흩어진 의료데이터 하나로 잇는 힘…의료 AI ‘미래 여는 열쇠’ [건강한겨레]

AI-driven medical paradigms will shift towards self-planning and acting AI, according to BCG managing director and senior partner Aushkan Afkami, who predicted this in a December report. AI will move beyond summarizing medical records to presenting disease risks and treatment directions based on patients' comprehensive information.

It will also perform multiple tasks across several stages. However, even the most advanced AI requires reliable data to function effectively. The report emphasized that the future of AI-based medicine relies not only on algorithms but also on the completeness of the data ecosystem that supports them. Korean medical institutions struggle with data scarcity and fragmentation.

National medical data is scattered among hospitals, public institutions, and individual businesses, often recorded using different codes and expressions. AI trained on such disparate and normatively inconsistent data may interpret the same information differently or produce biased results. The government and the Korean Health Information Management System (KHIMS) are jointly pursuing a national unified biobig data platform, digital health information interchange system, medical data standardization, and public hospital AI infrastructure.

These efforts aim to collect more data and create tangible changes in diagnosis and health management by linking fragmented information. The "2026 Annual Symposium," opening in Seoul's Coex on September 10-11, will discuss the conditions for this transformation. The first day will cover topics such as building clinical and genomic data banks, advancing advanced data analysis methods, and the case of the UK BioBank, which has accumulated genetic information, clinical records, lifestyle data, and physical measurements for long-term analysis.

The symposium will also examine the digital health information interchange system, which connects health information spread across various institutions, allowing patients to directly check and utilize their health records through the "My Health Record" app. The symposium will emphasize that bio-big data construction and medical information connection are not separate but integrated aspects of a single system, with research and treatment flowing seamlessly into one another.

The goal is for digital transformation benefits to extend beyond large hospitals in the capital, addressing disparities in healthcare access. Key discussions during the symposium include reducing regional and public hospital gaps through AI agents, establishing a medical referral system where neighborhood, community, and large hospitals share patient responsibilities according to condition.

However, simply installing the platform is not enough; information system maintenance, standardized records, operational staff, and continuous technical support are necessary. AI-generated incorrect information must be verified and corrected promptly, with clear responsibility assigned. The final condition for connection is public trust.

Medical information contains not only disease and medication records but also sensitive genetic and mental health data. As connection expands, concerns about data breaches and unauthorized use also increase. The symposium will explore the "secondary use" system, allowing the use of collected information for research, policy-making, and pharmaceutical safety evaluations.

Legal frameworks for digital health care will cover patients' rights to access their medical information and transfer it to other institutions, as well as protections for research and AI development. Experts from the OECD will present governance and regulatory conditions for responsible AI deployment in healthcare.