臺灣博碩士論文加值系統

研究目的
藥物不良反應的發生會導致不必要的住院、發病率和死亡率。因此藥物不良反應的早期偵測和準確預測對於藥物開發和患者安全至關重要，本研究擬藉由資料探勘技術及機器學習的方法，建立藥物不良反應預測模型。
研究方法
本研究為回溯性研究設計。收集南部某醫療體系之醫學中心及兩家區域醫院五年藥物不良反應通報，共5042筆資料進行分析。方法中資料探勘使用分群分析，機器學習方式則運用人工類神經網路，採取監督式學習方法建立器官、不良反應症狀及嚴重程度之預測模型。
研究結果
根據通料探勘由分群分析結果得知，可以分為六大主要器官。機器學習進一步以這六報表單可以得知藥物通報前三名為Iopromide、Levofloxacin、Vancomycin。藥物不良反應症狀通報前三名為紅疹(Rash)、皮膚炎(Dermatitis)、癢疹(Pruritus)。在資大群分類資料去進行分析器官、不良反應症狀及嚴重程度並分析預測模型之準確度。
結論與建議
研究結果顯示，使用資料探勘方式可以有效找到分群的規則，可針對特別顯著的六大器官族群進行預測並建立模型。在預測方面，可以得知使用ATC code和ICD code預測不良反應症狀，其有三大器官類別預測準確度可達0.7-0.8。而在ATC code和ICD code預測嚴重程度則有三大器官類別準確度可高達0.7-0.8。本研究結果發現，的確可以透過資料探勘方式及機器預測模型作為一個主動的藥物警訊偵測方式，也可從研究中建立早期預測藥物不良反應的基礎模型，提供未來在臨床實務上藥物不良反應管理之參考。

Purposes:
Poor management and monitoring of adverse drug reactions can lead to unnecessary hospitalization, morbidity and mortality. Early detection and accurate prediction of adverse drug reactions is importance for drug quality and patient safety. This research purposes are using data mining and machine learning methods to establish predictive models for adverse drug reactions.
Methods:
This research is a retrospective study. We collected records of adverse drug reactions in hospitals which contains a medical center and two regional hospitals in the past five years. We totally collected 5042 records. In this study ,we use cluster analysis for data mining. And we use artificial neural networks for machine learning, it is a supervised learning method to build predictive models of adverse organs, adverse symptoms and severity of adverse reactions.
Results:
According to the records, top three drugs are Iopromide, Levofloxacin, and Vancomycin. The top three symptoms of adverse drug reactions were Rash, Dermatitis, and Pruritus. In the process of data mining, we known from the results of cluster analysis that can be divided into six major organs. Machine learning uses these six clusters to analyze the accuracy of the predictive models of adverse organs, adverse symptoms and severity of adverse reactions.
Conclusions and Suggestions:
The result shows that using data mining can help us find the rules for grouping, and machine learning can predict and build models for the six particularly significant organ groups. In terms of prediction, ATC code and ICD code are used to predict adverse reaction symptoms, and the prediction accuracy of three organ categories can reach 0.7-0.8. In the severity of adverse reactions, the accuracy of the three major organ categories can be 0.7-0.8. Data mining and machine prediction models can be used as active pharmacovigilance methods, and being basic models for early prediction of adverse drug reactions.

