A portable, easy-to-use tympanometer developed by Duke University clinicians and engineers holds promise for allowing teachers, school nurses, and frontline healthcare providers to conduct school- and community-based screening to identify children in need of ear and hearing care.
While studying childhood hearing loss in remote and rural areas of Alaska, trained audiologist Dr Samantha Robler from Norton Sound Health Corporation and now Adjunct Associate Professor of Global Health Susan Emmett from Duke identified a gap in the screening tools that could be used to identify children at risk of hearing loss.
The most reliable tool to evaluate the state of the area immediately behind the ear drum, also referred to as the middle ear, is a tympanometer. It is used to identify middle-ear diseases such as ear infections, perforated ear drums and other conditions. But it is expensive and require trained audiologists, like Robler, to use and interpret the results. And, therefore, rarely included in screening programmes.
“Tympanometry is an essential tool for evaluating infection-related hearing loss, but a mobile and lay-friendly tympanometer to support mass screening programs just doesn’t exist,” said Robler.
Seeking to address this problem, Robler and Emmett approached Mark Palmeri, Professor of the Practice in the Department of Biomedical Engineering (BME) at Duke’s Pratt School of Engineering and Director of the BME Design Fellows programme, about developing an innovative solution to make tympanometry accessible, affordable and easy to use.
The assembled team brought together clinicians, engineers, data scientists, and Alaska-based Norton Sound Health Corporation, to build a solution. And to jump start the work, they received a Duke MEDx (Medicine + Engineering) grant in 2020.
“MEDx has helped to find a home for this project through seed funding,” said Palmeri. “The project truly has been a labour of love and through the support of MEDx, we were able to get this project off the ground.”
Palmeri and his team, including Duke graduate students and undergraduate BME Design Fellows, first developed a novel mobile health (mHealth) tympanometer with a machine learning interface that the team believes will have broad applicability and the potential to improve diagnoses in community health settings.
Machine learning is an integral part of this platform as it allows a layperson to acquire the data and then have the device interpret the data to guide clinical decisions.
With additional funding from the Duke Global Health Institute, MD-PhD students Felix Jin and Ouwen Huang, from the Medical Scientist Training Programme, took on the next part of the project developing and evaluating the use of machine learning in automating the mHealth tympanometer, under Palmeri’s guidance.
The team then developed a hardware prototype that could be used with a smartphone. Duke student engineering teams demonstrated “proof-of-concept” of the prototype, showing that it meets the requirements of low cost and ease of use, is comfortable when placed in a child’s ear, and works quickly enough to be an effective screening tool.
“This project has given me and my peers an opportunity to use our individual skills and watch the project come together,” said Alexander Chen, a Duke BME undergraduate student.
The multidisciplinary team is now working on the subsystems of the device to fine-tune specifications and find creative ways to drive down cost. By the end of 2022, the team expects to have the device ready for beta testing and aims to incorporate the device in clinical trials in 2023.
“Right now, there isn’t a tool that is available and competitive in the market like the tympanometer we have developed,” said Palmeri. “COVID at-home testing has really accelerated the desire for these accessible and cost-effective tests. The long-term goal would be to implement this in our communities and clinically, and then be able to offer these screening tests as an option for parents and caregivers to purchase and use at home.”
The team believes that the mHealth tympanometer could benefit children across the world and ensure they are equally able to reach their full potential.
“Our innovative approach that combines low-cost mHealth technology with machine learning represents a great leap forward to address childhood hearing loss in underserved populations globally,” added Emmett. “It is exciting to see this new collaboration, supported by MEDx, make significant progress on such an important public health issue.”
Adapted by Karl Bates from New mHealth tympanometer developed by Duke researchers could address preventable hearing loss in children.