AI Medical Devices Demand Rigorous Clinical Evidence, Vuno CEO Urges Compliance
Table of Contents
A new framework for evaluating AI-driven medical technologies is gaining traction, but ensuring patient safety and fostering continued innovation hinges on strict adherence to regulatory guidelines, according to industry leaders.
The rapid advancement of artificial intelligence is transforming healthcare, but unlike traditional medical devices, AI-based systems require a unique approach to validation. This was the central message delivered by Vuno CEO Lee Ye-ha during a presentation at the Korea Future Bio and Health Forum on Thursday, hosted by Medigate News and the office of New Reform Party Rep. Lee Joo-young. Lee emphasized the critical need for medical institutions to comply with guidelines issued by the Ministry of Health and Welfare to support both the development of cutting-edge medical technology and, crucially, patient safety.
The Challenge of AI Validation
Traditional medical devices and pharmaceuticals benefit from established clinical evidence built over years of research and testing. New products are often iterations of existing, proven technologies. AI-based medical devices, however, present a different challenge. Their performance is heavily influenced by the data they are trained on and the methods used to verify their accuracy. This inherent variability necessitates a robust system for gathering and evaluating clinical evidence.
“Clinical evidence is very important when looking at AI medical devices from a regulatory standpoint,” Lee stated. “The standard for judging whether they actually have the same level of clinical effectiveness rather than just being the same product because it has the same input and same outcome must be established.”
The Rise of Advanced Medical Technology Systems
Fortunately, a pathway for gathering this crucial evidence is emerging through the implementation of advanced medical technology systems. These systems allow for the temporary use of innovative AI devices in clinical settings, even before full regulatory approval and reimbursement are secured. This approach, which includes options for reimbursement, non-insurance coverage, or selective reimbursement, is enabling real-world testing and data collection.
Vuno’s own DeepCARS, the first AI-based medical device in Korea to receive deferred evaluation as a new medical technology, exemplifies this progress. DeepCARS is an AI-powered solution designed to predict the risk of cardiac arrest within 24 hours by analyzing vital signs – blood pressure, pulse, respiration, and body temperature – of hospitalized patients. Since its launch in August 2022, the technology has been deployed in 140 hospitals, monitoring over 50,000 beds.
Demonstrating Real-World Impact
Early results are promising. According to Lee, a university hospital study demonstrated that the use of DeepCARS guidance led to approximately a 50% reduction in in-hospital cardiac arrests and a corresponding decrease in ward mortality rates. Vuno is currently conducting randomized controlled trials (RCTs) at four university hospitals to further solidify this evidence base, achieving Level 2a clinical trial status – surpassing the Level 3 evidence typically seen from overseas companies.
Previously, validating medical devices relied heavily on retrospective data analysis. However, the advanced medical technology system now facilitates prospective research, allowing for the assessment of utility and effectiveness with actual patient data.
Despite the positive momentum, challenges remain. Unlike traditional medical devices, where similarity in object, purpose, and method often equates to regulatory equivalence, AI devices exhibit significant performance variations based on their training and verification processes. This complexity demands a more nuanced regulatory framework.
A key concern raised by Lee is the potential for misuse and the need for robust risk management. “If advanced medical technologies are misunderstood and used indiscriminately to make profits, patient medical costs may increase and patient safety risks may increase as a result,” he warned. He urged both companies and the government to prioritize responsible implementation and diligent oversight, including active investigation of billing practices to ensure compliance with established guidelines.
The success of the advanced medical technology system, and the future of AI in healthcare, ultimately depends on a commitment to rigorous clinical evidence, proactive risk management, and unwavering dedication to patient safety.
