"MeDiCU-AI" to Support ICU Patient Discharge Decisions, Officially Adopted for the First Time in Japan at Kansai Medical University General Medical Center
MeDiCU Inc. has announced the first official adoption in Japan of its AI programs, "MeDiCU-AI Discharge/Transfer Support Program" and "Symptom Description Drafting Support Program," at Kansai Medical University General Medical Center. These AI tools, leveraging the "OneICU" database, aim to assist in ICU patient discharge decisions and streamline medical documentation, enhancing healthcare services and reducing staff burden.
📋 Article Processing Timeline
- 📰 Published: April 15, 2026 at 22:00
- 🔍 Collected: April 15, 2026 at 13:31
- 🤖 AI Analyzed: April 19, 2026 at 10:47 (93h 15m after Collected)
## Press Release Information
Title: "MeDiCU-AI" to Support ICU Patient Discharge Decisions, Officially Adopted for the First Time in Japan at Kansai Medical University General Medical Center
Subtitle:
Company Name:
Industry:
Body (First 8000 characters):
MeDiCU Inc. (Headquarters: Osaka City, Osaka, CEO: Takahiro Kinoshita, hereinafter "MeDiCU"), a provider of "OneICU," the world's largest database for emergency and critical care, will begin providing the "MeDiCU-AI Discharge/Transfer Support Program" to assist in deciding when patients admitted to the ICU should be discharged, and the "Symptom Description Drafting Support Program" to streamline physician documentation, to Kansai Medical University (Hirakata City, Osaka, Chairman: Toshio Yamashita, Chancellor: Tatsuo Kinashi, hereinafter "Kansai Medical University"). This is an extremely rare case globally, and the first in Japan, of a program that supports ICU patient discharge decisions using AI trained on large-scale multi-institutional medical data (※1).
Background of Formal Adoption
Kansai Medical University and MeDiCU have been conducting joint research since 2024 on "Creation of an acute event prediction model using vital information and examination of its usefulness." This multi-institutional study involves ICUs (Intensive Care Units) from 39 university and general hospitals nationwide, has built the "OneICU database" by collecting approximately 200,000 cases (※2) of vital information, blood test results, and medication data, and has already reported research findings in numerous academic papers and conference presentations. Furthermore, MeDiCU has developed an AI to support decisions for transferring patients admitted to emergency and critical care centers/ICUs to general wards, using the "OneICU database." They have revealed that this model can grasp patient conditions with higher accuracy than existing indicators.
The model underwent a demonstration adoption at Kansai Medical University General Medical Center starting in August 2025. Its safety and accuracy in assisting physician judgments in clinical settings have been demonstrated, leading to its formal adoption as the "MeDiCU-AI Discharge/Transfer Support Program" from April 2026. Additionally, during the demonstration adoption, the "Symptom Description Drafting Support Program," which generates drafts of symptom descriptions by AI, was also developed. The utilization of both programs is expected to lead to the provision of more advanced medical services and a reduction in the burden on medical personnel.
※1: MeDiCU Research
※2: As of March 2026
■ About MeDiCU-AI "Discharge/Transfer Support Program"
ICU beds are limited, and patient turnover is necessary depending on the occurrence of severe cases or the schedule of surgeries performed within the hospital. While decisions on allocating limited beds to which patients are made carefully based on the patient's condition and the physician's clinical experience, it often happens that patients' conditions worsen after discharge from the ICU, requiring re-admission. Re-admission tends to prolong the ICU stay and lead to higher medical costs. Furthermore, research results show that patients who are re-admitted have approximately 4 times the risk of mortality compared to those who are not (※3).
This program is the first AI product in Japan that calculates the probability of a patient experiencing a sudden deterioration within 48 hours of ICU discharge, based on past statistical information. By utilizing this program, it has become possible to initiate decisions on whether to continue ICU treatment or discharge the patient earlier, without relying solely on the physician's clinical experience (※4).
※3: Ponzoni CR, et al.Ann Am Thorac Soc. 2017;14:1312-1319.
※4: Comparison between January 1, 2025 - May 31, 2025 (pre-demonstration adoption) and August 1, 2025 - December 31, 2025 (demonstration period).
■ About MeDiCU-AI "Symptom Description Drafting Support Program"
Symptom descriptions are documents required for medical billing, detailing the specific content of symptoms, their severity, test results, and treatment progress. To demonstrate to a third party the basis for requiring advanced treatment, it is necessary to compile details of the patient's condition severity, procedures performed, and medications administered in chronological order. Physicians involved in emergency and critical care spend a significant amount of time creating these extensive documents in their busy schedules.
