The Global Burden of Hearing Loss and Future Hearing Aids

The Global Burden of Hearing Loss and Future Hearing Aids

For the original blog post that began this discussion, please go here.

September 1
Authors: Alan Archer-Boyd & Saima Rajasingam.

The simultaneous discussions took place directly after Nick Lesica’s keynote talk on the global burden of hearing loss and the application of artificial intelligence (AI) to solving this problem. One discussion, chaired by Alan Archer-Boyd, covered hearing devices of the future, and the other, chaired by Saima Rajasingam focussed on hearing diagnostics and services of the future. Both specifically discussed solutions and ideas for low-income countries. This post is based on the recollections and opinions of the chairs.

Discussion 1: Hearing devices of the future – Overcoming barriers of stigma, logistics, costs and efficacy

The discussion began with the question of what a future hearing device would look like. The main options for future designs are the something similar to the behind-the-ear (BTE) designs most common among hearing-device manufacturers, and eye-glasses. BTE designs are relatively discrete, which can reduce stigma, an important factor in hearing-device uptake. In addition, they have been the dominant form factor in hearing device design for at least thirty years. However, their compact design potentially limits their ability to measure and obtain information from the listener, such as eye movement and visual information from cameras. Progress has been made in obtaining eye-position data from sensors in the ear (Hládek et al., 2018). Eye-glasses potentially have the physical space and form-factor to include many more sensors and obtain much more useful information about a listeners’ surroundings and listening intent than can current hearing devices.

Many big tech companies are staking part of their future success on the popularity of augmented reality. However, previous eye-glass hearing aids (e.g. RadioEar, 1970) have not been successful, due mainly to the difficulty users had putting them on, and Google’s augmented reality “Google Glass” product has had limited and variable success, due in part to privacy concerns (e.g. Motti et al., 2015). It remains unclear what form future hearing devices will take. What is clear is that future hearing devices will be much more connected to each other and other devices in the environment, in the so-called “Internet of Things”. They will also off-load more and more of their computational needs to connected smartphones and cloud-based systems.

Increased connectivity and cloud-based services require good mobile internet coverage in both urban and rural areas. Mobile coverage in low-income countries continues to improve, with a 7% coverage gap in 2020, down from 24% in 2014 (GSMA report, 2020). This suggests that location-based and cloud-computing services employed by future hearing devices will work well in most areas.

Future hearing devices are likely to shift from proprietary to general-purpose microchips. A major reason for this is that the processing latency (a major concern for hearing device manufacturers) of general purpose chips has decreased to be similar to that of the specialized chips used previously. This development will open up hearing-device algorithm development to many, possibly leading to greater innovation and a faster development process.

AI is also likely to speed up development of new algorithms in at least two ways. First, auditory models developed using machine learning (ML) techniques are edging closer to producing responses similar to human listeners in both classic psychoacoustic experiments from speech intelligibility in noise to sound source localization. Fully functioning models of human auditory perception will, among other outcomes, speed up algorithmic development by allowing engineers to quickly ascertain whether new systems will provide benefits to listeners across multiple acoustic scenarios, without initially going through the much slower process of developing an experiment and recruiting real listeners. These models could be extended to produce responses to higher-level percepts, such as listening comfort.

AI could also be crucial in developing fully integrated and automated fitting processes for new devices, and this point was also raised in the second discussion.

Hearing devices developed specifically for low- and middle-income countries should also take into account linguistic factors not common in high-income countries, such as tonal languages. This is one specific example of how hearing device developers should tailor their development process to the problems that need to be solved in low-income countries, rather than the current business model, in which hearing devices are developed for easily accessible high-income markets, and other countries receive older and obsolete versions of these devices.

Discussion 2: Hearing diagnostics and services of the future – Ensuring wide and equitable access to hearing healthcare

Participants highlighted that even with the development of affordable hearing aid technologies, ensuring uptake and accessibility in underserved populations remains a challenge.

A key area for development is training of mid-level workers (e.g. audiology technicians, healthcare workers) to deliver screening & educational interventions. Audiologists and ENT specialists are required, but task shifting to community health workers can help broaden the scope of healthcare.

Advocacy on governmental and community level is essential in creation of sustainable hearing care – it is clear that audiology “camps” or “missions” cannot provide continued care and may not always be culturally aware. This advocacy and engagement is required from all involved in the development of hearing technology & care if sustainable programs of hearing healthcare are needed. Although remote screening and self-fitting technologies will be an integral part of future hearing care, it is essential that treatable ear disease is detected and managed early – requiring access to a healthcare worker.

Immediate priorities are

  1. Access to affordable test technology: including calibration of devices and access to automated protocols. The development of good quality self-fitting hearing devices can also help free-up the limited human resources currently available.
  2. Development of approaches to shift attitudes towards hearing technology amongst underserved populations: access to culturally tailored resources for patients and healthcare workers, and interventions informed by research into attitudes towards hearing aids/care.
  3. Training of audiology technicians/mid-level workers to broaden access to a hearing care professional.

Where do we go from here?

Clearly, the task of alleviating the global burden of hearing loss is a large and multifaceted one, and in many ways we are just at the beginning. Above all, clinicians and technologists need to communicate with each other, and the potential users and service providers in order to achieve a solution that is fit for purpose. Nick Lesica has suggested that the burden of this work should not be left to private companies and governments – the research community should get involved and take their research outcomes out of the lab and into the real world. There are many challenges to doing this, not least the difficulty of developing research commercially while under pressure to publish. However, the benefit of doing this may be great indeed.

Further reading:

References

  • GSMA (2020). The state of mobile internet connectivity report 2020.
  • Hládek, Ľ., Porr, B., & Brimijoin, W. O. (2018). Real-time estimation of horizontal gaze angle by saccade integration using in-ear electrooculography. Plos one, 13(1), e0190420.
  • Motti, V. G., & Caine, K. (2015, January). Users’ privacy concerns about wearables. In International Conference on Financial Cryptography and Data Security (pp. 231-244). Springer, Berlin, Heidelberg.