A human-like AI avatar will read your heart monitor results aloud and explain them in plain English, not medical jargon. The project, called DCODE, builds a large language model that can automatically interpret a single-lead Holter ECG report and deliver the explanation through a realistic digital avatar. This matters because patients often cannot understand their own ECG results. The graphs are complex, the terminology is technical, and the confusion can cause unnecessary anxiety or lead people to miss early signs of heart trouble. In 2024, over 428,000 people were on NHS waiting lists for cardiac care in England and Northern Ireland, showing how strained cardiology services already are. If DCODE succeeds, consumers will receive a clear, empathetic explanation of their heart data without needing a clinician to interpret it for them. The project will produce a demonstrator to be trialled with local clinicians, who will check that the AI does not generate false positives or false negatives that could alarm or mislead. The direct-to-consumer offering could help people manage their cardiovascular health more effectively and reduce pressure on healthcare services.
View original technical description
DCODE is an Ai Consumer Health project looking to develop a Large Language Model (LLM) that will enable to automatic explanation of Electrocardiograms for the consumer. We're pleased to be building DCODE in the Coventry and West Midlands area as we have excellent support in the creative sector and great outreach to the local healthcare services. In 2024, over 428,000 people were on NHS waiting lists for cardiac care in England and Northern Ireland, reflecting significant demand for cardiology services and challenges in meeting growing patient needs. Consumers often struggle to understand electrocardiograms (ECGs) because they are complex medical tests that produce detailed graphs of the heart's electrical activity. The technical nature of the results, filled with intricate waveforms and medical jargon, can be confusing for those without a medical background. This lack of understanding can impact consumers by causing unnecessary anxiety or misinterpretation of their heart health status. It may also hinder effective communication with healthcare providers, leading to misinformed decisions about treatment options or adherence to medical advice. The inability to comprehend ECG results can prevent individuals from recognizing early signs of heart issues, potentially delaying crucial interventions. Ultimately, this gap in understanding can negatively affect a person's ability to manage their cardiovascular health effectively. Understanding ECG is a complex art and explaining it to a consumer in a manner they may understand is complex. By using a Large Language Model and understanding the sequence and impact of specific ECG terminology the project seeks to develop a solution that can machine read in the outputs report of a single lead Holter style ECG and communicate the results to the consumer using AI derived avatars. We will provide the output of DCODE to our customers through our Direct-to-Consumer offering. Using a human Ai avatar to read the results of the Holter ECG to the consumer, providing them with a highly realistic human presentation and explanation of their data, is a novel and creative solution blending creative services, Ai and augmented reality. Humanoid AI avatars offer a powerful way to explain medical conditions by combining advanced technology with human-like interaction. Avatars simplify complex information using natural language and empathetic communication, enhancing understanding and trust. The output is a demonstrator that can then be trialled with local clinicians to ensure that the output is correct and does not contain errors or interpretations that may alarm through false positive or false negative.
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