臺灣博碩士論文加值系統

史蒂芬強生症候群，為一種由藥物引起導致皮膚發生廣泛性脫落、壞死及黏膜糜爛等不良反應，並有造成生命危險的可能。由於藥物是史蒂芬強生症候群(SJS)的一項重要起因，首要與最重要治療史蒂芬強生症候群(SJS)的步驟是停止任何可能導致它的藥物。但也有廣泛分佈於各種不同的醫療領域的許多潛在的藥物可能會導致史蒂芬強生症候群(SJS)，而且只有極少數的醫師熟悉這些藥物。有一個例子是在醫師未察覺相關藥物會引起SJS的情形下，醫師未能發現必須停止使用這些藥物，並給予病人抗生素治療，使得SJS病人的病情越來越嚴重。這種狀況使得患者重新住院治療 SJS。
目前在台灣，尚未出現針對SJS患者再次住院的案件的相關研究。最近的一項研究激勵，成功地定義六種醫師常用來治療SJS病人的處方(用藥)行為，這項研究是從台灣國民健康保險資料庫取出的相同SJS案例資料，提出一個資料探勘的方法來調查藥物和用藥行為之間可能會影響SJS患者發生再次住院的關連性。
這項研究將首先以規則為基礎的分類方法來分類再次住院的患者的處方歷史。藥物的關連行為的確立是藉由使用藥物的行為改變的基本特性。然後，藥物配對將利用A-priori演算法來獲得，而其關聯度將藉由選定的客觀衡量來取得。最後，基於關聯度來列出與排序藥物配對，將有益於醫師開立處方使得減少SJS患者復發與再次住院。

Complications of drug therapy are the most common type of adverse events in hospitalized patients. The example includes Stevens-Johnson syndrome (SJS), a life-threatening skin reactions to medications. Since drugs are the important causes of SJS, the first and most important step in treating SJS is to discontinue any medication that may be causing it. But the potential drugs that may lead SJS spread extensively in various medical fields and very few physicians are familiar with all these drugs. There is a case when physician is not aware about the drugs causing SJS, the physician failed to identify the “must be discontinued drugs” and treat the patient with more antibiotics which make the SJS patient’s condition worse. This condition makes the patient re-hospitalized for treating SJS.
Currently in Taiwan, the emergence of re-hospitalization case in SJS patients has not yet known. Motivating by recent study that successfully defined six kinds of prescription behavior that commonly acted by physicians in treating SJS patient, this study, by taking the same SJS case data from Taiwan’s National Health Insurance database, proposes a data mining approach to investigate association pattern between drugs within these prescription behavior that may affect the occurrence of re-hospitalization case in SJS.
This study will first classify prescription history of re-hospitalized patients through rule-based classification method. Secondly, using the basic properties of prescription behavior identify drug’s association pattern. Then, pair of drugs will be generated by employing A-priori algorithm and its degree of association will be obtained by using selected objective measure. Finally, by listing and ranking up this pair of drugs based on its degree of association, it can assist physician in prescribing drugs to increase the awareness and reduce recurrence and re-hospitalization case in SJS patients.

中文摘要 ii
Abstract iii
Acknowledgements iv
Table of Contents v
List of Figures viii
List of Tables ix
Chapter 1 Introduction 1
1.1 Background 1
1.2 Motivation 2
1.3 Objective 3
1.4 Organization of Thesis 4
Chapter 2 Literature Review 5
2.1 Data Mining 6
2.2 Rule-Based Classification Algorithm 8
2.3 Association Analysis 9
2.4 Alternative Objective of Interesting Measure 10
Chapter 3 Research Methodology 13
3.1 Concept Phase 13
3.2 Design Phase 14
3.2.1 Data Pre-processing 14
3.2.2 Classification of prescription behavior 14
3.2.3 Generating drugs pair candidates 18
3.2.4 Calculating degree of association between drugs 21
3.2.5 Terminology Summary 23
Chapter 4 Model Implementation 24
4.1 Data pre-processing 25
4.1.1 Obtaining prescription record 25
4.1.2 Drug and its compositions code 26
4.2 Classifying patient’s prescription behavior 28
4.3 Discovering drug association 29
4.3.1 Generating drugs pair 29
4.3.2 Measuring degree of association 31
4.4 Result Analysis and Discussion 38
4.4.1 Result Analysis 38
4.4.2 Discussion 41
Chapter 5 Conclusion and Future Research 43
5.1 Conclusion 43
5.2 Future Research 44
Reference 46
Appendix A-1. Classification result for all prescription records of re-hospitalized patients 49
Appendix B-1. Calculation result of association between drugs under {f10} ∧ {f01} behavior using asymmetric measures 53
Appendix B-2. Calculation result of association between drugs under {f11} ∧ {f10} behavior using asymmetric measures 54
Appendix B-3. Calculation result of association between drugs under {f11} ∧ {f01} behavior using asymmetric measures 55
Appendix C-1. Calculation result of association between drugs under {f11} ∧ {f10} behavior using six objective measures 56
Appendix C-2. Calculation result of association between drugs under {f11} ∧ {f01} behavior using six objective measures 56
Appendix D-1. Ranking result of drugs pair under {f11} ∧ {f10} association behavior using six objective measures 57
Appendix D-2. Ranking result of drugs pair under {f11} ∧ {f01} association behavior using six objective measures 57
Appendix E-1. A composite ranking of drugs pair under {f11} ∧ {f10} behavior obtained from sums of six objective measure ranks 58
Appendix E-2. A composite ranking of drugs pair under {f11} ∧ {f01} behavior obtained from sums of six objective measure ranks 59
Appendix F-1. Ranking of drugs pair under {f11} ∧ {f10} association behavior using six objective measures 60
Appendix F-2. Ranking of drugs pair under {f11} ∧ {f01} association behavior using six objective measures 61

[1]Wolf, R., Orion, E., Marcos, B., and Matz, H., "Life-threatening acute adverse cutaneous drug reactions," Clinics in Dermatology, Vol. 23, pp. 171-181, 2005.
