到院前心肺功能停止之資料分析

Data Analysis of Out-of-Hospital Cardiac Arrest

簡佳苓(Chia-Ling Chien)

指導教授 ： 黃乾綱

國立臺灣大學/工學院/工程科學及海洋工程學研究所/碩士(2014年)

https://doi.org/10.6342/NTU.2014.11195

摘要

「到院前心肺功能停止」（Out-of-Hospital Cardiac Arrest, OHCA）定義為病患在被送到醫院前，呼吸、心跳皆已經停止者，其相關指標不論台灣或國外的研究，皆為當今緊急醫療體系評估的重要基礎。目前關於OHCA資料之相關文獻，大多僅使用傳統生物統計方式──Utstein Style，但都有其侷限性，且因國家地域和實施救護決策之不同，而有不一樣的爭議。如何挑出真正重要的影響因子和釐清其錯縱複雜的因果關係，尚未在新北市OHCA資料庫驗證。因此本研究針對新北市OHCA資料庫從2010~2013年間的11,010筆案件進行相關資料分析，讓研究結果更具可靠性及代表性。
　　本研究分為兩大部分：流行病學分析及資料探勘分析。流行病學方面，以生物統計方式──卡方檢定（Chi-squre）和勝算比（Odds Ratio）分析各項因子和存活率之間的關係，發現其他文獻未確認之重要特徵，再進一步建立OHCA地理資訊系統，進行空間和時間軸資料分析，提供醫療人員一個快速查詢視覺化資料分布的平台，更有助於基礎的統計分析和後續決策。資料探勘方面則打破傳統方式，運用貝氏網路（Bayesian network）技術，探討處置是否有其必要性，如使用喉罩（LMA）、插管（ETI）、施行基礎急救術（Basic Life Support, BLS）或高級急救術（Advanced Life Support, ALS）對於存活的影響。研究結果顯示，特徵間彼此相依性甚高，在內科不建議電擊下，ALS相對於BLS服務勝算比達1.19和1.25倍，在救護時段處置時間、送醫時間、整體時間拉長時，勝算比更可以高達1.50倍左右，表示時間越長，實施ALS服務更有顯著性的效果和必要性。
　　由於目前缺乏完善的救護決策和資源分配，在資源有限的情況下，希望利用這一套的分析方法，找出潛在有用的資訊知識，利用其探勘結果提供決策分析的參考資訊，期望對於醫學領域增加資訊化的價值，提升OHCA存活率及保障需要緊急救護服務的人民。

關鍵字

緊急醫療救護 ； 地理資訊 ； 貝氏網路 ； 到院前心肺功能停止

並列摘要

Out-of-Hospital Cardiac Arrest (OHCA) is defined as the patients before they were taken to the hospital, breathing, heartbeat have stopped. Whether it’s related indicators in Taiwan or abroad, is an important foundation of today's emergency medical service. The current literatures on the OHCA mostly use only the traditional method of Utstein Style, but it has some limitations. Therefore, how to pick out the really important factor and to clarify its complicated causal relationship has not been validated in New Taipei City OHCA database. This study used 11,010 cases between 2010 and 2013 for data analysis.
The study is divided into two parts: Epidemiological Analysis and Data Mining Analysis. First, we used Chi-square test and Odds Ratios to analyze the relationship between various factors and survival rate, and further established OHCA Geography information systems, spatial data and timeline analysis provide a quick check of medical personnel to visualize data distribution platform. Second, breaking the traditional way, we used Bayesian network technology to explore the necessity of the intervention. The results show that under non-trauma and non-shockable people, Advanced Life Support relative to Basic Life Support for survival rate, the odds ratio is 1.19 and 1.25 times. And in the handling time, run time, total time, when the time is the longest, the overall odds ratio can be as high as 1.50 times more, which means that the longer the time, the implementation of ALS services has more significant effect and necessity.

並列關鍵字

emergency medical services ； geographic information ； Bayesian network ； Out-of-Hospital Cardiac Arrest

參考文獻

4. Chang, M.-Y. and M. Lin, Predictors of survival and hospital outcome of prehospital cardiac arrest in southern Taiwan. Journal of the Formosan Medical Association= Taiwan yi zhi, 2005. 104(9): p. 639-646.

7. Lin, J.N., et al., Analysis of factors associated with successful cardiopulmonary resuscitation in non-traumatic dead-on-arrival patients in emergency department. Kaohsiung J Med Sci, 2002. 18(2): p. 84-90.

2. Turakhia, M. and Z.H. Tseng, Sudden cardiac death: Epidemiology, mechanisms, and therapy. Current Problems in Cardiology, 2007. 32(9): p. 501-546.

3. Cooper, J.A., J.D. Cooper, and J.M. Cooper, Cardiopulmonary resuscitation: history, current practice, and future direction. Circulation, 2006. 114(25): p. 2839-49.

6. Hu, S.C., Out-of-Hospital Cardiac-Arrest in an Oriental Metropolitan City. American Journal of Emergency Medicine, 1994. 12(4): p. 491-494.