臺灣博碩士論文加值系統

台灣投入DRG研究及準備實施，已歷10餘年，共試行「論病例計酬」50餘項，並發佈了「台灣版DRG分類」。一旦實施了DRG支付制度，所有住院病患將於出院時，才得以知道病患此次住院的DRG，亦即在出院時才得知病患此次住院可申報的醫療費用；如此一來造成醫院於病患照護過程中，無法預先得知病患未來出院時的DRG，便無法根據可獲得之醫療資源來加以有效利用，如此恐會造成過多醫療資源的浪費，或是發生醫院為節省成本，減少必要的檢驗檢查、縮短病患住院天數等情形。
本研究目的是希望能在病患尚未出院時，就能利用C4.5演算法與倒傳遞類神經網路預測出闌尾切除術病患出院時的DRG。病患的ICD-9、合併症或併發症都是在病患出院後才能得知，因此我們使用病患的檢驗檢查資料來作預測。本研究的資料來源為兩家區域級醫院，研究結果顯示，利用C4.5演算法所建構的決策樹及倒傳遞類神經網路所建構出的預測模式，準確率分別可達97.84%及98.70%。
根據預測出的病患DRG碼，醫院便可預先瞭解闌尾切除術病患此次住院可向健保局申報的費用，使醫院更能有效運用醫院資源，作最適度的安排。對病患而言，可提升病患就醫之醫療品質，如此將可創造醫院與病患雙贏的局面。

The Bureau of National Health Insurance (NHI) in Taiwan has endeavored for decades in the preparation of carrying out DRGs payment system, bringing case-payment system with more than 50 items into action and issuing three versions of DRG classification, TDRG-I、TDRG-II and TDRG-III. Once NHI puts DRG payment system into practice, the hospitals will face a serious condition that they will not know the reimbursement of inpatients until the patients discharge. Since of the uncertainty of payment, such a condition will cause the hospitals to over-utilize medical resources or to reduce the necessary exams or the length of stay for saving costs.
In this thesis, we predict the DRG of an inpatient with appendectomy through the models constructed by C4.5 algorithm and back-propagation neural network (BPN) before they discharge. Instead of the information, such as ICD-9, complication, and comorbidity, which can be got only when the inpatient discharges, we utilize the tests of an inpatient which can be got while he gets the admission note. After implementing the above two models, we conduct the experiments with the data from two hospitals. The results showed that the accuracies of the two classification models constructed by C4.5 and BPN can be up to 97.84% and 98.70%, respectively.
The results imply that the hospital can estimate the reimbursement of appendectomy patients accurately. According to the predicted DRG code, the hospitals can know how many the medical expenses they can acquire. Thus, the hospitals can effectively arrange the medical resources with certainty in payment and quality. Finally, this method is expected to improve the medical quality of patients and create a win-win situation between hospitals and patients.

致謝 i
Abstract ii
Chinese Abstract iii
List of Figures v
List of Tables vi
Chapter 1 Introduction 1
1.1 Research Background and Motivation 1
1.2 Problem Statement and Research Objective 2
1.3 Research Contribution 2
Chapter 2 Literature Review 3
2.1 Diagnosis related groups (DRGs) 3
2.1.1 Basics of Diagnosis Related Groups 3
2.1.2 Clinical Definition of Diagnosis Related Groups 4
2.1.3 Related Research of Diagnosis Related Groups 6
2.1.4 Appendectomy of Diagnosis Related Groups 7
2.2 Classification Methods 9
Chapter 3 Research method and model construction 15
3.1 The Prediction Attributes of Diagnosis Related Groups 15
3.2 Model Construction of Diagnosis Related Groups 16
3.2.1 Data Source and Data Processing 16
3.2.2 Classification Model Construction 18
Chapter 4 Research Result Analysis 23
4.1 Model Evaluation Index 23
4.2 Model Analysis and Evaluation 24
4.2.1 Model Analysis for all Attributes 24
4.2.2 Model Analysis for Selection Attributes 29
4.2.3 Model Analysis for Bagging Method 32
Chapter 5 Conclusion and Research Limitation 36
5.1 Conclusion 36
5.2 Research Limitation and the Future Research Direction 37
Reference 38
Appendix 43

[1]A.C. Tan and D. Gilbert, “Ensemble machine learning on gene expression data for cancer classification,” Applied bioinformatics. Vol. 2, No. 3 Suppl, pp. s75-s83, 2003.