[2]WHO, "Medicines: safety of medicines - adverse drug reactions," No.293, 2008.
[3]Harr, T. and French, L. E. (2010, 07 December 2011). Toxic epidermal necrolysis and Stevens-Johnson syndrome. 5. Available: http://www.ojrd.com/content/5/1/39
[4]Ghislain, P.-D. and Roujeau, J.-C., "Treatment of severe drug reactions: Stevens-Johnson Syndrome, Toxic Epidermal Necrolysis and Hypersensitivity syndrome," Dermatology Online Journal, Vol. 8, 2002.
[5]Roujeau, J.-C., Kelly, J. P., Naldi, L., Rzany, B., Stern, R. S., Anderson, T., Auquier, A., Bastuji-Garin, S., Correia, O., Locati, F., Mockenhaupt, M., Paoletti, C., Shapiro, S., Shear, N., Schopf, E., and Kaufman, D. W., "Medication Use and the Risk of Stevens–Johnson Syndrome or Toxic Epidermal Necrolysis," New England Journal of Medicine, Vol. 333, pp. 1600-1608, 1995.
[6]劉宜佩, "發展一探勘模式於醫師用藥行為探討-以史蒂芬氏強生症候群為例," Master, 工業管理系, 國立台灣科技大學, 2011.
[7]Garcia-Doval, I., LeCleach, L., Bocquet, H., Otero, X. L., and Roujeau, J. C., "Toxic epidermal necrolysis and Stevens-Johnson syndrome: does early withdrawal of causative drugs decrease the risk of death?," Archives of Dermatology, Vol. 136, p. 323, 2000.
[8]Canlas, R. D., "Data Mining in Healthcare: Current Applications and Issues," School of Information Systems & Management, Carnegie Mellon University, Australia 2009.
[9]Shi, Y. W., Min, F. L., Liu, X. R., Zan, L. X., Gao, M. M., Yu, M. J., and Liao, W. P., "Hla-B alleles and lamotrigine-induced cutaneous adverse drug reactions in the Han Chinese population," Basic Clin Pharmacol Toxicol., Vol. 109, pp. 42-6. doi: 10.1111/j.1742-7843.2011.00681.x. Epub 2011 Mar 16., 2011.
[10]Wang, Q., Zhou, J. Q., Zhou, L. M., Chen, Z. Y., Fang, Z. Y., Chen, S. D., Yang, L. B., Cai, X. D., Dai, Q. L., Hong, H., and Wang, H. X., "Association between HLA-B*1502 allele and carbamazepine-induced severe cutaneous adverse reactions in Han people of southern China mainland," Seizure., Vol. 20, pp. 446-8. Epub 2011 Mar 11., 2011.
[11]Wu, X. T., Hu, F. Y., An, D. M., Yan, B., Jiang, X., Kwan, P., Stefan, H., and Zhou, D., "Association between carbamazepine-induced cutaneous adverse drug reactions and the HLA-B*1502 allele among patients in central China," Epilepsy Behav., Vol. 19, pp. 405-8. Epub 2010 Sep 15., 2010.
[12]Han, J., Kamber, M., and Pei, J., Data mining: concepts and techniques: Morgan Kaufmann, 2011.
[13]Hand, D., Mannila, H., Smyth, P, Principles of data mining: MIT, 2001.
[14]Tan, P.-N., Steinbach, M., and Kumar, V., Introduction to Data Mining, Pearson International ed.: Pearson Addison Wesley, 2006.
