Introduction: Tinnitus affects ~14% of the population, often causing distress and affecting quality of life. Despite its frequency in outpatient clinics, understanding of its pathophysiology and optimal management strategies remains poor, and there is limited guidance available to clinicians. Aim: To audit UK tinnitus care pathways to gain better understanding of the presentation, investigation, and management of patients with tinnitus. Methods: A collaborative multicentre retrospective audit was conducted nationally across ENT, Audiology and AVM services within the UK. Anonymised data were collected on adults presenting between August-October 2022. These included: demographics, medical history, tinnitus features, associated symptoms, examination findings, audiometry, investigation results, management, and outcomes. Results: Thirty-six sites submitted data for 1,136 patients. 824 patients (73%) were found to have some form of hearing loss on audiometric testing, with nearly two-fifths of these (38%) receiving hearing aid provision. The most common management strategies were self-management techniques (52%), relaxation techniques (26%), sound generators (5%) and masking devices (4%). 46 patients (4%) received CBT. Almost all patients (94%) had no tinnitus questionnaires used for documenting the severity of their symptoms, and only two patients in the whole cohort had follow-up questionnaires used to monitor progression or improvement. Conclusions: Previous reviews of tinnitus practice have relied on questionnaires sent to departments rather than reviewing granular data in hospital records. This work provides analysis of a large group of patient journeys through tinnitus services. It highlights a gap in monitoring and follow-up of these patients and will be useful in shaping future pathways.

Introduction Ensuring that patients are well-informed in making health decisions has become increasingly pressing, particularly in light of resource constraints faced by the NHS. The emergence of artificial intelligence (AI) and natural language processing technologies, such as ChatGPT, offers potential solutions for delivering accessible patient information. This study explores the application of ChatGPT as a patient information tool, focusing on patients undergoing Functional Endoscopic Sinus Surgery (FESS) in the UK. Methods To evaluate the effectiveness of ChatGPT in providing patient information, the authors devised three common patient queries related to FESS. These questions were presented to both ChatGPT and three authors (including validation by a supervising Consultant) to generate a 150-word response. 20 qualified clinicians were blinded to responses and subsequently completed a 5-point Likert scale questionnaire to evaluate each response. Results When comparing mean scores between author and ChatGPT responses, it was found that there was no statistically significant difference for Accuracy, Completeness, Clarity or Appropriateness in any of the 1-3 questions asked. When explaining FESS, ChatGPT responses scored ≥50% on accuracy, clarity and appropriateness. ChatGPT responses scored lower in all areas when asked to described the alternatives to surgery. When explaining the risks of surgery, ChatGPTs responses scored higher on average. Conclusions This study establishes a foundational assessment of ChatGPT’s potential utility as a source of patient information within UK ENT departments. Notably, the study finds no significant disparities in the evaluations of accuracy, completeness, clarity, and appropriateness between ChatGPT-generated responses and those crafted by medical experts.