臺灣博碩士論文加值系統

研究背景: 手術室是一個高成本、高營收的醫療單位，隨著人口高齡化影響造成平均餘命延長，在世界各地的醫院都面臨手術需求增加的趨勢，手術室效率會影響醫院營運的績效與服務品質。找出最佳的手術時間預測模型來準確預估手術的時間，增加手術室的效率及利用率以提升醫院的收入，是目前手術管理的目標。
研究目的:建置手術時間預測模型，透過預測模型的導入，找出最貼近真實的手術時間，以提升手術室最大的效率。
研究方法:本研究以某區域醫院2021年1月1日至2022年6月30日手術室使用時間登錄資料進行分析，篩選出骨科常見手術術式。第一階段以敘述統計，分析影響手術時間之因素包括病人年齡、性別、來源、執刀醫師、醫師預測時間、手術術式、麻醉方式、ASA麻醉風險等級等;第二階段則運用線性回歸、決策樹建置模型進行預測分析，並以決策樹為基礎建置分位數回歸預測模型，最終透過預測評估指標RMSE均方誤差比較三種預測模型準確性，找出本研究醫院最適合的手術時間模型。
研究結果:本研究共有13691筆資料，為處理手術科別間變異性大問題，本研究以區域醫院佔比最大之骨科手術為主，經條件篩選後共納入866筆。敘述統計結果中，實際手術時間為155.98±56.025分鐘，醫師預估手術時間為91.66±38.108分鐘，研究結果發現，透過線性回歸分析發現最有影響力之因子為手術術式、執刀醫師及醫師預估時間;在決策樹的部分最有影響力之因子為手術術式以及執刀醫師。最後透過預測評估指標 RMSE比較後發現，分位數回歸模型之RMSE為5.62，決策樹預測模型之RMSE為30.93，線性回歸模型之RMSE為35.35，分位數回歸模型有最低的RMSE。
研究結論:
(1)本研究實證提出一個以決策樹為基礎之分位數回歸預測模型。
(2)在本研究模型比較後，分位數回歸模型為最佳預測手術時間模型。
(3)若能掌握造成手術時間異常關鍵因素，預測手術時間模型可取代醫師手術預測方法。
(4)本研究提出之骨科手術時間預測模型，可以提供未來相關手術時間預測模型建置框架，外推應用至該研究醫院其他科別手術預測。

Background: The Operating room (OR) are high-cost, high-revenue healthcare units. With the influence of an aging population and increased life expectancy, hospitals worldwide are facing a growing demand for surgeries. The efficiency of operating rooms affects the performance and service quality of hospitals.
Objective: The objective of this study is to identify the optimal surgical time prediction model to accurately estimate surgical durations, thereby enhancing the efficiency of operating rooms and increasing hospital revenue.
Method: This study analyzed surgical time registration data from a regional hospital in a specific area between January 1, 2021, and June 30, 2022. Common orthopedic surgical procedures were selected for analysis. In the first stage, descriptive statistics were used to analyze factors affecting surgical time, including patient age, gender, referral source, operating surgeon, physician-predicted time, surgical procedure, anesthesia method, and ASA anesthesia risk level. In the second stage, linear regression and decision tree models were employed for predictive analysis. A quantile regression prediction model was built based on the decision tree. Finally, the root mean square error (RMSE) was used to compare the accuracy of the three prediction models and identify the most suitable surgical time model for the study hospital.
Results: A total of 13,691 data points were collected for this study. Due to the large variability between surgical departments, the focus was placed on orthopedic surgeries, which accounted for the largest proportion in the regional hospital. After applying the inclusion criteria, 866 data points were included. Descriptive statistics revealed that the actual surgical time was 155.98±56.025 minutes, while the physician-predicted time was 91.66 ±38.108 minutes. The study found that the surgical procedure, operating surgeon, and physician-predicted time were the most influential factors through linear regression analysis. For the decision tree model, the most influential factors were the surgical procedure and operating surgeon. Comparing the RMSE values, the quantile regression model achieved the lowest value of 5.62, followed by the decision tree model with 30.93, and the linear regression model with 35.35.
Conclusion:
(1) This study proposes a quantile regression prediction model based on a decision tree.
(2) Among the models compared in this study, quantile regression model is the best for predicting surgical time.
(3) If key factors causing abnormal surgical time can be identified, prediction model can replace physician prediction methods.
(4) The orthopedic surgical time prediction model proposed in this study can provide a framework for future development of related surgical time prediction models and can be extrapolated for surgical prediction in other departments of the research hospital.