[15]Koh, H. C. and Tan, G., "Data mining applications in healthcare," Journal of Healthcare Information Management, Vol. 19, p. 65, 2011.
[16]Yoo, I., Alafaireet, P., Marinov, M., Pena-Hernandez, K., Gopidi, R., Chang, J.-F., and Hua, L., "Data Mining in Healthcare and Biomedicine: A Survey of the Literature," Journal of Medical Systems, pp. 1-18, 2011.
[17]Golub, T. R., Slonim, D. K., Tamayo, P., Huard, C., Gaasenbeek, M., Mesirov, J. P., Coller, H., Loh, M. L., Downing, J. R., and Caligiuri, M. A., "Molecular classification of cancer: class discovery and class prediction by gene expression monitoring," science, Vol. 286, p. 531, 1999.
[18]Sanchez, M. A., Uremovich, S., and Acrogliano, P., "Mining Tuberculosis Data," Data Mining and Medical Knowledge Management: Cases and Applications, Medical Information Science Reference, New York, 2009.
[19]Mukhopadhyay, A., Maulik, U., and Bandyopadhyay, S., "Refining Genetic Algorithm Based Fuzzy Clustering through Supervised Learning for Unsupervised Cancer Classification," in Evolutionary Computation, Machine Learning and Data Mining in Bioinformatics. Vol. 5483, Pizzuti, C., Ritchie, M., and Giacobini, M., Eds., ed: Springer Berlin / Heidelberg, 2009, pp. 191-202.
[20]Becquet, C., Blachon, S., Jeudy, B., Boulicaut, J.-F., and Gandrillon, O., "Strong-association-rule mining for large-scale gene-expression data analysis: a case study on human SAGE data," Genome Biology, Vol. 3, pp. research0067.1 - research0067.16, 2002.
[21]Kaur, H. and Wasan, S. K., "Empirical study on applications of data mining techniques in healthcare," Journal of Computer Science, Vol. 2, pp. 194-200, 2006.
[22]Kahramanli, H. and Allahverdi, N., "Mining Classification Rules for Liver Disorders," International Journal of Mathematics and Computers in Simulation, Vol. 3, 2009.
[23]Aramaki, E., Imai, T., Miyo, K., and Ohe, K., "Patient Status Classification by using Rule based Sentence Extraction and BM25-kNN based Classifier," The University of Tokyo Hospital, Japan2006.
[24]Nguyen, A. N., Lawley, M. J., Hansen, D. P., Bowman, R. V., Clarke, B. E., Duhig, E. E., and Colquist, S., "Symbolic rule-based classification of lung cancer stages from free-text pathology reports," Journal of the American Medical Informatics Association, Vol. 17, pp. 440-445, 2010.
[25]Chen, T.-J., Chou, L.-F., and Hwang, S.-J., "Application of a data-mining technique to analyze coprescription patterns for antacids in Taiwan," Clinical Therapeutics, Vol. 25, pp. 2453-2463, 2003.
[26]Tai, Y. M. and Chiu, H. W., "Comorbidity study of ADHD: applying association rule mining (ARM) to National Health Insurance Database of Taiwan," Int J Med Inform., Vol. 78, pp. e75-83. Epub 2009 Oct 22., 2009.
[27]Agrawal, R. and Srikant, R., "Fast algorithms for mining association rules," in Proceedings of the 20th VLDB conference, 1994, pp. 487-499.
[28]Li, J., Fu, A. W.-c., He, H., Chen, J., Jin, H., McAullay, D., Williams, G., Sparks, R., and Kelman, C., "Mining risk patterns in medical data," presented at the Proceedings of the eleventh ACM SIGKDD international conference on Knowledge discovery in data mining, Chicago, Illinois, USA, 2005.
[29]Tan, P. N., Kumar, V., and Srivastava, J., "Selecting the right objective measure for association analysis," Information Systems, Vol. 29, pp. 293-313, 2004.
[30]Ohsaki, M., Kitaguchi, S., Okamoto, K., Yokoi, H., and Yamaguchi, T., "Evaluation of Rule Interestingness Measures with a Clinical Dataset on Hepatitis," in Knowledge Discovery in Databases: PKDD 2004. Vol. 3202, Boulicaut, J.-F., Esposito, F., Giannotti, F., and Pedreschi, D., Eds., ed: Springer Berlin / Heidelberg, 2004, pp. 362-373.
[31]Ngan, V. and Oakley, D. A. (2003, Stevens Johnson Syndrome & Toxic Epidermal Necrolysis. Available: http://dermnetnz.org/reactions/sjs-ten.html
[32]Hamilton, H. J., "Itemset/Probability Based Interestingness Measures Generator," ed, 2002.