The "Symptom Description Drafting Support Program" uses AI to analyze vital sign and electronic health record data to automatically generate drafts of symptom descriptions. By utilizing this program, the time required for creating symptom descriptions can be reduced to approximately 1/4 of the previous time (※5), leading to a significant reduction in physician burden, as well as the expectation of preventing omissions in medical billing.
※5: Comparison of creation time when symptom descriptions were created conventionally for the same disease versus when MeDiCU-AI was used (Conventional: approx. 40 minutes, MeDiCU-AI: approx. 10 minutes).
Title: "MeDiCU-AI" to Support ICU Patient Discharge Decisions, Officially Adopted for the First Time in Japan at Kansai Medical University General Medical Center
Subtitle:
Company Name:
Industry:
Body (First 8000 characters):
MeDiCU Inc. (Headquarters: Osaka City, Osaka, CEO: Takahiro Kinoshita, hereinafter "MeDiCU"), a provider of "OneICU," the world's largest database for emergency and critical care, will begin providing the "MeDiCU-AI Discharge/Transfer Support Program" to assist in deciding when patients admitted to the ICU should be discharged, and the "Symptom Description Drafting Support Program" to streamline physician documentation, to Kansai Medical University (Hirakata City, Osaka, Chairman: Toshio Yamashita, Chancellor: Tatsuo Kinashi, hereinafter "Kansai Medical University"). This is an extremely rare case globally, and the first in Japan, of a program that supports ICU patient discharge decisions using AI trained on large-scale multi-institutional medical data (※1).
Background of Formal Adoption
Kansai Medical University and MeDiCU have been conducting joint research since 2024 on "Creation of an acute event prediction model using vital information and examination of its usefulness." This multi-institutional study involves ICUs (Intensive Care Units) from 39 university and general hospitals nationwide, has built the "OneICU database" by collecting approximately 200,000 cases (※2) of vital information, blood test results, and medication data, and has already reported research findings in numerous academic papers and conference presentations. Furthermore, MeDiCU has developed an AI to support decisions for transferring patients admitted to emergency and critical care centers/ICUs to general wards, using the "OneICU database." They have revealed that this model can grasp patient conditions with higher accuracy than existing indicators.
The model underwent a demonstration adoption at Kansai Medical University General Medical Center starting in August 2025. Its safety and accuracy in assisting physician judgments in clinical settings have been demonstrated, leading to its formal adoption as the "MeDiCU-AI Discharge/Transfer Support Program" from April 2026. Additionally, during the demonstration adoption, the "Symptom Description Drafting Support Program," which generates drafts of symptom descriptions by AI, was also developed. The utilization of both programs is expected to lead to the provision of more advanced medical services and a reduction in the burden on medical personnel.
※1: MeDiCU Research
※2: As of March 2026
■ About MeDiCU-AI "Discharge/Transfer Support Program"
ICU beds are limited, and patient turnover is necessary depending on the occurrence of severe cases or the schedule of surgeries performed within the hospital. While decisions on allocating limited beds to which patients are made carefully based on the patient's condition and the physician's clinical experience, it often happens that patients' conditions worsen after discharge from the ICU, requiring re-admission. Re-admission tends to prolong the ICU stay and lead to higher medical costs. Furthermore, research results show that patients who are re-admitted have approximately 4 times the risk of mortality compared to those who are not (※3).
This program is the first AI product in Japan that calculates the probability of a patient experiencing a sudden deterioration within 48 hours of ICU discharge, based on past statistical information. By utilizing this program, it has become possible to initiate decisions on whether to continue ICU treatment or discharge the patient earlier, without relying solely on the physician's clinical experience (※4).
※3: Ponzoni CR, et al.Ann Am Thorac Soc. 2017;14:1312-1319.
※4: Comparison between January 1, 2025 - May 31, 2025 (pre-demonstration adoption) and August 1, 2025 - December 31, 2025 (demonstration period).
■ About MeDiCU-AI "Symptom Description Drafting Support Program"
Symptom descriptions are documents required for medical billing, detailing the specific content of symptoms, their severity, test results, and treatment progress. To demonstrate to a third party the basis for requiring advanced treatment, it is necessary to compile details of the patient's condition severity, procedures performed, and medications administered in chronological order. Physicians involved in emergency and critical care spend a significant amount of time creating these extensive documents in their busy schedules.
The "Symptom Description Drafting Support Program" uses AI to analyze vital sign and electronic health record data to automatically generate drafts of symptom descriptions. By utilizing this program, the time required for creating symptom descriptions can be reduced to approximately 1/4 of the previous time (※5), leading to a significant reduction in physician burden, as well as the expectation of preventing omissions in medical billing.
※5: Comparison of creation time when symptom descriptions were created conventionally for the same disease versus when MeDiCU-AI was used (Conventional: approx. 40 minutes, MeDiCU-AI: approx. 10 minutes).