Detecting "Heart Failure Signs" in Just 15 Seconds — From "Listening" to "Diagnosing": AI x Stethoscope Doubles Diagnosis and Increases AF Detection by 3.5 Times in Demonstration

A research team from Imperial College London has tested a card-shaped AI stethoscope that simultaneously captures heart sounds and ECG, analyzing them with cloud AI, in primary care settings. In real-world data from 12,725 symptomatic patients, the group using the device was linked to diagnoses of heart failure at 2.33 times, atrial fibrillation at 3.45 times, and valvular disease at 1.92 times the rate, with the ability to triage risk in just 15 seconds. However, challenges in operation were identified, such as a high rate of false positives (about two-thirds) and a decrease in usage by approximately 70% of the GPs who introduced it within a year. It is suggested that the device should be applied for triaging symptomatic patients rather than asymptomatic screening. On social media, while some welcome it as a "game changer in primary care," there are also cautious voices pointing out issues with false positives, over-testing, and the limitations of auscultation findings. The key to success lies in early detection leading to earlier treatment, as well as the integration of workflows and the design of data governance.