急診壅塞即時監測系統之實證性研究

研究背景與動機目前國內各大醫院急診壅塞之情況極為嚴重，但並無統一之標準可供界定何謂急診壅塞，亦無良好之監測系統可以顯示急診室壅塞之即時情況與壅塞成因，作為醫院或主管機關因應之參考。本研究試圖藉由建構目標醫學中心急診室的即時性數據蒐集與運算程式，對常用之各項急診壅塞指標作分析與驗證。以目標醫學中心通報急診滿床記錄，作為壅塞參考基準，並經由比對找出最能反應該院急診壅塞之指標，以建構急診壅塞之即時監測系統，且作為將來監測國內急診壅塞與因應策略之參考。研究方法經由急診各項人力與需求指標的即時蒐集，以電腦同步轉換為各項壅塞指標，並與官方急診壅塞資料—高雄市緊急醫療資訊整合中心（Emergency Medical Operation Center，EMOC）通報壅塞紀錄比對，測試各指標之準確率。資料蒐集比對完成後，以ROC 曲線(Receiver Operating Characteristic Curve)分析，衡量目前常用的壅塞指標何者對於目標醫學中心壅塞資料之解釋程度較高。 II 研究結果目前常用之急診壅塞指標中，以EDWIN、Work Score、Occupancy Level 對目標醫學中心壅塞的解釋程度較高。經比較分析，建議以 EDWIN 為主建構目標醫學中心之急診壅塞即時監測系統，並輔以 Work Score 與Occupancy Level 以提高系統監測之準確率。研究中訂出各指標之最佳門檻值，分別為：EDWIN 0.884259259259259、Work score：11.106、Occupancy Level：0.865546218487395。結論與建議 EDWIN、Work score、Occupancy Level 最適宜用以監測目標醫學中心急診室是否發生壅塞，並可以EDWIN 為主，輔以Occupancy Level 及Work Socre 建構急診壅塞監測與預警系統，以提供醫院管理階層監控與因應時之決策輔助。未來並可考慮推廣及建構區域內緊急醫療資源監測與通報系統，作為緊急救護時之資源分配、病患送治與救護車分流參考。

關鍵字

急診壅塞壅塞指標；即時監控；救護車分流EDWIN ； Work score ； NEDOCS ； Occupancy Level

並列摘要

Background: Emergency Department (ED) crowding is a global problem with increasing intensity in Taiwan. However, there is no universal definition or criteria about crowding, and real-time surveillance system is also absent. Therefore, it is difficult in forming the strategy of solution or even responding to this problem. Throughout a real-time collecting and analyzing program of the crowding-related data, the purpose of the study is to verify six commonly-used ED crowding index by the official databank, the Emergency Medical Operation Center（EMOC）crowding announcement records, and looks for the most effective ones in monitoring ED crowding of the target ED. The study also aims on developing a real-time surveillance system for the target ED for more effective crowding monitoring, and hopefully that it can also be used in other hospital. With district-wide employment of this real-time surveillance system in most hospital, the information can provide a more efficient utilization of the emergency medical system and helps to save life and resources. Method: The related data from target ED was collected automatically in a 10-minute period from the Hospital information System（HIS）and transformed into 6 ED crowding indices. These indices were then correlated with the official crowding records of the EMOC and their sensitivity and specificity in relationship to occurrence of crowding was marked. Then we use the Receiver Operating Characteristic (ROC) Curve analysis to identify the most useful indices. Results: EDWIN、Work score、Occupancy Level have the highest interpretative ability to the target ED. After general consideration, EDWIN may be used as the main index in building the real-time surveillance system for the target ED, assisting by Occupancy level and Work score for more accuracy. The cut-off value for each index are: EDWIN 0.884259259259259,Work score 11.106, and Occupancy Level 0.865546218487395. Conclusion: The commonly-used ED crowding indices can be used to monitor the occurrence of crowding in target ED and build up a real-time surveillance system for the target ED.