臺灣博碩士論文加值系統

用藥安全儼然已成為病人安全的關鍵性指標之一。近年來，有關於抗凝血劑在使用上的安全性，引起美國醫療機構評鑑聯合會(JCAHO)相當大的關注。而在所有的抗凝血劑中，Warfarin一直都是美國藥物不良事件(ADE)排行中的前十大藥品之ㄧ。
本研究採用多種監督式學習技術來建立Warfarin給藥劑量預測模式，包括類神經網路(ANN)、模式樹(M5)、支援向量機(SVM)、以及k-最鄰近演算法(kNN)。同時，為了提昇預測準確率，本研究進一步將Bagging效能提昇技術與多重分類器的概念納入實驗中。而在效能評估的部分，則是以醫師所開立的Warfarin劑量為比較基準。實驗資料來源為587位曾接受Warfarin治療的住院病患之病歷。
最後，整體評估結果顯示，無論何種技術的預測效能都明顯優於醫師預測之準確度。在所有預測模式中，Bagging整合多重分類技術所建構之分類器預測效能最佳，因為此模式之平均絕對誤差(MAE)及標準差皆為最低。整體而言，本研究所建構之預測模式不僅能支援醫師給藥劑量之決策，更能有效降低藥物不良事件的發生率。

Drug safety has become one of the most critical indicators in patient safety. In recent years, the Joint Commission on Accreditation of Healthcare Organizations (JCAHO) has paid considerable attention to the safety of anticoagulant administration. Among all anticoagulants, Warfarin has long been listed among the top ten drugs causing adverse drug events (ADE).
This paper investigates a number of supervised learning techniques, including artificial neural network (ANN), model tree (M5), support vector machines (SVM), and k-nearest neighbors (kNN), to construct a Warfarin dosage prediction model. To achieve higher prediction accuracy, we further consider both Bagging algorithms and ensemble classifiers in experimental evaluation. For performance evaluation, the initial dose of Warfarin prescribed by clinicians is established as the baseline of our study. The experimental data collection consists of complete historical records of 587 inpatients who received Warfarin treatment.
The overall evaluation results show that all of the learning based systems are significantly more accurate than the baseline. Among all prediction models, Bagging VOTE with four classifiers is suggested as the most effective prediction model due to its lower mean absolute errors (MAE) and variation of errors. The investigated models can not only facilitate clinicians in dosage decision-making, but also help reduce patient risk from adverse drug events.

第一章 緒論………………………………………………………………1
1.1 研究背景……………………………………………………………1
1.2 研究動機……………………………………………………………2
1.3 研究目的……………………………………………………………4
第二章 文獻探討…………………………………………………………6
2.1 用藥安全……...……………………….………………….…….6
2.1.1用藥安全與臨床藥物動力監測……………………………….…6
2.1.2決策支援系統於用藥安全之應用……………………………….7
2.2 抗凝血劑-Warfarin………………………………………………….......10
2.2.1 Warfarin藥物介紹………………………...……………….10
2.2.2 Warfarin抗凝血強度監測…………………………………….11
2.2.3 INR理想目標範圍………………………………………...….12
2.2.4 Warfarin之常見副作用……………………………………….…12
2.3 資料探勘技術………………………………………………...………13
2.3.1 類神經網路……………………………………………………..13
2.3.2 模式樹…………………………………………………………..18
2.3.3 支援向量機……………………………………………………..19
2.3.4 kNN演算法……………………………………………………..21
2.3.5 Bagging分類效能提昇技術……………………………………21
第三章 研究方法………………………………………………………………..23
3.1 資料收集與研究變數定義…………………………………………...23
3.2 資料前處理……………………………………...……………………24
3.3 研究架構…………………………………...…………………………27
3.4 參數設定…………………………………………………………...…30
3.5 評估程序與準則……………………………………………………...30
第四章 實驗結果與分析………………………………………………………..32
4.1 各單一分類器與基準值效能評估…………………………………...32
4.2 結果評估：DS-all資料集……………………………………………33
4.3 結果評估：DS-Inter資料集…………………………………………34
4.4 結果評估：DS-W.O. inter資料集……………………………………35
4.5 綜合討論……………………………………………………………...36


第五章 研究結論與建議………………………………………………………..37
5.1 研究結論……………………………………………………………...37
5.2 未來研究方向與建議………………………………………………...38
參考文獻…………………………………………………………………………39
附錄………………………………………………………………………………44
附錄一 ANN參數設定與說明…………………………………………...44
附錄二 SVM參數設定與說明…………………………………………...45

英文部分
Ackroyd-Stolarz, S., Hartnell, N., & MacKinnon, N. J. (2005). Approaches to improving the safety of the medication use system. Healthcare Quarterly, 8, 59-64.
