers of artificial neural networks to automatically extract features, recognise complex patterns, as well as generate predictions.
Computer vision, a part of AI that focuses on understanding visual information, has been used to analyse various forms of medical imaging in radiology and ophthalmology. AI has the potential to transform how ear disease is screened and triaged in rural and remote areas.; for example, AI systems trained on otoscopic images to detect pathology, classify severity, and support healthcare workers in making decisions at the frontline.
The DrumBeat.ai project is a collaborative initiative involving researchers and clinicians from the University of Sydney, Royal Darwin Hospital, Westmead Hospital, Queensland Children’s Hospital, and Microsoft’s AI for Good Research Lab, each bringing unique expertise to improve ear health care in remote areas.
The goal of the project is to use AI to enhance ear health care in underserved regions by enabling faster, more accurate screening and triage of ear diseases. DrumBeat.ai leverages advanced AI algorithms to analyse otoscopic images, identifying conditions such as acute and chronic otitis media, and glue ear.
A typical scenario might involve a healthcare worker in a remote clinic using a digital video otoscope to capture an image of a child’s ear during a routine checkup. This image is instantly analysed by DrumbBeat.ai. Within seconds, the model provides a prediction, classifying the type of otitis media, its severity, and whether the case requires an otolaryngologist review. For example, the AI model can detect chronic suppurative otitis media and classify it as needing referral to an otolaryngologist. This immediate feedback empowers the healthcare worker to make informed decisions about care, streamlining the triage process and ensuring that children who need can be flagged for review.
The next phase of the DrumBeat.ai project is to refine, test and integrate this AI tool into daily clinical workflows. By incorporating feedback from healthcare providers and collaborating closely with specialists in otolaryngology and audiology, DrumBeat.ai aims to become an integral part of clinical workflows. These efforts will help overcome barriers to accessing expert care in remote areas by streamlining the triage process. If you’re interested in learning more, please visit www.drumbeat.ai.
About the Author: Dr Al-Rahim Habib is an otolaryngology – head and neck surgery registrar from Brisbane with a PhD in artificial intelligence.
