運用決策樹建立骨科手術時間預測模型－以某區域教學醫院為例

Using the Decision Tree Establishment Orthopedics Surgery Time Forecast Model: Take Some Region Teaching Hospital as the Example

劉翠燕(Tsui-Yen Liu) ； 林伯堅(Bojain Lin) ； 吳文祥(Wen-Hsiang Wu)

《源遠護理》 13卷3期 (2019/11) Pp. 31-40

https://doi.org/10.6530/YYN.201911_13(3).0005

摘要

背景：手術排程會直接影響手術室的使用效率及營運效益，而手術時間預測是手術排程系統最重要的資訊，大部分醫院手術系統的時間以醫師經驗決定排程，與實際時間產生落差。目的：掌握影響手術時間的變數，才能估計出合宜的手術時間以作為排程決策之參考。方法：運用資料探勘方法中的決策樹找出影響手術時間的變數，由於系統要能夠估計醫師預排手術時間的可能性，必須先計算其機率，透過Beta分佈演算出時間參數，掌握α及β估計值可計算出手術起始時間及其機率，再與醫師實際手術時間相互驗證。結果：找到影響手術時間的變數包括手術術式、麻醉方式、所使用的骨材等。結論與實務運用：實際運用在臨床上醫師的手術時間排程，以此模型估計出排程手術時間的機率大小及是否在估計時間的合理範圍內，經由臨床實際情況不斷的修正將排程改善至效率最佳化。

關鍵字

手術時間 ； 決策樹分析 ； Beta分佈

並列摘要

Background: Surgical scheduling directly affects the efficiency of surgeons' operating and the oper¬ation room. Surgery time prediction is the most important information of the surgical scheduling system. Most hospital surgical time in the systems is determined by the doctor's experience; therefore, the actual time has wide variation. Purpose: To grasp the variables that affect the operation time, to better estimate the appropriate operation time to make better scheduling decisions. Methods: Use the decision tree in the data exploration to identify key variables that affect the operation time. The system needs to collect time estima¬tion before surgery to compare with actual time of the surgery to establish the predication model. Through the Beta distribution calculation time parameters, use the α and β numerical estimation calculus to compare predicted end time with actual end time of each surgeon. Results: The variables that affected the time of surgery included surgical procedures, anesthesia methods, physicians, and bone materials used. Conclusion and practical application: To better arrange surgeons' surgery schedule base on the estimated probability of operation time, to know if the surgery is within the reasonable range of estimated time, and to improve the scheduling through continuous correction of clinical actual situation. To optimize efficiency.

並列關鍵字

surgery time ； decision tree ； Beta distribution

參考文獻

Mei, T. C. (2016). A study on the effect on scheduling and management of surgeries with the introduction of excellent medical information. Studies in Health Technology and Informatics, 225, 824-825. doi:10.3233/978-1-61499-658-3-824

江淑如（2011）．以法則式分類器預測手術時間 ·碩士論文，中興大學資訊科學與工程學系研究所學位論文，台中市。

林怡君（2003）·運用模擬技術於手術室排程管理—以某醫學中心為例．碩士論文，臺灣大學醫療機構管理研究所，台北市。

林重賢（2002）．手術時間預測模式建立．碩士論文，臺灣大學醫療機構管理研究所學位論文，台北市。

林承俊、胡雅涵（2017）．運用資料探勘技術建構植牙手術成敗預測模型．醫學與健康期刊，6（2），57-69。