AI stethoscope study published in The Lancet

The Lancet has published a large NHS study which found that using an AI-enabled stethoscope could help detect serious heart conditions faster and more frequently in primary care.

The clinical trial, called TRICORDER, led by researchers at Imperial’s National Heart and Lung Institute and Imperial College Healthcare NHS Trust, and supported by the Imperial Biomedical Research Centre (BRC), examined if the technology could detect heart conditions such as heart failure, arrhythmias, and valve disease.

The year-long study, published by The Lancet on 28 January 2025, found that although the technology worked well, it didn’t significantly increase the overall number of heart failure diagnoses.

This was primarily because too many GPs did not use the device consistently in routine clinical practice. When doctors used the AI stethoscope as intended, the technology detected these cardiovascular conditions faster and more frequently.

Dr Patrik Bachtiger, one of the researchers who led the study said: “Our trial suggests AI tools like smart stethoscopes help us detect heart conditions earlier, but only if they are used and properly integrated into everyday clinical practice.

“Although the technology performed well, the impact it could have in the real world depends on doctors being able to use the technology easily during busy clinics.”

Results first published in September 2025 show that patients examined with the Eko device were twice as likely to be diagnosed with heart failure.

They were also three and a half times more likely to be diagnosed with atrial fibrillation, an abnormal heart rhythm which can increase the risk of having a stroke, and almost twice as likely to receive a diagnosis of heart valve disease.

The stethoscope was trialled with 12,725 patients who showed symptoms of heart failure – breathlessness, fatigue or swelling of the lower legs and/or feet, across 205 GP surgeries in north west London.

These patients were compared to patients where AI stethoscopes were not used.

Professor Nicholas Peters, who led the team said: “Studies like this are essential for patients, health systems, industry, regulators and researchers to bring the most promising health innovations into routine patient care.”

Results also showed that 70% of GP surgeries given the smart stethoscopes in the study stopped using them, or used them infrequently, after 12 months,

Dr Mihir Kelshiker, co-lead author, added: “Some practices used the device less over time because it added extra steps to routine care and was not well integrated with existing electronic health record systems.

“Despite its diagnostic accuracy, it was these practical barriers that limited the overall impact of the technology.”

Also, two-thirds of people identified by the AI stethoscope as having suspected heart failure did not have it when given a further blood test or heart scan, which could lead to unnecessary anxiety and tests for some people.

Researchers said that the technology should only be used for patients with symptoms of suspected heart problems, and not for routine checks.