Structural Heart Disease Screening in Community Settings: AI + ECG Wearables vs Traditional Echocardiography

Abstract

Background: Structural heart disease (SHD) is a disease that is often undiagnosed in civil settings because it has a narrow provision of echocardiography and qualified staff. The potential of early detection is open to artificial intelligence (AI) improvements of wearable ECG technologies that could provide a potentially affordable, scalable technology. Objective: To reflect the success of AI-augmented ECG wearable screening over conventional community-based echocardiography in detecting previously unknown SHD. Method: A participatory community based research involved adulthood ages between 40-85 years of age in primary care clinics and at a community health screening. The control group was provided with a single-lead wearable ECG control group plus an AI-based detection algorithm as well as a portable echocardiography by trained sonographers. The accuracy of the diagnosis was assessed against echocardiography which has confirmed accuracy after confirming by cardiologists. The main outcomes were the sensitivity and specificity to finding left-ventricular hypertrophy, valvular malformations and atrial enlargement. Results: AI-ECG devices showed good diagnostic results (sensitivity 82 per cent, specificity 78 per cent), especially with the ability to diagnose left-ventricular hypertrophy and atrial enlargement. Conventional portable echocardiography exhibited greater overall accuracy (92 to 85 specificity and sensitivity respectively) but it was much more time consuming, expensive and required more operator capability. AI-wearables detected 71% of those found with echocardiography and identified one more group previously undetected by electrical abnormalities, which needed to be followed up. Conclusion: AI-enhanced ECG monitors represent an effective scalable screening instrument of structural heart disease in the community. Although Echocardiography has been classified as the gold standard of diagnosis, AI-wearables have the potential to significantly increase early detection potentials in situations when the resources are scarce.

Ähnliche Arbeiten