New Diabetes Treatment Study: Hope Ahead

A New Study Aims to Train AI to Help Manage Hearing Loss and Tinnitus Referrals

For the estimated 7 million people in the UK with tinnitus and 11 million with hearing loss, access to specialist care is often a long and frustrating wait. A new pilot study, led by the Royal Cornwall Hospitals NHS Trust, is testing whether an artificial intelligence (AI) system can help clear the backlog. The trial, titled “Artificial Intelligence for the Automated Diagnosis, Triage, and Assessment of Patients With Hearing Loss and Tinnitus,” will see if a trained AI can match a clinician’s ability to recommend the next best step for a patient—such as discharge, imaging, or a face-to-face appointment.

The Bottleneck in Hearing and Tinnitus Care

Long waiting lists for ear, nose, and throat (ENT) services are a significant problem across the UK. Tinnitus alone generates over one million GP appointments annually, and referrals to hospitals can result in delays exceeding a year for a first assessment. This situation worsened after the COVID-19 pandemic. In response, the Royal Cornwall Hospitals NHS Trust developed a virtual ENT clinic. Patients take a validated in-person hearing test and complete online questionnaires. A clinician reviews the data and decides on management without an initial physical consultation. This model successfully reduced waiting times, but it created a new bottleneck: clinicians must still manually review every single case.

How the AI Clinical Trial Will Work

This observational study builds directly on the virtual clinic model. When a patient is processed through the standard virtual clinic, their anonymized data—including hearing test results and questionnaire responses—will also be analyzed by an AI system. The AI will be trained using explainable AI methods, a type of machine learning designed to provide clear reasoning for its decisions, not just a black-box recommendation.

The core of the study is a comparison. For each of the 1,500 participants, the AI will generate an outcome recommendation (e.g., discharge, refer for MRI, schedule face-to-face consultation). Separately, a clinician will review the same case and make their standard decision. The research team will then measure the concordance rate—how often the AI’s suggestion matches the clinician’s. Crucially, the treating clinician will not see the AI’s recommendation during their assessment, ensuring the final patient care decision remains a human one.

Who Can Participate in This Research?

The study will recruit exclusively from the pool of patients already being assessed in the Royal Cornwall Hospitals’ virtual hearing loss and tinnitus clinic. The main inclusion criteria are straightforward: patients must be assessed in that clinic, have symptoms of hearing loss and/or tinnitus, and be able to provide informed consent.

The exclusion criteria are designed to ensure data quality and ethical compliance. Individuals under 18, those unable to consent, patients without sufficient English proficiency where translation cannot be arranged, and cases with insufficient data for analysis will not be included.

The Goal: An AI Assistant, Not an AI Replacement

The researchers are clear that the AI tool is intended as a clinical recommendation assistant. If the study shows a high level of agreement between the AI and clinicians, the tool could be integrated into the virtual clinic workflow. In practice, this would mean the AI could screen and provide initial recommendations for a large volume of straightforward cases. A clinician would then review and sign off on these recommendations rapidly, allowing them to devote more time and attention to complex cases that require nuanced human judgment.

The potential benefits are two-fold: accelerating access to care for all patients by streamlining triage, and improving job satisfaction for clinicians who can focus on the work that most requires their expertise.

Current Status and Future Implications

As of now, the trial’s status is listed as NOT_YET_RECRUITING. The team is preparing to enroll the target of 1,500 participants. This is a pilot study, meaning its primary aim is to gather evidence on whether the AI approach is viable and accurate compared to the current gold standard of clinician review.

For the field of hearing health, a successful outcome could provide a scalable model for managing common yet burdensome conditions like hearing loss and tinnitus. It represents a pragmatic use of technology to address a systemic healthcare challenge: overwhelming demand for specialist services. For patients, the hope is that such tools, working under clinician supervision, could mean shorter waits, faster reassurance, and quicker access to appropriate treatment or support.

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

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