Intelligent Intensive Care Unit Improved Quality of Critical Care
Since 2017, our hospital has implemented the Kaohsiung Veterans General Hospital Intelligent ICU Care and Decision Support System, integrating information technology to optimize clinical decision-making, thereby enhancing patient care quality and safety. Addressing challenges inherent in intensive care units (ICUs)—such as delayed medical decisions, prolonged lengths of stay, and high infection risks—the introduction of this intelligent system has effectively overcome these difficulties and significantly improved treatment efficacy and patient safety.
The system integrates data from HIS (Hospital Information System), NIS (Nursing Information System), and ICCA (IntelliSpace Critical Care and Anesthesia) to provide standardized data. It leverages Artificial Intelligence (AI) for condition prediction, demonstrating notable success in forecasting risks of cardiogenic shock, acute kidney injury (AKI), and mortality. The system's built-in AI algorithms continuously monitor patients' vital signs and are coupled with a remote alert system, enhancing the early detection of critical conditions and response speed. Through these technologies, medical staff can rapidly identify potential acute conditions or sepsis risks, enabling preemptive interventions and effectively reducing the likelihood of patient deterioration.
Beyond predictive capabilities, the system incorporates management of bundled care for urinary catheters, central venous catheters, and endotracheal tubes. It utilizes Power BI to monitor metrics like length of stay and disease progression, facilitating meticulous and standardized patient management. The synergy between these technologies and systems not only bolsters patient safety but also improves care efficiency and alleviates the burden on healthcare staff.
Following the system's implementation, the rate of poorly controlled patient pain decreased by 28.5%, the rate of poorly controlled delirium dropped by 22.7%, and the quality sleep rate increased by 17.1%. The average ICU length of stay decreased from 7.53 days to 5.65 days, resulting in estimated annual medical cost savings of approximately NT$107 million. Furthermore, the system has contributed to improved patient mortality rates and significantly reduced ventilator days and the risk of line-associated infections, enhancing overall patient outcomes and safety.
Looking ahead, our hospital will continue to refine the AI prediction models and strengthen staff education and training to maximize system effectiveness. We also plan to promote the adoption of intelligent healthcare solutions in other medical institutions, further elevating the quality of critical care and reducing the healthcare burden.
The Intelligent Multi-Agent System for Severe Sepsis Early Detection
To enhance patient safety and care quality, our hospital has specifically developed The Intelligent Multi-Agent System for Severe Sepsis Early Detection. This system aims to overcome the challenges of timely sepsis detection often faced by traditional manual assessment. It integrates multi-modal data from electronic medical records, laboratory reports, and physiological monitoring equipment. Through a series of intelligent agents specifically designed for early sepsis warning and treatment assistance—such as the "Lactate Value Monitoring Intelligent Agent" that monitors lactate levels, and the "Antibiotic Management Intelligent Agent" and "Fluid Therapy Intelligent Agent" which provide antibiotic and fluid recommendations—the system achieves precise risk assessment and real-time alerts for severe sepsis. This innovative technology not only provides strong decision support for the medical team but also significantly improves the efficiency of critical care and patient outcomes.
The implementation of this system has yielded remarkable results. In terms of workload reduction, the system's intelligent recommendations have significantly decreased clerical work by over 30% and overall clinical workload by 20% for medical staff. Especially in emergency situations, the decision-making burden has been reduced by over 50%, enabling medical personnel to make critical judgments more quickly and confidently. Economically, by shortening the average hospital stay by 1.6 days, the system saves over NT$17 million annually for patient families and over NT$107 million for society as a whole. Additionally, reducing unnecessary antibiotic use saves over NT$9 million annually in drug costs, substantially lowering medical and societal expenditures. Regarding care quality and safety, after system implementation, the execution rate of rapid fluid resuscitation initiated by the Lactate Value Monitoring Intelligent Agent for severe sepsis patients in the emergency department with elevated lactate levels increased from 51.06% to 68.75%, an improvement of 34.65%. The median time for the first antibiotic administration for severe sepsis patients in the emergency department decreased from 203.94 minutes to 169.96 minutes, a reduction of 34 minutes. Concurrently, the execution rate for vasopressor (Norepinephrine) administration after the first episode of hypotension significantly increased from 17.25% to 38.15%, a remarkable improvement of 121.16%. All these data reached statistical significance (P < 0.05), fully demonstrating the system's exceptional performance in enhancing patient safety and treatment efficacy.
Overall, the successful implementation of The Intelligent Multi-Agent System for Severe Sepsis Early Detection has not only effectively addressed the difficulties in early diagnosis and risk assessment of sepsis but has also significantly optimized medical resource allocation and improved overall care quality through its automated monitoring, precise early warning, and treatment assistance functions. This innovative technology provides patients with more timely and accurate treatment and offers medical teams a more efficient and safer working environment. It is a tangible representation of our hospital's commitment to enhancing patient safety and achieving medical excellence.
International Conference of ISQua in Seoul 2023