Osaka Acute Care Medical Center Introduces 'MeDiCU-AI' to Support ICU Patient Condition Monitoring

The Osaka Acute Care Medical Center (Osaka City, Sumiyoshi Ward; President: Takeshi Shimazu, Hospital Director: Yuichiro Togi), under the Osaka Prefectural Hospital Organization, will introduce 'MeDiCU-AI,' an AI system developed by MeDiCU Inc. (Headquarters: Osaka City, Osaka Prefecture; CEO: Takahiro Kinoshita), to support decisions on ICU patient discharge at its Advanced Emergency and Critical Care Center, ICU, and Pediatric HCU (collectively referred to as the 'Critical Care Department').

Background of Introduction

The Osaka Acute Care Medical Center is designated by Osaka Prefecture as a 'Core Disaster Base Hospital' and plays a central role in mass casualty reception and regional patient transfer coordination during large-scale disasters. In particular, in the event of the anticipated Nankai Trough mega-earthquake, thousands of critically ill patients are expected within Osaka Prefecture alone. Efficiently managing approximately 50 ICU beds under limited time and resources is a critical challenge directly linked to regional survival rates.

Beyond large-scale disasters like the Nankai Trough earthquake, efficient use of limited critical care resources is also essential during sudden surges in medical demand, such as emerging or re-emerging infectious disease pandemics. To sustainably provide appropriate care to patients who truly require intensive treatment, it is essential to continuously monitor critically ill patients' conditions and safely transfer those with stabilized conditions to general wards, thereby maintaining the admission capacity of the Critical Care Department.

However, traditionally, monitoring bed occupancy and patient conditions in the Critical Care Department has heavily relied on clinical experience and verbal communication among physicians and nurses. Mechanisms to visualize and predict in real time 'which bed will become available' and 'which patient is approaching transfer-readiness' have not been sufficiently established. This issue not only affects bed management and operational efficiency during normal operations but also risks delaying the rapid assessment of available patient capacity and regional transfer coordination during emergencies.

The newly introduced MeDiCU-AI is an AI system that supports decision-making in critical patient management and bed utilization by visualizing risk scores in real time based on statistical data from previously treated patients. Beyond enhancing bed management in daily clinical practice, MeDiCU-AI is expected to contribute significantly to maximizing the effective use of limited critical care bed resources during large-scale disasters or pandemics, helping build a system capable of admitting 'as many critically ill patients as possible.'

By utilizing an AI system that comprehensively visualizes ICU patient conditions, we aim to establish a seamless critical care support system from routine operations to emergencies, and contribute to realizing a medical system capable of saving as many critically ill patients as possible, even during large-scale disasters, through optimal use of limited critical care resources.

Overview of MeDiCU-AI Implementation at Osaka Acute Care Medical Center

The system centrally visualizes in real time the certainty of discharge to general wards for all approximately 50 beds in the Critical Care Department, comparing each patient with historical statistical data. By enabling physicians, nurses, and bed management staff to share the same information platform, it facilitates rapid discharge and transfer coordination and immediate awareness of available bed capacity.

Additionally, the AI analyzes patient data, including vital signs, and scores the risk of clinical deterioration after ICU discharge that may lead to re-admission to the Critical Care Department, based on statistical outcomes from past patient cases. By transferring appropriate patients to general wards, the system increases bed turnover in the Critical Care Department and secures capacity for the next critically ill patient.

In the event of a disaster causing a surge of injured and ill patients, this system enables rapid assessment of the current number of available beds in the Critical Care Department and continuously evaluates the number of beds that could become available within the next few hours. This allows for rapid information sharing with DMAT (Disaster Medical Assistance Team) and swift regional transfer coordination from the onset of a disaster, facilitating the early establishment of a critical patient admission system.

Furthermore, even in resource-constrained disaster scenarios, the system supports bed management in the Critical Care Department beyond reliance on experience alone by objectively visualizing each patient's condition. By systematically transferring patients with low re-admission risk to general wards, the system aims to rapidly reserve ICU beds for more severely injured disaster victims.

Highlighting patients with high discharge uncertainty by color per ward

Comment from Dr. Satoshi Fujimi, Director of the Advanced Emergency and Critical Care Center, Osaka Acute Care Medical Center

During the novel coronavirus pandemic, we experienced the difficulty of securing sufficient medical resources for critically ill patients due to a sudden surge in cases. There were instances where the balance between medical supply and demand was severely disrupted, such as re-admissions due to patient deterioration after ICU discharge and shortages of beds for critically ill patients.

Through these experiences, we strongly recognized the importance of objectively assessing patient conditions using data and supporting timely discharge and transfer to maximize the effective use of limited critical care resources.

With the introduction of MeDiCU-AI, we aim to more comprehensively understand the overall status of the Critical Care Department, enhance routine bed management, and build a system capable of rapidly and appropriately admitting critically ill patients during surges in medical demand, such as large-scale disasters or infectious disease outbreaks.

As a core disaster base hospital, we will continue to strengthen our healthcare delivery system by leveraging data and technology, ensuring that local residents can access medical care with confidence.

Comment from Takahiro Kinoshita, CEO of MeDiCU

In disaster medical settings, continuously and accurately monitoring the status of critically ill patients within the hospital is extremely important. Personally, as an emergency physician and DMAT member, I was deployed to the 2016 Kumamoto earthquakes. In chaotic disaster medical environments, I strongly felt that clearly visualizing 'where critically ill patients are' leads to rapid decision-making and patient survival.

MeDiCU has supported the visualization of critically ill patients and operational efficiency using AI, leveraging approximately 200,000 clinical cases collected from emergency and critical care settings across Japan. We are confident that deploying this system in disaster medical settings will greatly support rapid decision-making by healthcare providers and significantly contribute to building 'disaster-resilient healthcare.'

About 'MeDiCU-AI'

MeDiCU-AI is an AI system developed using 'OneICU,' the world's largest database in emergency and critical care, built by MeDiCU, to support solving various challenges faced in clinical settings. In medical environments, rapid decision-making is required under time-critical conditions, while numerous challenges exist, such as increasing workloads for healthcare providers due to staffing shortages and the complexity of medical data management.

In addition to the program being introduced at Osaka Acute Care Medical Center, MeDiCU is actively developing multiple programs tailored to the specific challenges of individual medical institutions. Moving forward, we will continue to listen to healthcare providers and contribute to improving operational efficiency and enhancing the quality of medical care.