TLDR: Hallym University Sacred Heart Hospital in South Korea has successfully implemented ‘HAI,’ a generative artificial intelligence platform, to automate the drafting of electronic medical records (EMRs). This innovative system, the first of its kind in Korea to cover the entire inpatient care cycle, is projected to save doctors up to 30 days of administrative work annually, freeing up valuable time for patient care, research, and teaching.
At Hallym University Sacred Heart Hospital in Anyang, Gyeonggi Province, the process of writing patient medical records has been revolutionized, now taking mere seconds. This significant advancement is attributed to ‘HAI,’ a generative artificial intelligence platform developed by Hallym University Medical Center.
When neurology professor Lee Min-woo activates the ‘HAI’ button within the hospital’s electronic medical record (EMR) program, the day’s progress notes are instantly generated. These notes are comprehensive, including sections on patient-reported symptoms, examination findings, clinical assessments, and future care plans. After a quick review and final approval, an inpatient’s daily record is complete.
This marks a groundbreaking achievement as Hallym University Medical Center is the first in Korea to apply AI across the entire inpatient care cycle, from admission to discharge. While some hospitals have utilized AI for specific parts of the record-keeping process, HAI’s comprehensive application is unique. The initiative comes at a crucial time, addressing a shortage of residents who traditionally managed the bulk of record-keeping, a situation exacerbated by last year’s labor disputes between doctors and the government.
To develop HAI, the hospital integrated 97 categories of patient data from its information system, encompassing nursing notes and major test results from the preceding 24 hours, to generate initial drafts. The AI was meticulously trained by comparing its output with records previously written by doctors, allowing for fine-tuning to match the precise style and detail expected by clinicians. Professor Lee, who heads the HAI task force, stated, ‘We’re continually refining it so the notes come out exactly as we want without any changes.’
Also Read:
- AWS Pioneers AI-Enhanced Patient Profiles for Transformative Healthcare Insights
- Highmark Health and Abridge Forge Alliance to Transform Prior Authorization with AI
The system is currently operational in neurology, obstetrics and gynecology, and otolaryngology for specific conditions, with plans for a hospital-wide expansion in the near future. Professor Lee estimates that in neurology, where writing an inpatient EMR typically takes about 10 minutes, HAI saves him approximately 90 minutes daily, translating to an impressive 20 to 30 days saved annually. Hospital officials project that if adopted across all departments, the system could save approximately 83,000 hours each year. Professor Lee emphasized the broader impact, noting, ‘That time can go toward seeing more patients, research or teaching.’
