某地區醫院運用管制圖尋找影響全人工膝關節手術費用之研究

A Research by a Certain District Hospital Using Control Charts to Find out the Influences on Fees for Total Knee Arthroplasty Surgery

我國於2018年正式邁入高齡化社會，造成2000年至2017年接受全人工膝關節置換手術增加26％，健保署則透過定額支付制度來減少財政支出壓力，個案醫院因此藉由實證醫學修正臨床路徑而改善全人工膝關節置換手術後血液血漿點數高之現況。首先運用個案醫院2016年接受全人工膝關節置換手術共320筆資料，運用資料探勘－決策樹分析之CART演算法發現4個年齡層會影響手術時間。透過變異數分析結果得知，特殊材料點平均數差異最大，故再考量是否使用自費特殊材料後，可依此8群透過個別值－移動全距（I-MR）管制圖觀察總醫療點數後，發現17筆異常個案中有13筆是血液血漿點數偏高。最後，驗證實證醫學證據於關節內使用Tranexamic acid（TXA）可減輕此問題，血液血漿點數平均下降962點（83％）、總醫療點數平均下降1,341點（1％）。本研究運用決策樹分組、透過管制圖監控醫療點數並找出變異之過程可有效管控醫療耗用，更重要的是驗證臨床路徑改善血液血漿點數之成效，挑戰更高醫療品質。

Taiwan officially became an aged society in 2018, resulting in an increase of 26% in annual numbers of total knee arthroplasty surgeries performed from 2000 to 2017. Under such circumstances, National Health Insurance Administration tried to reduce the increasing pressure on fiscal expenditure by adapting a fixed amount payment system. This very hospital, therefore, did try to improve the current situation of high blood plasma points after total knee arthroplasty surgery by modifying the clinical pathway through empirical medicine. First, the Classification And Regression Tree algorithm of decision tree analysis was applied to a total of 320 total knee arthroplasty surgeries performed at the hospital in 2016, using data mining- decision tree, four age groups affecting the operation time were identified. According to the analysis of variance result, it is known that the special material points have the largest difference in means. Further, by taking into account whether the special materials were paid by patients, eight groups were found for monitoring the total medical points. After observing the total medical points of each group through Individual-moving range control chart, we found 13 higher blood plasma point cases out of 17 deviant cases. Lastly, with the evidence-based medicine, we verified that using Tranexamic acid in the articulation could rectify the situation. The rectified blood plasma points decreased by 962 points (83%) and the total medical points decreased by 1,341 points (1%) on average. This study used the grouping of decision tree, monitored the medical points through the control chart and found out the process of variation, which could effectively manage the medical utilization and, more importantly, reduce the blood transfusion as well as medical care hours. Thus due to its success, it looks that similar trials could be implemented in other medical care procedures for further improvements.