"Send the Myocardial Infarction Patient to This Hospital"... AI Opens the Golden Hour Path

现场 at the "Daegu-Gyeongbuk Emergency Medical AX Technology Demonstration"
AI Suggests Differential Diagnoses and Test Items, and Drafts Prescriptions
"Emergency Room Ping-Pong" Disappears... Pilot Project for Transport System Innovation Expands

"The patient is alert, with a blood pressure of 140 over 100, a pulse rate of 105 bpm, a respiratory rate of 22 breaths per minute, oxygen saturation at 96%, and a temperature of 36.5 degrees Celsius."

On the 12th, at the Daegu Kyungpook National University Hospital auditorium, the 119 emergency medical team checked the condition of a patient complaining of chest pain during the 'Daegu-Gyeongbuk Smart Transport System' technology demonstration. The mobile terminal screen of the emergency team shown on the screen displays results such as patient vital signs collected by artificial intelligence (AI), electrocardiogram analysis, and severity classification. Ministry of Health and Welfare

When the 119 emergency medical technician measured the patient’s vital signs and read the results aloud, the artificial intelligence (AI) on the tablet instantly recognized the speech and began automatically entering the patient’s symptoms and necessary information as text. When the patient said, "My chest hurts a lot," the AI immediately converted this to "chest pain" and entered it accordingly. Accompanying symptoms such as "numbness in the arm" and "cold sweat" were also recorded automatically. When the emergency worker pressed the "save" button, the AI evaluated the severity and urgency of the condition and classified it as "Pre-KTAS Level 2 - Cardiac pain." The emergency worker then took a photo of the patient’s ECG on the spot and uploaded it, after which the AI analyzed the ECG waveform and diagnosed "ST-segment elevation myocardial infarction." This is a situation requiring emergency vascular intervention.

At the same time, an alert for a "time-sensitive emergency patient (Fast Track) occurrence" went off on the monitor in the emergency room at Kyungpook National University Hospital. Without the need for the emergency worker to call each hospital to check for patient acceptance, the AI system matched the patient’s estimated diagnosis with the real-time status of medical resources in nearby hospitals and simultaneously sent acceptance requests to six emergency rooms capable of optimal treatment. Once the hospital clicked the "accept" button, the ECG data measured by the emergency worker was shared in real time, allowing the cardiology team to complete surgical preparations even before the patient arrived. This is the moment when the tragedy of so-called "emergency room ping-pong," where patients lost critical time while searching for a hospital to accept them, is resolved by AI technology.

The government is expanding its "Emergency Transport System Innovation Pilot Project" to the Daegu and Gyeongbuk regions to address delays in patient transport due to emergency room refusals. From the moment an emergency patient boards the ambulance to the time they receive treatment in the emergency room, AI is deployed to quickly and accurately analyze the patient's condition and find the hospital best suited for treatment and transport accordingly.

The Ministry of Health and Welfare held a "Daegu-Gyeongbuk Transport Protocol Revision Meeting," presided over by Minister Jeong Eun-kyeong, at Kyungpook National University Hospital in Daegu on the 12th, along with a technology demonstration for the AI-based "Daegu-Gyeongbuk Smart Transport System."

Through the Korean ARPA-H project, Kyungpook National University Hospital is developing the "SAVE-R" platform, which uses multimodal AI technology to connect the entire process from ambulance to emergency room into a single information network. When recommending a transfer hospital, not only the number of available beds is considered, but also the AI-classified severity of the patient, the required final treatment, the congestion level of each hospital’s emergency room, the availability of specialized treatment teams, the transport distance, and the patient’s past medical history, all of which are comprehensively calculated.

Professor Ryu Hyunwook of the Department of Emergency Medicine at Kyungpook National University Hospital explained, "Currently, the emergency patient transfer process requires calling hospitals one by one to check for acceptance, which causes delays and makes it difficult to make rational emergency room acceptance decisions, as unreliable patient information is repeatedly relayed by phone." He added, "The key feature of the SAVE-R platform is that by applying AI technology, the emergency team can more accurately assess the patient’s condition and, at the same time, analyze hospital resources to quickly recommend the optimal hospital for transfer." Based on the information delivered in this way, hospitals can begin preparing treatment in advance, ultimately reducing the time required to provide final treatment.