[2]B.A. Flores and J.A. Gonzalez, “Data Mining with Decision Trees and Neural Networks for Calcification Detection in Mammograms,” MICAI 2004: Advances in Artificial Intelligence: Third Mexican International Conference on Artificial Intelligence, Mexico City, Mexico, April 26-30, 2004. Proceedings.
[3]Bureau of National Health Insurance, www.nhi.gov.tw, Diagnosis Related Groups III for uploading form of coding service system.
[4]Bureau of National Health Insurance, www.nhi.gov.tw, Diagnosis Related Groups III for classification table.
[5]Bureau of National Health Insurance, www.nhi.gov.tw, Diagnosis Related Groups III for payment project.
[6]C. Guerra-Salcedo and D. Whitley, “Feature Selection mechanisms for ensemble creation: a genetic search perspective,” Freitas AA (Ed.) Data Mining with Evolutionary Algorithms: Research Directions–Papers from the AAAI Workshop, 13-17. Technical Report WS-99-06. AAAI Press, 1999.
[7]Cong-Jian Jian, ”Introduce to Diagnosis Related Groups/ Prospective Payment System,” Journal of Taiwan Medical, Vol. 37, No. 6, pp. 105-108, 1994.
[8]D. Milosevic, D. Batinic, P. Konjevoda, N. Blau, N. Stambuk, L. Nizic, K. Vrljicak and D. Batinic, “Analysis of calcium, oxalate, and citrate interaction in idiopathic calcium urolithiasis in children,” Journal of chemical information and computer sciences, Vol. 43, No. 6, pp. 1844-1847, 2003.
[9]D.A. Forgione, T.E. Vermeer, Krishnamurthy Surysekar, J.A. Wrieden and C.A. Plante, “The Impact of DRG-Based Payment System on Quality of Health Care in OECD Countries,” Journal of Health Care Finance, Vol. 31, No. 1, pp. 41-54, 2004.
[10]D.C. Hsia, W.M. Krushat, A.B. Fagan, J.A. Tebbutt and R.P. Kusserow, “Accuracy of diagnostic coding for Medicare patients under the prospective-payment system, ”The New England Journal of Medicine, Vol. 318, No. 6, pp. 352-355,1988.
[11]D.D. Lewis and J. Catlett, “Heterogeneous Uncertainty Sampling for Supervised Learning,” In Proceedings of the 11th International Conference on Machine Learning, pp. 148-156, 1994.
[12]D.W. Simborg, “DRG Creep,” The New England Journal of Medicine, Vol. 304, pp. 1602-1604, 1981.
[13]D.Z. Louis, E.J. Yuen, M. Braga, A. Cicchetti, C. Rabinowitz, C. Laine and J.S. Gonnella, “Impact of a DRG-based hospital financing system on quality and outcome of care in Italy,” Health Service Research, Vol. 34, No. 1, pp. 405-415, 1999.
[14]Glenn Rouse, “Sonography of Appendicitis: A Review,” Journal of Diagnostic Medical Sonography, Vol. 5, No. 2, pp. 57-60, 1989.
[15]Health and National Health Insurance Annual Statistics Information Service, http://www.doh.gov.tw/statistic/index.htm.
[16]I.H. Witten & E. Frank, Data mining：practical machine learning tools and techniques, Morgan Kaufmann, 2005.
[17]J. Han, and M. Kamber, Data mining concepts and techniques, Morgan Kaufmann, 2001.
[18]J.R. Quinlan, C4.5: programs for machine learning, Morgan Kaufmann, 1993.
[19]Jenn-Lung Su, Guo-Zhen Wu and I-Pin Chao, “The approach of data mining methods for medical database,” Engineering in Medicine and Biology Society, Proceedings of the 23rd Annual International Conference of the IEEE Vol. 4, pp. 3824-3826, 2001.
[20]Kuei Han, “Case Payment for Health Care and Preparing for Implementation of TDRG in Taiwan,” Journal of Healthcare Management, Vol. 6, No. 1, pp. 20-36, 2005.
[21]L. Breiman, “Bagging predictors,” Machine Learning, Vol. 24, No. 2, pp. 123-140, 1996.
[22]L.B. Russel, and C.L. Manning, “The effect of prospective payment on Medicare expenditure,” The New England Journal of Medicine, Vol. 320, pp. 439-444, 1989.
[23]L.O. Hall, X. Liu, K.W. Bowyer, and R. Banfield, “Why are Neural Networks Sometimes Much More Accurate than Decision Trees: An Analysis on a bio-Information Problem,” Systems, Man and Cybernetics, IEEE International Conference on Vol. 3, pp. 2851-2856, 2003.
