The Cost-effectiveness of Artificial Intelligence Acute Kidney Injury Prediction Auxiliary Software (Acura AKI)

"Huede" AI Aided AKI Prediction Software, Acura AKI, uses machine learning algorithms to predict the risk of AKI within the next 24 hours and provide a ranking of feature importance. By using Acura AKI, physicians can assess the risk of AKI, focusing on high-risk patients to provide care decisions. This study will be conducted in a prospective randomized clinical trial in adult ICUs, implementing the Acura AKI system for predicting AKI. The study aims to determine whether early prediction and intervention using the Acura AKI system can improve the outcomes of critically ill patients with adverse kidney conditions. The study endpoint is to evaluate the cost-effectiveness of using Acura AKI, including the incidence of AKI, dialysis rates, mortality rates, length of hospital stay, and treatment costs.

"Huede" AI Aided AKI Prediction Software, Acura AKI, uses machine learning algorithms to predict the risk of AKI within the next 24 hours. It has undergone cross-hospital validation at four medical centers in Taiwan (Taichung Veterans General Hospital, Mackay Memorial Hospital, National Cheng Kung University Hospital, and Kaohsiung Medical University Hospital), successfully obtaining invention patents in Taiwan and the United States, as well as receiving a software medical device license from the Taiwan Food and Drug Administration. Acura AKI is installed on the hospital's servers, where it processes patient physiological data, laboratory parameters, and medication information to infer the risk of AKI occurring within 24 hours. It also provides a ranking of feature importance. By using Acura AKI, physicians can assess the risk of AKI, focusing on high-risk patients to provide care decisions. This study will be conducted in a prospective randomized clinical trial in adult ICUs, implementing the Acura AKI system for predicting AKI. In the intervention group with Acura AKI system, physicians will be proactively notified via sending alarm message when Acura AKI identifies a high-risk patient population. After receiving alarm message, physicians and pharmacists will provide feedback and recommendations, including blood pressure, fluid management, infusion options, medication adjustment suggestions, and dialysis recommendations. The study aims to determine whether early prediction and intervention using the Acura AKI system can improve the outcomes of critically ill patients with adverse kidney conditions. Additionally, the researchers will collect 20ml of urine from Acura AKI identified patients to test for urinary biomarkers predictive of AKI then verify the performance of Acura AKI with these urinary biomarkers. The study endpoint is to evaluate the cost-effectiveness of using Acura AKI, including the incidence of AKI, dialysis rates, mortality rates, length of hospital stay, and treatment costs.