We collect preoperative patient data, including demographics, comorbidities, laboratory results, and surgical descriptions, and employ machine learning, deep learning, and natural language processing techniques to predict post-surgical mortality risk. It results in highly accurate predictions with an AUROC of 0.96. Even during real-world usage in surgical settings, it maintains an AUROC of 0.83.
Features / strengths
“Anesthesia risk prediction system” has successfully implemented at Far Eastern Memorial Hospital. It assists the physician to identify high-risk patients and take proactive interventions, therefore, enhancing the quality of care and efficiently managing healthcare resources.
This project highlights the important role of textual surgical descriptions in predicting post-surgical mortality. This technology holds a significant potential to bring transformative changes in the healthcare industry.
Specification in detail
Far Eastern Memorial Hospital