Decaf Helps Slow Parkinson’s Progression

AI Steps into the Hearing Clinic

More than one million general practitioner appointments in the UK each year are for tinnitus, a figure that starkly illustrates the immense pressure on hearing health services. For patients, a referral to a hospital specialist often means waiting over a year. A new observational study led by the Royal Cornwall Hospitals NHS Trust aims to test whether artificial intelligence can help clear this backlog. The trial, titled “Artificial Intelligence for the Automated Diagnosis, Triage, and Assessment of Patients With Hearing Loss and Tinnitus,” will assess if an AI can safely and accurately replicate clinician decisions, potentially freeing up specialists to focus on the most complex cases.

A Virtual Clinic’s Search for Greater Efficiency

The research builds directly on an existing innovation at the Royal Cornwall Hospitals NHS Trust: a virtual ear, nose, and throat clinic. In this model, patients first complete a validated in-person hearing test and online questionnaires. A clinician then reviews the data remotely to decide on the next step, which could be discharge with advice, a request for imaging, or a face-to-face consultation. Initial results showed this approach could manage most patients virtually, significantly cutting wait times. However, the system still requires a clinician to review every single case, consuming time that could be spent on patients needing direct, hands-on care.

This new study introduces an AI component designed to support, not replace, clinical judgment. The core question is whether an AI system can analyze the same patient data—audiograms and questionnaire responses—and produce a triage recommendation that matches what a clinician would decide.

How the AI Clinical Trial Will Work

This is a pilot comparative study with a planned enrollment of 1,500 participants. All participants will be patients already being assessed in the Royal Cornwall virtual hearing loss and tinnitus clinic. They must be 18 or older, have symptoms of hearing loss or tinnitus, and be able to provide informed consent. The study excludes individuals who cannot consent or who lack sufficient English proficiency where translation is unavailable.

The intervention is the AI system itself. Using “explainable AI” methods, the system will be trained on historical virtual clinic data to learn the patterns of clinician decision-making. For each new patient in the study, the AI will generate a recommendation. Crucially, the managing clinician will not see the AI’s suggestion; they will make their independent assessment based on standard protocol. Later, researchers will compare the AI’s output against the clinician’s “gold standard” decision to measure their level of agreement.

The trial status is currently NOT_YET_RECRUITING. The primary goal is to measure the concordance between AI and clinician recommendations.

The Potential Impact on Patients and Hearing Healthcare

Hearing loss affects an estimated 11 million people in the UK, and tinnitus impacts around 7 million. Both conditions are strongly linked to reduced quality of life, mental health challenges, and employment difficulties. The demand for services far outpaces supply, a problem worsened by the COVID-19 pandemic backlog.

If the AI tool demonstrates a high degree of alignment with clinician judgment, the proposed next step is to integrate it as a clinical recommendation assistant. In practice, this could mean the AI provides a first-pass analysis of patient data, flagging straightforward cases for rapid processing and highlighting complex cases for immediate clinician attention. The goal is a more efficient triage system that accelerates access to appropriate care for everyone. Patients with simple cases might receive management advice more quickly, while those with red flags or complicated presentations could see a specialist sooner.

This study represents a pragmatic approach to using AI in medicine. It does not seek autonomous diagnosis but aims to create a tool that augments human expertise, addressing a specific systemic bottleneck in hearing healthcare. The use of explainable AI is also critical, as it aims to provide clinicians with the system’s reasoning, fostering trust and allowing for informed oversight.

Source:
AI-assisted Diagnosis, Triage and Assessment of Hearing Loss and Tinnitus (ClinicalTrials.gov: NCT07586098)

This article is for informational purposes only. Consult a qualified professional for personalised advice.