血流感染(Bloodstream Infections, BSI)為醫療照護中具高致死率與延長住院風險的重要感染類型，尤其與中心靜脈導管(central venous catheter, CVC)使用密切相關。為提升偵測效率與臨床反應速度，人工智慧(Artificial Intelligence, AI)技術已廣泛導入感染監測、預測與決策支援系統中，成為醫療品質管理的新興趨勢。國際應用方面，美國Mayo Clinic與Kaiser Permanente等機構已建立AI演算法自動分析電子病歷與實驗室數據，辨識高風險中心導管相關血流感染(Central Line-Associated Bloodstream Infection, CLABSI)個案，提供臨床即時預警。亞洲地區如新加坡與韓國亦導入自然語言處理(Natural Language Processing, NLP)與監測模型整合臨床紀錄，有效提升風險預測準確度。臺灣醫界也逐步擴展將AI運用於感染管制。例如:於急診建置「Sepsis AI模型」，快速辨識病人導致敗血症之風險等級，供臨床醫師迅速應變處理；自醫院的電子病歷系統彙集所有相關資料，將臺灣疾病管制署訂定之醫療照護相關感染(Health care–associated infection, HAI)判定定義導入AI系統，協助判斷是否符合收案標準進行收案，提升即時性的同時，使臨床可以提早介入各項感染管制防治措施；將監測結果運用PBI(Power BI)資料視覺化工具，呈現醫療照護相關感染的變化趨勢，協助檢討分析，迅速而有效的協助感染管制決策與精準回饋。未來整合AI預測模型、快速微生物診斷與標準化照護流程，將有助於建立更高效、安全且成本效益佳的感染管理系統，持續提升臨床照護品質。

Bloodstream infections (BSI) represent a major category of healthcare-associated infections. This medical condition is characterized by high mortality and an increased risk of prolonged hospitalization, particularly when associated with the use of central venous catheters (CVC). To enhance detection efficiency and the speed of clinical response, artificial intelligence (AI) technologies have been widely introduced into infection surveillance, prediction, and decision-support systems, emerging as a new trend in healthcare quality management. Internationally, institutions such as Mayo Clinic and Kaiser Permanente in the United States have developed AI algorithms capable of automatically analyzing electronic medical records and laboratory data to identify high-risk cases of central line-associated bloodstream infection (CLABSI), thereby providing real-time clinical alerts. In Asia, countries including Singapore and South Korea have incorporated natural language processing (NLP) and surveillance models into clinical record systems, which has effectively improved the accuracy of risk prediction. In Taiwan, the medical community has also progressively expanded the application of AI in infection control. For example, emergency departments have implemented a “sepsis AI model” to rapidly stratify patients by sepsis risk levels, enabling physicians to respond promptly. Hospitals have also integrated data from electronic medical record systems and applied the Taiwan Centers for Disease Control’s definitions for healthcare–associated infection (HAI) into AI systems to assist in case determination, thereby improving timeliness and enabling earlier clinical interventions with appropriate infection prevention and control measures. In addition, surveillance results have been visualized using the Power BI platform, facilitating the presentation of HAI trends, supporting review and analysis, and providing rapid, accurate feedback to guide infection control decision-making. In the future, the integration of AI prediction models, rapid microbial diagnostics, and standardized care pathways will contribute to the establishment of a more efficient, safer, and cost-effective infection management system, thereby continuously enhancing the quality of clinical care.