Aha, D. W., Kibler, D., & Albert, M. K. (1991). Instance-based learning algorithms. Machine Learning, 6(1), 37-66.
Ansell, J., Hirsh, J., Hylek, E., Jacobson, A., Crowther, M., & Palareti G. (2008). Pharmacology and management of the vitamin K antagonists: American College of Chest Physicians Evidence-Based Clinical Practice Guidelines (8th ed.). Chest, 133, 160S-198S.
ASHP. (1993). ASHP Report:guidelines on preventing medication errors in hospitals. American Journal of Health-System Pharmacy, 50, 305-314.
ASHP. (1998). ASHP statement on the pharmacist's role in clinical pharmacokinetic monitoring. American Journal of Health-System Pharmacy, 55(16), 1726-1727.
Bates, D.W., Cohen, M., Leape, L. L., Overhage, J. M., Shabot, M. M., & Sheridan, T. (2001). Reducing the frequency of errors in medicine using information technology. Journal of the American Medical Informatics Association, 8(4), 299-308.
Bates, D. W., Teich, J. M., Lee, J., Seger, D., Kuperman, G. J., & Boyle, D. (1999). The impact of computerized physician order entry on medication error prevention. Journal of the American Medical Informatics Association, 6(4), 313-321.
Bobb, A., Gleason, K., Husch, M., Feinglass, J., Yarnold, P. R., & Noskin, G. A. (2004). The epidemiology of prescribing errors: the potential impact of computerized prescriber order entry. Archives of Internal Medicine, 164(7), 785-792.
Breiman, L. (1996). Bagging predictors. Machine Learning, 24(2), 123-140.
Breiman, L., Friedman, J. H., Olshen, R. A., & Stone, C. J. (1984). Classification and Regression Trees. Belmont, California, CA: Wadsworth International Group.
Byrne, S., Cunningham, P., Barry, A., Graham, I., Delaney, T., & Corrigan, O. I. (2000). Using neural nets for decision support in prescription and outcome prediction in anticoagulation drug therapy. Intelligent Data Analysis In Medicine and Pharmacology, a Workshop at the 14th European Conference on Artificial Intelligence.
Camps-Valls, G., Porta-Oltra, B., Soria-Olivas, E., Martin-Guerrero, J. D., Serrano-Lopez, A. J., Perez-Ruixo, J. J., et al. (2003). Prediction of cyclosporine dosage in patients after kidney transplantation using neural networks. IEEE Transactions on Biomedical Engineering, 50(4), 442-448.
Chipman, H. A., George, E. I., & McCulloch, R. E. (1998). Bayesian CART model search. Journal of the American Statistical Association, 93(443), 935-960.
Colombet, I., Bura-Rivière, A., & Chatila, R. (2004). Personalized versus non-personalized computerized decision support system to increase therapeutic quality control of oral anticoagulant therapy: an alternating time series analysis. BMC Health Services Research, 4(1), 27.
Dean, B., Schachter, M., Vincent, C., & Barber, N. (2002). Prescribing errors in hospital inpatients: their incidence and clinical significance. Quality & Safety in Health Care, 11(4), 340-344.
Dzeroski, S., & Zenko, B. (2004). Is combining classifiers with stacking better than selecting the best one? Machine Learning, 54(3), 255-273.
Ernst, F. R., & Grizzle, A. J. (2001). Drug-related morbidity and mortality: updating the cost-of-illness model. Journal of the American Pharmaceutical Association, 41(2), 192-199.
Fitzgerald, R. J. (2009). Medication errors: the importance of an accurate drug history. British Journal of Clinical Pharmacology, 67(6), 671-675.