致謝............................................................................................................................................ i
中文摘要...................................................................................................................................ii
Abstract.....................................................................................................................................iii
目錄........................................................................................................................................... v
圖目錄...................................................................................................................................... vi
表目錄.....................................................................................................................................vii
第一章 緒論........................................................................................................................... 1
第一節 研究背景與動機............................................................................................... 1
第二節 研究目的與重要性........................................................................................... 2
第二章 文獻探討................................................................................................................... 3
第一節 手術室管理.......................................................................................................... 3
第二節 手術時間預測模型.............................................................................................. 5
第三章 研究設計與方法......................................................................................................... 9
第一節 研究架構.............................................................................................................. 9
第二節 研究對象與資料來源........................................................................................ 10
第三節 研究變項與操作型定義.................................................................................... 11
第四節 資料處理流程與統計分析方法........................................................................ 14
第四章 研究結果................................................................................................................... 16
第一節 手術資料基本特質分佈.................................................................................... 16
第二節 線性回歸與決策樹手術時間預測.................................................................... 18
第三節 分位數回歸手術時間預測................................................................................ 39
第四節 手術時間預估模型之誤差比較........................................................................ 60
第五章 討論與建議............................................................................................................... 63
第一節 研究發現............................................................................................................ 63
第二節 研究限制............................................................................................................ 65
第三節 研究結論與建議................................................................................................ 66
第六章 參考文獻................................................................................................................... 67

英文文獻
Ronald Gabel, John Kulli, B. Stephen Lee, Deborah G. Spratt, Denham S. Ward .(1999). Operating room management,Butterworth-Heinemann.
Eijkemans, M. J., Van Houdenhoven, M., Nguyen, T., Boersma, E., Steyerberg, E. W., & Kazemier, G. (2010). Predicting the unpredictable: a new prediction model for operating room times using individual characteristics and the surgeon's estimate. The Journal of the American Society of Anesthesiologists, 112(1), 41-49
Devi, S. P., Rao, K. S., & Sangeetha, S. S. (2012). Prediction of surgery times and scheduling of operation theaters in optholmology department. Journal of medical systems, 36, 415-430.
van Veen-Berkx, E., Bitter, J., Elkhuizen, S. G., Buhre, W. F., Kalkman, C. J., Gooszen, H. G., & Kazemier, G. (2014). The influence of anesthesia-controlled time on operating room scheduling in Dutch university medical centres Linfluence du temps controle par lanesthesie sur le programme operatoire dans les centres medicaux dune universite neerlandaise. Can J Anaesth, 61, 524-532.
Hosseini, N., Sir, M. Y., Jankowski, C. J., & Pasupathy, K. S. (2015). Surgical duration estimation via data mining and predictive modeling: a case study. AMIA annual symposium proceedings, 2015, 640.
Wu, H.-L., Chang, W.-K., Hu, K.-H., Langford, R. M., Tsou, M.-Y., & Chang, K.-Y. (2015). A quantile regression approach to estimating the distribution of anesthetic procedure time during induction. Plos one, 10(8), e0134838.
Xiang, W., Yin, J., & Lim, G. (2015). A short-term operating room surgery scheduling problem integrating multiple nurses roster constraints. Artificial intelligence in medicine, 63(2), 91-106.
Edelman, E. R., Van Kuijk, S. M., Hamaekers, A. E., De Korte, M. J., Van Merode, G. G., & Buhre, W. F. (2017). Improving the prediction of total surgical procedure time using linear regression modeling. Frontiers in medicine, 4, 85.
Tuwatananurak, J. P., Zadeh, S., Xu, X., Vacanti, J. A., Fulton, W. R., Ehrenfeld, J. M., & Urman, R. D. (2019). Machine learning can improve estimation of surgical case duration: a pilot study. Journal of medical systems, 43, 1-7.
Martinez, O., Martinez, C., Parra, C. A., Rugeles, S., & Suarez, D. R. (2021). Machine learning for surgical time prediction. Computer Methods and Programs in Biomedicine, 208, 106220.
Yuniartha, D. R., Masruroh, N. A., & Herliansyah, M. K. (2021). An evaluation of a simple model for predicting surgery duration using a set of surgical procedure parameters. Informatics in Medicine Unlocked, 25, 100633.
Yeo, I., Klemt, C., Melnic, C. M., Pattavina, M. H., De Oliveira, B. M. C., & Kwon, Y.-M. (2022). Predicting surgical operative time in primary total knee arthroplasty utilizing machine learning models. Archives of Orthopaedic and Trauma Surgery, 1-9.
Miller, L. E., Goedicke, W., Crowson, M. G., Rathi, V. K., Naunheim, M. R., & Agarwala, A. V. (2023). Using machine learning to predict operating room case duration: A case study in otolaryngology. Otolaryngology–Head and Neck Surgery, 168(2), 241-247.

中文文獻
林重賢（2002）。手術時間預測模式建立。未出版之碩士論文，國立臺灣大學醫療機構管理研究所，台北市。
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。未出版之碩士論文，國立臺灣大學醫療機構管理研究所，台北市。
陳德芳（2006）。建立手術時間預測模式來從事電腦化手術室排程。未出版之碩士論文，國立臺灣大學醫療機構管理研究所，台北市。
童麗清（2013）。運用系統模擬技術縮短手術房病人等候時間之研究—以中部某區域教學醫院手術室為例。未出版之碩士論文，東海大學工業工程與經營資訊學系，台中市。
劉翠燕, 林伯堅, & 吳文祥（2019）。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。源遠護理， 13（3），頁 31-40。
吳慧雯（2020）。探討影響東部區域教學醫院手術排程相關性分析。未出版之碩士論文，長庚大學商管專業學院碩士學位學程在職專班醫務管理組，桃園縣。
莊子葳（2022）。手術室使用時間分析－以某區域醫院為例。未出版之碩士論文，國立陽明交通大學醫務管理研究所，新竹市。