第一章 緒論 12
第一節 研究背景與動機 13
第二節 研究目的 18
第三節 研究流程 19
第四節 名詞解釋及縮寫簡稱對照表 20
第二章 文獻探討 23
第一節：藥物不良反應(Adverse Drug Reaction) 23
第二節：資料探勘 45
第三節：機器學習法 54
第四節：總結 64
第三章 研究方法 66
第一節 研究架構 66
第二節 研究流程及研究樣本 67
第三節 研究工具 69
第四節 研究變項 73
第五節 研究分析 75
第四章-研究結果 78
第一節 描述性統計 78
第二節、使用資料探勘方式：分群圖(Clustergram) 97
第三節、使用資料探勘方式：分群分析(Cluster analysis) 99
第四節、探討ANN預測模型之準確性分析 107
第五章-研究討論與結論 134
第一節、通報配對組數分析討論及分群圖 134
第二節、使用資料探勘方式：分群分析 136
第三節、探討ANN預測模型之準確性分析 137
第四節、研究限制 149
第五節、結論與未來研究發展 150
參考文獻 151
附錄 156

Alayed, N., Alkhalifah, B., Alharbi, M., Alwohaibi, N., & Farooqui, M. (2019). Adverse Drug Reaction (ADR) as a Cause of Hospitalization at a Government Hospital in Saudi Arabia: A Prospective Observational Study. Curr Drug Saf, 14(3), 192-198. doi:10.2174/1574886314666190520105330
Aldrich, C. (2020). Process Variable Importance Analysis by Use of Random Forests in a Shapley Regression Framework. Minerals, 10(5). doi:10.3390/min10050420
Badillo, S., Banfai, B., Birzele, F., Davydov, II, Hutchinson, L., Kam-Thong, T., . . . Zhang, J. D. (2020). An Introduction to Machine Learning. Clin Pharmacol Ther, 107(4), 871-885. doi:10.1002/cpt.1796
Bailey, C., Peddie, D., Wickham, M. E., Badke, K., Small, S. S., Doyle-Waters, M. M., . . . Hohl, C. M. (2016). Adverse drug event reporting systems: a systematic review. Br J Clin Pharmacol, 82(1), 17-29. doi:10.1111/bcp.12944
Behera, S. K., Das, S., Xavier, A. S., Velupula, S., & Sandhiya, S. (2018). Comparison of different methods for causality assessment of adverse drug reactions. Int J Clin Pharm, 40(4), 903-910. doi:10.1007/s11096-018-0694-9
C A Naranjo, U. B., E M Sellers, P Sandor, I Ruiz, E A Roberts, E Janecek, C Domecq, D J Greenblatt. (1981). A method for estimating the probability of adverse drug reactions.
C. Lee Ventola, M. (2018). Big Data and Pharmacovigilance:Data Mining for Adverse Drug Events and Interactions. P&T®, Vol. 43 No. 6.
Center, T. N. A. R. (2020). Taiwan National Adverse Drug Reaction Notification System Setup. Retrieved from https://adr.fda.gov.tw/
Charles H. Chesnut, I., MD, Stuart Silverman, MD, Kim Andriano, PhD, Harry Genant, MD, Alberto Gimona, MD, Steven Harris, MD, Douglas Kiel, MD, Meryl LeBoff, MD, Michael Maricic, MD, Paul Miller, MD, Caje Moniz, MD, Munro Peacock, MD. (2000). A randomized trial of nasal spray salmon calcitonin in postmenopausal women with established osteoporosis the prevent recurrence of osteoporotic fractures study. The American Journal of Medicine, Volume 109(Issue 4), Pages 267-276.
Christopher McMaster, D. L., Claire Keith,Parnaz Aminian,Albert Frauman. (2019). A Machine-Learning Algorithm to Optimise Automated Adverse Drug Reaction Detection from Clinical Coding. Drug Saf, 42(6), 721-725. doi:10.1007/s40264-018-00794-y
Chun Yen Lee, Y.-P. P. C. (2019). Machine learning on adverse drug reactions for pharmacovigilance. Drug Discov Today, 24(7), 1332-1343. doi:10.1016/j.drudis.2019.03.003
Dathe, K., Hultzsch, S., Pritchard, L. W., & Schaefer, C. (2019). Risk estimation of fetal adverse effects after short-term second trimester exposure to non-steroidal anti-inflammatory drugs: a literature review. Eur J Clin Pharmacol, 75(10), 1347-1353. doi:10.1007/s00228-019-02712-2