Floares, A. G., Floares, C., Cucu, M., Marian, M., & Lazar, L. (2004). Optimal drug dosage regimens in cancer chemotherapy with neural networks. Journal of Clinical Oncology, 22(14), 2134.
Freeman, J. A., & Skapura, D. M. (1992). Neural networks algorithms, applications, and programming techniques. New York: Addison-Wesley Publishing Company.
Freund, Y., & Schapire, R. (1997). A decision-theoretic generalization of on-line learning and an application to boosting. Journal of Computer and System Sciences, 55(1), 119-139.
Gouveia, W. A., Shane, R., & Clark, T. (2003). Computerized prescriber order entry: power, not panacea. American Journal of Health-System Pharmacy, 60(18), 1838.
Horton, J. D., & Bushwick, B. M. (1999). Warfarin therapy: evolving strategies in anticoagulation. American family physician, 59(3), 635-646.
Hu, P. J., Cheng, T. H., Wei, C., Yu, C. H., Chan, A., & Wang, H. Y. (2007). Managing clinical use of high-alert drugs: a supervised learning approach to pharmacokinetic data analysis. IEEE Transactions on Systems, Man, and Cybernetics-Part A: Systems and Humans, 37(4), 481-492.
ISMP. (2008). ISMP’s list of high-alert medications. Retrieved August 18, 2010, from http://www.ismp.org/Tools/highalertmedications.pdf
Kanjanarat, P., Winterstein, A. G., & John, T. E. (2003). Nature of preventable adverse drug events in hospitals: a literature review. American journal of health-system pharmacy, 60(17), 1750-1759.
Kaushal, R., Shojania, K. G., & Bates, D. W. (2003). Effect of computerized physician order entry and clinical decision support systems on medication safety: a systematic review. Archives of Internal Medicine, 163(12), 1409-1416.
Kretschmann, E., & Apweiler, R. (2001). Automatic rule generation for protein annotation with the C4.5 data-mining algorithm applied on peptides in Ensembl. Proceedings of the German Conference on Bioinformatics (pp. 53-57). Braunschweig, Germany.
Lassetter, J. H., & Warnick, M. L. (2003). Medical errors, drug-related problems, and medication errors: a literature review on quality of care and cost issues. Journal of Nursing Care Quality, 18(3), 175-183.
Leape, L. L., Bates, D.W., Cullen, D. J., Cooper, J., Demonaco, H. J., Gallivan, T., et al. (1995). System analysis of adverse drug events. ADE Prevention Study Group. Journal of the American Medical Association, 274(1), 35-43.
Leape, L. L., & Berwick, D. M. (2005). Five years after “To Err Is Human”: what have we learned? Journal of the American Medical Association, 293(19), 2384-2390.
Linkins, L. A., Choi, P. T., & Douketis, J. D. (2003). Clinical impact of bleeding in patients taking oral anticoagulant therapy for venous thromboembolism: a meta-analysis. Annals of Internal Medicine, 139(11), 893-900.
Martin, B., Filipovic, M., Rennie, L., & Shaw, D. (2010). Using machine learning to prescribe Warfarin. Proceedings of the 14th International Conference on Artificial Intelligence: Methodology, Systems, Applications, 6304, 151-160.
Mukta, P., & Usha, A. K. (2007). Neural networks and statistical techniques: A review of application. Expert Systems with Application, 36(1), 2-17.
Narayanan, M., & Lucas, S. (1993). A genetic algorithm to improve a neural network to predict a patient’s response to warfarin. Methods of Information in Medicine, 32(1), 55-58.
Palareti, G., & Cosmi, B. (2009). Bleeding with anticoagulation therapy–who is at risk, and how best to identify such patients. Thrombosis and haemostasis, 102(2), 268-278.
Palareti, G., Leali, N., & Coccheri, S. (1996). Bleeding complications of oral anticoagulant treatment: an inception-cohort, prospective collaborative study (ISCOAT). Italian Study on Complications of Oral Anticoagulant Therapy. Lancet, 348(9025), 423-428.
