AI-enabled stethoscopes show promise for improving diagnosis of cardiovascular conditions

A large-scale NHS study has tested the real-world impact of an artificial intelligence (AI)-enabled stethoscope in primary care.

The clinical trial, led by researchers at Imperial's National Heart and Lung Institute and Imperial College Healthcare NHS Trust, examined the effectiveness of introducing an AI-enabled stethoscope designed to help doctors identify early signs of serious heart conditions, such as heart failure, arrhythmias, and valve disease, in primary care clinics within the NHS. 

Published in The Lancet, the study found that introducing the technology didn’t significantly increase the overall number of heart failure diagnoses among all patients, primarily because many physicians did not use the device consistently in routine practice. However, when doctors used the AI stethoscope as intended, the technology was associated with substantially faster and more frequent detection of these cardiovascular conditions, indicating that the device performed well, but its real-world impact was limited by low uptake and workflow challenges.

Early detection of heart disease is a global health priority. Conditions like heart failure, irregular heart rhythms, and valve problems are common and treatable, but often only diagnosed after an emergency hospital visit. New technologies, such as AI-enabled stethoscopes, could help doctors detect these conditions earlier during routine primary care appointments. While these tools demonstrate strong technical performance and have regulatory approval, few have had a meaningful impact on everyday care. Studies that combine real-world testing, randomised trials, and practical implementation are needed to determine whether the technology works and how it can be used effectively in busy healthcare settings.

The elusive combination of frontline care, randomisation and real-world data in this study is essential for patients, health systems, industry, regulators and researchers to realise impact from the most promising health innovations.  

Professor Nick Peters, senior author and Professor of Cardiac Electrophysiology at Imperial's NHLI, said "The elusive combination of frontline care, randomisation and real-world data in this study is essential for patients, health systems, industry, regulators and researchers to realise impact from the most promising health innovations.  Sadly, failure to achieve this is the reason why too many effective innovations fail to make it into routine patient care."

The TRICORDER trial is the first cluster randomised controlled implementation trial (RCIT) of a clinical AI technology on a national scale, involving 205 NHS GP practices and more than 1.5 million registered patients in the UK. The AI stethoscope integrates three algorithms designed to detect heart failure, irregular heart rhythms, and valve disease during routine appointments. Over the 12-month study period, doctors performed nearly 13,000 AI-assisted cardiac examinations. While the overall heart failure detection rates were similar between the groups (1,342 new cases in the AI group versus 1,984 in usual care), subgroup analyses showed that patients examined with the AI stethoscope had significantly higher detection rates – nearly twice as many new heart failure cases and three times as many detections of irregular heart rhythms – compared to patients who were not examined with the AI device, suggesting that the increased detection was associated AI stethoscope examinations.

 The research was supported by the NIHR Imperial Biomedical Research Centre, a translational research partnership between Imperial College Healthcare NHS Trust and Imperial College London, and Imperial's British Heart Foundation Centre of Research Excellence.

This article is based on press materials supplied by The Lancet.