Doherty, M. J. (2009). Algorithms for assessing the probability of an Adverse Drug Reaction. Respiratory Medicine CME, 2(2), 63-67. doi:10.1016/j.rmedc.2009.01.004
Edwards, I. R., & Aronson, J. K. (2000). Adverse drug reactions: definitions, diagnosis, and management. The Lancet, 356(9237), 1255-1259. doi:10.1016/s0140-6736(00)02799-9
EMA. (2012). European Medicines Agency recommends limiting longterm use of calcitonin medicines. Retrieved from
EMH. (2020a). Erythromycin: update on known risk of infantile hypertrophic pyloric stenosis. Retrieved from https://www.gov.uk/drug-safety-update/erythromycin-update-on-known-risk-of-infantile-hypertrophic-pyloric-stenosis
EMH. (2020b). European Erythromycin: caution required due to cardiac risks (QT interval prolongation); drug interaction with rivaroxaban. Retrieved from https://www.gov.uk/drug-safety-update/erythromycin-caution-required-due-to-cardiac-risks-qt-interval-prolongation-drug-interaction-with-rivaroxaban
FDA. (2013). FDA Panel: Long-Used Osteoporosis Drug Too Risky. Retrieved from https://www.webmd.com/osteoporosis/news/20130307/fda-panel-osteoporosis-drug-too-risky
FDA. (2020). FDA Warns that Using a Type of Pain and Fever Medication in Second Half of Pregnancy Could Lead to Complications. Retrieved from https://www.fda.gov/news-events/press-announcements/fda-warns-using-type-pain-and-fever-medication-second-half-pregnancy-could-lead-complications
Garbade, D. M. J. (2018). Understanding K-means Clustering in Machine Learning. Retrieved from https://towardsdatascience.com/understanding-k-means-clustering-in-machine-learning-6a6e67336aa1
Harpaz, R., DuMouchel, W., Shah, N. H., Madigan, D., Ryan, P., & Friedman, C. (2012). Novel data-mining methodologies for adverse drug event discovery and analysis. Clin Pharmacol Ther, 91(6), 1010-1021. doi:10.1038/clpt.2012.50
Iasella, C. J., Johnson, H. J., & Dunn, M. A. (2017). Adverse Drug Reactions: Type A (Intrinsic) or Type B (Idiosyncratic). Clin Liver Dis, 21(1), 73-87. doi:10.1016/j.cld.2016.08.005
ICH. (1995). Clinical Safety Data Management:Definitions and Standards for Expedited Reporting.
Imai, S., Takekuma, Y., Kashiwagi, H., Miyai, T., Kobayashi, M., Iseki, K., & Sugawara, M. (2020). Validation of the usefulness of artificial neural networks for risk prediction of adverse drug reactions used for individual patients in clinical practice. PLoS One, 15(7), e0236789. doi:10.1371/journal.pone.0236789
Jason Lazarou, M. B. H. P., MD, PhD; Paul N. Corey, PhD. (1998). Incidence of Adverse Drug Reactions in Hospitalized Patients A Meta-analysis of Prospective Studies. JAMA, Vol 279, No. 15.
Jeong, E., Park, N., Choi, Y., Park, R. W., & Yoon, D. (2018). Machine learning model combining features from algorithms with different analytical methodologies to detect laboratory-event-related adverse drug reaction signals. PLoS One, 13(11), e0207749. doi:10.1371/journal.pone.0207749
Jiawei Han, M. K., Jian Pei. (2011). Data Mining Concepts and Techniques (3rd ed.).
Jonathan R. Nebeker, M., MD; Paul Barach, MD, MPH; and Matthew H. Samore, MD. (2004). Clarifying adverse drug events a clinician's guide to terminology documentation and reporting. American College of Physicians, 795-801.