Pedersen, C. A., Schneider, P. J., & Scheckelhoff, D. J. (2004). ASHP national survey of pharmacy practice in hospital settings: Monitoring and patient education-2003. American Journal of Health-System Pharmacy, 61(5), 457-471.
Quinlan, J. R. (1992). Learning with continuous classes. Proceedings of the 5th Australian Joint Conference on Artificial Intelligence (pp. 343-348). Hobart , Australia.
Robins, C. H., & Robins, C. R. (1970). The eel family Dysommidae (including the Dysomminidae and Nettodaridae) its osteology and composition, including a new genus and species. Proceedings of the Academy of Natural Sciences of Philadelphia, 122(6), 293-335.
Rumelhart, D. E., Hinton, G. E., & Williams, R. J. (1986). Learning representations by back-propagating errors. Nature, 323, 533-536.
Schaufele, M. K., Marciello, M. A., & Burke, D. T. (2000). Dosing practices of physicians for anticoagulation with warfarin during inpatient rehabilitation. American Journal of Physical Medication and Rehabilitation, 79(1), 69-74.
Schulman, S. (2003). Clinical practice: care of patients receiving long-term anticoagulant therapy. New England Journal of Medicine, 349(7), 675-683.
Setiabudi, E., Alwi, I., & Setiati, S. (2008). Oral anticoagulant treatment in management of elderly patients with atrial fibrillation: is it beneficial or detrimental? Acta medica Indonesiana, 40(1), 40-47.
Shirley, J., Paine, J. D., Sarah, G., & Benator, J. D. (2003). JCAHO initiative seeks to improve patient safety. Drug Benefit Trends, 15(1), 23-24.
Simon, H., (1999). Neural Networks: A Comprehensive Foundation (2nd ed.). USA: Prentice Hall.
Solomon, I., Maharshak, N., Chechik, G., Leibovici, L., Lubetsky, A., Halkin, H., et al. (2004). Applying an artificial neural network to warfarin maintenance dose prediction. Israel Medical Association Journal. 6(12), 770-771.
Stelfox, H. T., Palmisani, S., Scurlock, C., Orav, E. J., & Bates, D. W. (2006). The “To Err is Human” report and the patient safety literature. Quality & Safety in Health Care, 15(3), 174-178.
Tadros, R., & Shakib, S. (2010). Warfarin–indications, risks and drug interactions. Australian Family Physician, 39(6), 476-479.
The Joint Commision. (2010, July 1). 2010 National patient safety goals. Retrieved August 30, 2010, from http://www.jointcommission.org
The National Coordinating Council for Medication Error Reporting and Prevention. (n.d.). About medication errors. Retrieved October 5, 2010, from http://www.nccmerp.org
Tolle, K. M., Chen, H., & Chow, H. (2000). Estimating drug/plasma concentration levels by applying neural networks to pharmacokinetic data sets. Decision Support Systems, 30(2), 139-152.
Vapnik, V. (1995). The Nature of Statistical Learning Theory. New York: Springer-Verlag.
Visser, L. E., Penning-van Bees, F. J., & Kasbergen, A. A. (2002). Overanticoagulation associated with combined use of antibacterial drugs and acenocoumarol or phenprocoumon anticoagulants. Thrombosis and haemostasis, 88(5), 705-710.
Wang , Y., & Witten, I. H. (1997). Induction of model trees for predicting continuous classes. Proceedings of the European conference on machine learning (pp. 128-137). Prague, Czech Republic.
White, H. D., Gruber, M., & Feyzi, J. (2007). Comparison of outcomes among patients randomized to warfarin therapy according to anticoagulant control– results from SPORTIF III and IV. Archives of Internal Medicine, 167(3), 239-245.
You, J. H., Chan, F. W., Wong, R. S., & Cheng, G. (2005). Is INR between 2.0 and 3.0 the optimal level for Chinese patients on warfarin therapy for moderate-intensity anticoagulation? British Journal of Clinical Pharmacology, 59(5), 582-587.

中文部分
葉怡成(民92)。類神經網路模式應用與實作。臺北市：儒林。
楊瑛碧，鄭淑文(民99)。處方疑義與用藥安全之探討。藥學雜誌，26(1)，81-87。