The UK's NHS deploys an AI stethoscope from Mayo Clinic spinoff Eko in 200 GP surgeries to detect heart failure, atrial fibrillation, and valvular heart disease
Rollout of stethoscope with pioneering technology could save lives and ease the strain on the NHS X: @ft , @mihir_kelshiker , @subhaan_siddiq , @mihir_kelshiker , @mihir_kelshiker , and @mihir_kelshiker LinkedIn: Noor Ashraf and Connor Landgraf X: @ft : The stethoscope has not evolved over the past 200 years - but now AI technology has updated the essential piece of equipment and could help speed up the diagnosis and treatment of heart disease https://www.ft.com/... Mihir Kelshiker / @mihir_kelshiker : We are grateful to the hundreds of participating GPs and primary care teams - the warp drive 🌌 making change happen; the @NIHRresearch for funding the programme, and to the thousands of @NHS patients supporting and guiding us. @subhaan_siddiq : Fantastic example of how #AI can transform the #NHS & healthcare systems around the world. “In 15 seconds, the Eko stethoscope can detect 3 types of heart disease” Artificial intelligence promises to transform diagnosis of heart disease https://www.ft.com/... Mihir Kelshiker / @mihir_kelshiker : Innovation in healthcare means nothing without implementation 🤝. We are delighted that the @FT has recognised our AI-powered pathway transforming heart failure 💔 diagnosis & management, and the unique achievement of our direct-to-care implementation. https://giftarticle.ft.com/... Mihir Kelshiker / @mihir_kelshiker : TRICORDER 🚀 is a unique clinical and academic endeavour, driven by visionary leadership from @DocOfHearts. We are grateful for the support of senior clinical and executive colleagues in @imperialcollege @HealthierNWL @ImperialNHS Mihir Kelshiker / @mihir_kelshiker : The article captures the richness of our primary care site visits and our approach to collaboration and training on the potential benefits of an AI stethoscope. We need these real-world implementation trials to fully understand and unlock AI's impact on health and healthcare 🧠 . LinkedIn: Noor Ashraf : More than half the time, I have no idea what this guy is cooking up on the side when talking about the revolutionary uses of AI and making the case … Connor Landgraf : Heart failure should not be diagnosed in the emergency department. We're excited to be deploying this proactive model for heart failure detection …
Context & Ripple Effects
Eko had already built a cardiac monitoring platform combining a digital stethoscope and ECG, following its earlier funding for AI-enabled cardiac monitoring. This NHS deployment moves that approach into routine primary-care settings, where a standard clinical instrument can become a triage input rather than solely a listening device.
It also extends a longer NHS pattern of testing AI for earlier detection, from DeepMind's NHS eye-scan project to tools used at the point of care. The significance is operational: adoption now depends on how well the output fits GP referral and follow-up workflows.
First-order effects
- GP surgeries in the rollout gain an AI-assisted method for flagging heart failure, atrial fibrillation, and valvular heart disease during examinations; Eko gains deployment in a large public primary-care network.
- Clinicians must incorporate the device's findings into clinical judgment and onward-testing decisions, rather than treating the stethoscope as a standalone diagnosis.
Second-order effects
- The rollout creates pressure for competing cardiac-device and diagnostic-AI suppliers to show that their products can work within frontline care pathways, not just in specialist or hospital settings.
- Earlier flags may shift work downstream toward confirmatory testing and specialist referrals, making pathway capacity and result-handling as important as the screening device itself.
Third-order effects
- If such deployments prove useful in practice, AI-enabled versions of familiar medical hardware could become a distribution route for clinical AI, with procurement increasingly focused on workflow integration and evidence of outcomes.
- The broader constraint may shift from inventing detection models to governing their use: clinical accountability, validation across care settings, and the ability of health systems to absorb follow-up demand.
The trend: This is part of the shift from standalone clinical AI pilots toward AI embedded in everyday diagnostic tools and care workflows.