TAIPEI (TVBS News) — National Taiwan University Hospital (NTUH, 台大醫院) and Chunghwa Telecom (中華電信) unveiled their award-winning "Intelligent Voice Emergency Prediction System" (智慧語音急診預測系統) at a press conference on Thursday (Jan. 9). The system, which recently won the 21st National Innovation Award — Clinical Innovation Award (國家新創獎-臨床新創獎), converts patient-caregiver dialogues into text to predict emergency events within two weeks with nearly 90% accuracy.
Ho Yi-lwun (何奕倫), vice dean of the National Taiwan University College of Medicine (台大醫學院), highlighted the system's importance as Taiwan faces an aging population and increasing comorbidity cases. He stated that the emergency prediction system can reduce hospitalization or mortality risks. It analyzes keywords and routine physiological data to forecast emergency needs, minimizing unnecessary complications.
The system excels in acute myocardial infarction cases, accurately determining emergency needs within the critical six-hour window. This capability allows patient triage and precise treatment, reducing emergency visits and hospitalization duration. Hung Chi-sheng (洪啟盛) from the NTHU Telehealth Center (遠距照護中心) noted the center's collaboration with Chunghwa Telecom since its establishment in Aug. 2009, processing approximately 540,000 calls and caring for over 7,000 patients.
Hung explained that the AI system uses natural language processing and semantic analysis to integrate nursing records and predict emergency probabilities with about 90% accuracy. Wu Ming-shiang (吳明賢), NTUH superintendent, emphasized the system's role in enhancing patient safety, care quality, and survival rates by overcoming time and space limitations in telehealth.