AI Suggests Diagnostic Candidates and Connects to Prescription Sets

From the moment the patient arrives in the emergency room, "AEGIS" from Samsung Medical Center assists medical staff in decision-making, from gathering patient information to diagnosis and prescriptions. First, the patient's basic information, ambulance log, vital data, and initial nursing assessment provided by the emergency team are presented in an organized manner, and even the patient's past healthcare utilization history and medical records, called "MyData," are retrieved in real time. This allows medical staff to quickly grasp past medical history that the patient might not have mentioned.

The AEGIS AI analyzes the patient’s symptoms and medical history to suggest the diseases that emergency staff should prioritize. For a 72-year-old male patient with fever, shortness of breath, and low blood pressure, the system prioritized past history related to pulmonary diseases, hypertension, diabetes, and infections, while placing a relatively unrelated history of past cataract surgery at the end of the list.

Questions asked by the medical staff to the patient were also transcribed into text in real time. Interview details such as "I feel short of breath" and "I have had a cough and phlegm for a few days" were recorded automatically, and the AI used this information to draft the initial emergency room medical record.

The AI then predicted a high likelihood of sepsis and suggested differential diagnoses and sets of tests and prescriptions to be considered. When the staff selected the sepsis category, the initial prescription list—including fluids, blood tests, and imaging tests—appeared automatically. The doctor then added or removed items as appropriate for the patient’s condition to finalize the prescription. While AI cannot replace the judgment of medical staff, it quickly organizes vast amounts of data and reduces the possibility of errors, providing crucial support for medical decision-making.

Professor Cha Wonchul of the Department of Emergency Medicine at Samsung Medical Center said, "While the ultimate judgment and decision-making rest with the medical staff, in the emergency room, a difference of a few minutes can determine life or death, so the time-saving effect of AI is extremely important. Reducing the time spent on the phone allows staff to focus on observing the patient’s condition, thus improving the quality of care provided."

The diagnostic concordance rate for these AI platforms has already exceeded 90%, and the treatment concordance rate is also around 90%. The government expects that if such AI transformation (AX) is implemented at emergency medical sites, it will significantly reduce the work burden on hospital staff while enabling more patients to be safely cared for with limited emergency room beds and personnel.

On the 12th, at Kyungpook National University Hospital in Daegu, Jeong Eun-kyeong, Minister of Health and Welfare, received an explanation from Professor Ryu Hyunwook of the Department of Emergency Medicine at Kyungpook National University Hospital about the AI-based emergency medical platform "SAVE-R" to be applied to ambulances during the demonstration of the "Daegu·Gyeongbuk-type Smart Transport System." Ministry of Health and Welfare

Revision of Daegu-Gyeongbuk Transfer Protocols... Regionalized Response

Meanwhile, the meeting also discussed proposed revisions to the emergency patient transfer protocols for the Daegu and Gyeongbuk regions.

As a key hub in the Yeongnam region, Daegu plans to launch a "multi-institutional transfer and collaborative care network" this month, which will simultaneously request acceptance of Pre-KTAS Level 1–2 critical emergency patients at six regional and local centers. If no hospital within the region is able to accept the patient, a super-regional transfer system will be activated to identify hospitals nationwide.

Gyeongbuk’s protocols reflect its geographic characteristics. With a vast area where medical institutions are not evenly distributed and including mountainous terrain and Ulleungdo Island, helicopter transport is essential. When a critical emergency patient arises in a medically underserved area, doctor helicopters and fire department helicopters are mobilized, with 119 emergency services providing support. In both regions, the regional command centers function as a "control tower" to help select a receiving hospital for patients whom local facilities cannot accommodate.

Minister Jeong emphasized, "Results from the earlier pilot project in the Honam region showed a reduction in emergency patient deaths and significant improvements in the regional transfer system. The key is to establish a transfer system based on trust among local agencies and to develop and implement agreed-upon transfer protocols."

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