在世界各國醫療照護系統中，醫療浪費是普遍存在的問題，台灣自1995年實施全民健保之後，中央健康保險局虧損的問題也日益嚴重，在2002年實施醫療費用總額支付制度之後，支出上限的設定令醫院在成本上需要更準確的控制。醫院手術室約佔醫院約40%的收入與花費來源，為醫院最大的成本中心，手術排程為手術室重要的績效指標，本研究致力於改善手術超時預測之方法研究。本研究運用類神經網路以台北某醫學中心之手術資料建構手術超時模型，結果發現手術醫師年資、麻醉醫師年資、手術助手數量、是否為急診病患、病患是否住ICU加護病房、是否為假日病患、是否為住院病患、是否為麻醉病患、是否為緊急麻醉病患、是否為全身麻醉病患、病患性別、手術時間是否加班時間、以及科別最適合建構類神經網路，本研究建構出的網路ROC 值為0.637，和貝氏分類器比較，類神經網路有較好的效度。本研究所建構之醫療手術超時預測模型，將能提供手術排程人員在手術排程上的判斷依據。 Wasting in health care spending has become a common issue in health care systems worldwide. After the implementation of the national health insurance system in 1995, the Bureau of National Insurance’s deficit has also becoming a serious issue in Taiwan. With the implementation of the global budget System the budget of hospitals are limited by expenditure cap which increase the needs of more precise cost control for the managers in hospital. Operating rooms have been estimated to account for more than 40% of a hospital’s total revenues thus the largest cost center in a hospital (Denton. Et al. 2006). Scheduling is an important factors of operating room performance. This paper focus on predicting surgery overtime. The data set used for developing and testing the neural network was collected from a medical center in Taipei. We discovered that Years of service of the surgeon, Years of service of the anesthetist, numbers of the assistant during operation, emergency operation, emergency anesthesia, full body anesthesia, anesthesia, inpatient, ICU, overtime, sex of patients and sections are best fit for developing the neural network. The performance of the network was also evaluated by calculating the area under the Receiver Operating Characteristic (ROC) curve in the test set, 0.637. Compare with Na?ve Bayesian Classifier, the Neural network has better validity. We believe the surgery overtime prediction model we developed would serve as a prediction aid for scheduling surgery operations.