[24]M. Beller, R. Stotzka, T.O. Muller and H. Gemmeke, “An example-based system to support the segmentation of stellate lesions,” BVM - Bildverarbeitung fur die Medizin, pp. 475-479, 2005.
[25]M. Kantardzic, Data mining: concepts, Models, Methods, and Algorithms, Wiley IEEE Computer Society, 2003.
[26]M. Quartararo, P. Glasziou and C.B. Kerr, “Classification trees for decision making in long-term care,” The journals of gerontology. Series A, Biological sciences and medical sciences, Vol. 50, No. 6, pp. 298-302, 1995.
[27]M.J. Long, J.D. Chesney and R.P. Ament, “The Effect of PPS on Hospital Product and Productivity,” Medical Care, Vol. 25, No. 6, pp. 528-538, 1987.
[28]Mei-De Li, “Development and Evaluation of the decision Support System for Diagnostic Related Groups: Case Study of Respiratory System Disease,” Thesis for Department of Management of Information System, National Chung Cheng University, Taiwan, 2005.
[29]Michael J.A. Berry and G. Linoff, Data mining techniques: for marketing, sales, and customer support, Wiley, 1997.
[30]P.S. Maclin and J. Dempsey, “How to improve a neural network for early detection of hepatic cancer,” Cancer letters, Vol. 77, pp. 95-101, 1994.
[31]R.J. Roiger and M.W. Geatz, Data mining: a tutorial-based primer, Addison Wesley, 2003.
[32]Ray-Yi Chang and Chiu-Ling Lai, “Risk adjuster: the basis for capitation payment,” Taiwan Journal of Public Health, Vol. 23, No. 2, pp. 91-99, 2004.
[33]S. Walczak, “Artificial Neural Network Medical Decision Support Tool: Predicting Transfusion Requirements of ER Patients,” IEEE TRANSACTIONS ON INFORMATION TECHNOLOGY IN BIOMEDICINE, Vol. 9, No. 3, 2005.
[34]Su-Zhen Huang, “The Application of Data Mining Technique on Hospitalization Management: Classification of Hospital Stay Length for Appendectomy Inpatients” Thesis for Department of Business Administration, Southern Taiwan University of Technology, Taiwan, 2004.
[35]UniPattern Company, http://www.ibtechnology.biz/index.html.
[36]W.E. Pofahl, S.M. Walczak, E. Rhone and S.D. Izenberg, “Use of an artificial neural network to predict length of stay in acute pancreatitis,” The American surgeon, Vol. 64, No. 9, pp. 868-872, 1998.
[37]W.L. Wu and F.C. Su, “Potential of the back propagation neural network in the assessment of gait patterns in ankle arthrodesis,” Clinical biomechanics (Bristol, Avon), Vol. 15, No. 2, pp. 143-145, 2000.
[38]Wan-Ming Chen and Shiao-Chi Wu, “Simulating the financial impact of DRGs on hospitals in Taiwan,” Taiwan Journal of Public Health, Vol. 24, No. 4, pp. 306-314, 2005.
[39]Xing-Xiang Company, http://www.these.com.tw.
[40]Yi-Cheng Ye, Neural Network Implement, Ru-Llin Publishing House: Taiwan, 2003.
[41]Yi-Wen Tsai, Yi-Chou Chuang, Weng-Foung Huang, Lai-Chu See, Chung-Lin Yang and Pei-Fen Chen, “The effect of changing reimbursement policies on quality of in-patient care, from fee-for-service to prospective payment,” International Journal for Quality in Health Care, Vol. 17, No. 5, pp. 421–426, 2005.
[42]Zhao-Ming Huang, “Diagnosis Related Groups II Planning” DRG for information system conference, 2005.
[43]Zhe-Ming Yang, “The legislative experience of Diagnosis Related Groups in US,” Taiwan Journal of Public Health, Vol. 23, No. 2, pp. 79-90, 1996.
[44]Zhi-Yi You, “Review on Prospective Payment System/ Diagnosis Related Groups in US,” Journal of Taiwan Medical, Vol. 35, No. 5, pp. 26-30, 1992.
[45]Zhong-Fu Lan, “Payment System for National Health Insurance,” The research plan of Department of Health, Executive Yuan in Taiwan, 1997.
[46]Zi-Ling Ruan and Zhong-Fu Lan, “Introduce to Diagnosis Related Groups in Taiwan,” Journal of National Health Insurance, Vol. 19, pp. 10-13, 1999.