What Took Doctors 30 Minutes Now Takes Just 3... AI Transforming the Emergency Room Golden Hour [The Future of Medical AX]①

Zero Typing and Record-Keeping for Doctors in the ER
Beyond Imaging Interpretation: AI Supports Real Clinical Decisions
More Time for Patients, Less for Routine Tasks

Editor's NoteKorea's healthcare system stands at a major turning point. The entry into a super-aged society, the rise in chronic diseases, the shortage of essential medical personnel, and the widening regional healthcare gap have become challenges that are difficult to solve with the existing medical system alone. Both the government and the medical community are focusing on medical AX (AI Transformation), which goes beyond simply introducing artificial intelligence (AI) and aims to redesign the entire healthcare environment. The goal is not just for AI to support or replace the role of doctors, but to create a healthcare environment where everyone, regardless of region or income, can receive high-quality care. The Asia Business Daily explores how rapidly advancing AI is transforming the medical field and how it may reshape Korea’s healthcare system.

(ChatGPT Generated Image)

At 2 a.m., a man in his 70s named Kim, experiencing fever, difficulty breathing, and low blood pressure, was brought in by 119 emergency ambulance. As soon as the patient arrives, the emergency room monitor displays an emergency log containing his past diagnoses, medications, and treatment history, along with vital signs, electrocardiogram, and severity estimated by AI—all transmitted moments ago from the ambulance. This information is not presented randomly. AI prioritizes items most relevant to the patient’s current symptoms, while less related information is summarized. For example, Kim’s history of pulmonary disease, hypertension, diabetes, and infection symptoms is placed at the top, while less relevant histories such as cataract surgery are relegated to the end. Out of dozens of pieces of information related to the patient, only those immediately needed by the medical staff are selected. The questions and answers between the medical staff and the patient are recorded in real-time text. Interview responses such as 'I feel short of breath' or 'I've had a cough and phlegm for a few days' are automatically entered, and based on this, the AI drafts an initial emergency room medical record. The doctor does not need to type anything by hand.

The AI predicts that Kim is highly likely to have sepsis and generates a list of differential diagnoses that need to be considered first. Retrieval-Augmented Generation (RAG) technology instantly references emergency medicine guidelines and the hospital’s clinical protocols. It shows the medical rationale for determining the patient’s current state as sepsis, along with the necessary follow-up tests to confirm it. Once the physician agrees with this diagnosis, the AI immediately brings up sepsis tests and fluid therapy prescriptions. The system also highlights items missing from the physician’s selected prescription set and prompts, 'Review if anything needs to be added.' In the race-against-time environment of the emergency room, doctors and AI collaborate to diagnose the patient and make a final prescription.

In the past, it would have taken at least 20–30 minutes just to assess the patient’s condition and enter prescriptions. Now, the entire process is completed in just three minutes. Less than 30 minutes after getting out of the ambulance, Kim receives appropriate antibiotics and avoids the risk of sepsis progression. This scenario assumes the fully developed operation of the emergency medical AI R&D project 'AEGIS,' currently being developed with support from the Ministry of Health and Welfare and the Korea ARPA-H project.

Currently, in clinical practice, AI has evolved from merely assisting with interpretation of imaging tests such as X-rays and CT scans, to serving as a practical partner in medical decision-making. Without typing, voice recognition now enables real-time creation of medical records and drafts of clinical documents, while AI aggregates vast amounts of patient information to recommend appropriate diagnoses and even suggest initial prescriptions. According to the international journal NPJ Digital Medicine, the use of medical AI reduces imaging interpretation time by over 27%, and can decrease the workload of medical staff by up to 60%. Research also shows that using AI in breast cancer screening has increased cancer detection rates by 20%.

In the emergency room, chest X-ray analysis AI reviews small pneumothoraces or abnormal findings that physicians might miss. Electrocardiogram AI predicts left ventricular dysfunction that was previously hard to detect, providing additional insight into whether a patient’s shortness of breath is due to pulmonary or cardiac causes. If there are any findings that physicians have not noticed or are uncertain about, AI repeatedly prompts questions or highlights points that need to be checked. Thanks to AI, the accuracy and safety of medical care are improving.

The AEGIS system introduced in the earlier virtual scenario is still under development, but even now, its diagnostic and treatment concordance rates exceed 90% when compared to diagnoses made by real physicians. Led by Samsung Medical Center, this project is evolving into the world’s largest emergency medical database (DB), with de-identified and standardized emergency data from 17 major general hospitals nationwide. About 30 algorithms are being developed, including those for four key diseases such as myocardial infarction and stroke, where the golden hour is critical. When frontline physicians modify or provide feedback on AI suggestions, the model learns and accumulates more information, creating a virtuous cycle of improvement.

In the near future, AI is expected to support a broad range of tasks, including real-time medical information retrieval and remote collaborative care. It is fully possible that AI will control surgical robots to precisely assist surgeons or autonomously perform repetitive and simple procedures.

However, the medical community emphasizes that AI ultimately only supports physicians’ decisions; the final judgment is always made by the doctor, so AI cannot replace the physician’s role. Professor Cha Wonchul of the Department of Emergency Medicine at Samsung Medical Center explained, "If AI can process repetitive work such as reviewing test results, calculating complex figures, and typing medical records in just seconds, physicians will be able to spend more of their time on in-depth consultations with patients or critical medical decisions for treatment," adding, "This will raise the quality of medical services and patient care."

[The Future of Medical AX]

① From Diagnosis to Prescription in 3 Minutes... How AI Is Transforming the 'Golden Hour'

②"AI Is Technology That Allows Medical Staff to Spend More Time With Patients"

③Faster Than Human Eyes... AI Fills Medical Gaps at Regional Hospitals

④"AI That Never Tires—The Best Navigation System for Doctors"

⑤"Receiving Seoul-Level Care Even in Rural Areas"... Why the Government Is Rolling Up Its Sleeves for Medical AI

⑥"Built a Supercar but No Roads"... The Reality of Public Healthcare in the Provinces

⑦Overcoming Payment Gaps and Data Fragmentation... Conditions for Sustainable AX

This content was produced with the assistance of AI translation services.