A third of patients who black out suddenly are initially misdiagnosed, delaying treatment for conditions like epilepsy or dangerous heart rhythms. Transient loss of consciousness (TLOC) accounts for a large share of emergency visits, yet current care is inconsistent and often wrong. The research team has already built artificial intelligence tools that can interpret symptoms, spoken descriptions, and ECG readings, but none have been tested in real clinical settings. This project lays the groundwork for a future trial by mapping current practice, running workshops with patients and clinicians to define what an optimal AI-assisted pathway should look like, and modelling the costs of standard versus AI-supported care. If the pathway proves effective, it could reduce misdiagnosis, cut unnecessary hospital admissions, and ensure patients get the right specialist—cardiologist or neurologist—faster. The work itself produces three concrete outputs: a scoping review, workshop findings, and a health-economic model, all of which will be published and shared through patient networks.
View original technical description
Development work questions Members of the research team and others have developed a number of Artificial Intelligence (AI) solutions which could help more patients with Transient Loss of Consciousness to get correct advice and treatment more quickly. These include computer applications capable of interpreting TLOC symptoms, spoken TLOC descriptions and ECG interpretation. However, none of these have been tested in clinical practice. The integration of AI into the early management of TLOC is a complex intervention at the emergency / primary / specialist care interface requiring patient & medical expert consensus on what constitutes an optimal pathway and how AI can help to deliver it. Background Transient Loss of Consciousness (TLOC) is a common reason for unplanned presentations to emergency departments or primary care. Three causes cause over 90% of TLOC presentations (syncope, epilepsy, functional seizures) and require different investigations and treatments. Some patients only require advice, others need admission and urgent tests or specialist referral to cardiologists or neurologists. Guidelines for TLOC management have existed for > 10 years but there is great variability of care provision and ? of first diagnoses are incorrect, causing avoidable morbidity, mortality and ineffective use of healthcare resources. Aims and Objectives This project will provide the basis for a NIHR PGfAR project comparing the clinical and cost effectiveness of an optimised, AI-assisted early TLOC management pathway with current standards of care. Three interlinked work packages (WPs) aim to: (1) describe current practice, (2) identify the needs and expectations as well as the ideas about the use of AI of patients and professionals; (3) estimate the cost of standard and AI-assisted care delivery and the potential of a future clinical effectiveness trial to deliver value for money. Development work plan The completion of a scoping review of current practice (WP1) will inform the topic guide for six Knowledge Exchange (KE) workshops (3 with members of three patient partner organisations, 3 with a range of professionals involved in TLOC care) (WP2). Both review and KE workshops will feed into the health economic work including the calculation of the costs of delivery of standard or AI-assisted TLOC care, a Value-of-Information (VoI) analysis of a future Randomized Controlled Trial and a Rapid Assessment of Need for Evidence (RANE) (WP3). Timelines for delivery We will complete: the scoping review by month 4, the KE workshops by month 10. the analysis of KE workshops and the health economic work in month 16 The completed works will be summarised in a "Next Steps" in month 16. Impact and Dissemination The ultimate goal of this project is to provide the basis for a future clinical effectiveness study of an AI assisted TLOC pathway, it will produce three independent valuable outputs which will be published in scientific journals and communicated via the networks of our patient partners: A scoping review of current TLOC care, a description of the outcomes of the KE workshops about an AI assisted TLOC pathway and the health economic modelling of standard and AI-assisted care.
Plain English summaries and category classifications on this site are generated by AI and may not perfectly reflect the original research.
Is something wrong? Let us know