Kriegeskorte, N., & Golan, T. (2019). Neural network models and deep learning. Curr Biol, 29(7), R231-R236. doi:10.1016/j.cub.2019.02.034
Liao, P. J., Mao, C. T., Chen, T. L., Deng, S. T., & Hsu, K. H. (2019). Factors associated with adverse drug reaction occurrence and prognosis, and their economic impacts in older inpatients in Taiwan: a nested case-control study. BMJ Open, 9(5), e026771. doi:10.1136/bmjopen-2018-026771
Liu, X., Ma, J., Huang, L., Zhu, W., Yuan, P., Wan, R., & Hong, K. (2017). Fluoroquinolones increase the risk of serious arrhythmias: A systematic review and meta-analysis. Medicine (Baltimore), 96(44), e8273. doi:10.1097/MD.0000000000008273
MedDRA. (2020a). MedDRA Introductory Guide Retrieved from
MedDRA. (2020b). Welcome to the ICH MedDRA website. Retrieved from https://www.meddra.org/how-to-use/support-documentation/english/welcome
Merel, S. E., & Paauw, D. S. (2017). Common Drug Side Effects and Drug-Drug Interactions in Elderly Adults in Primary Care. J Am Geriatr Soc, 65(7), 1578-1585. doi:10.1111/jgs.14870
Murchison, L., De Coppi, P., & Eaton, S. (2016). Post-natal erythromycin exposure and risk of infantile hypertrophic pyloric stenosis: a systematic review and meta-analysis. Pediatr Surg Int, 32(12), 1147-1152. doi:10.1007/s00383-016-3971-5
Onakpoya, I. J., Heneghan, C. J., & Aronson, J. K. (2016). Post-marketing withdrawal of 462 medicinal products because of adverse drug reactions: a systematic review of the world literature. BMC Med, 14, 10. doi:10.1186/s12916-016-0553-2
Overman, R. A., Borse, M., & Gourlay, M. L. (2013). Salmon calcitonin use and associated cancer risk. Ann Pharmacother, 47(12), 1675-1684. doi:10.1177/1060028013509233
Patton, K., & Borshoff, D. C. (2018). Adverse drug reactions. Anaesthesia, 73 Suppl 1, 76-84. doi:10.1111/anae.14143
Pham, M., Cheng, F., & Ramachandran, K. (2019). A Comparison Study of Algorithms to Detect Drug-Adverse Event Associations: Frequentist, Bayesian, and Machine-Learning Approaches. Drug Saf, 42(6), 743-750. doi:10.1007/s40264-018-00792-0
Ramesh, A. N., Kambhampati, C., Monson, J. R., & Drew, P. J. (2004). Artificial intelligence in medicine. Ann R Coll Surg Engl, 86(5), 334-338. doi:10.1308/147870804290
Robert S. Bresalier, M. D., Robert S. Sandler, M.D., Hui Quan, Ph.D. (2005). Cardiovascular Events Associated with Rofecoxib in a Colorectal Adenoma Chemoprevention Trial. The new england journal of medicine, 1092-1102.
Sakaeda, T., Tamon, A., Kadoyama, K., & Okuno, Y. (2013). Data mining of the public version of the FDA Adverse Event Reporting System. Int J Med Sci, 10(7), 796-803. doi:10.7150/ijms.6048
Singh, P., Agrawal, M., Hishikar, R., Joshi, U., Maheshwari, B., & Halwai, A. (2017). Adverse drug reactions at adverse drug reaction monitoring center in Raipur: Analysis of spontaneous reports during 1 year. Indian J Pharmacol, 49(6), 432-437. doi:10.4103/ijp.IJP_781_16
Smith, W. (2013). Adverse drug reactions(allergy-side effect-intolerance. AustRAliAn FAmily PhysiciAn, Vol. 42.
Sun, W., Cai, Z., Li, Y., Liu, F., Fang, S., & Wang, G. (2018). Data Processing and Text Mining Technologies on Electronic Medical Records: A Review. J Healthc Eng, 2018, 4302425. doi:10.1155/2018/4302425
TaiwanFDA. (2014). Taiwan National Adverse Drug Reactions Reporting System. Retrieved from https://www.fda.gov.tw/tc/siteContent.aspx?sid=4240
TaiwanFDA. (2020). Statistics of adverse drug notifications. Retrieved from https://www.tdrf.org.tw/apply04/
Tan, Y., Hu, Y., Liu, X., Yin, Z., Chen, X. W., & Liu, M. (2016). Improving drug safety: From adverse drug reaction knowledge discovery to clinical implementation. Methods, 110, 14-25. doi:10.1016/j.ymeth.2016.07.023
Tiftikci, M., Ozgur, A., He, Y., & Hur, J. (2019). Machine learning-based identification and rule-based normalization of adverse drug reactions in drug labels. BMC Bioinformatics, 20(Suppl 21), 707. doi:10.1186/s12859-019-3195-5
Tripathi, M. (2020). Underfitting and Overfitting in Machine Learning. Data Science Foundation.
Usama Fayyad, G. P.-S., and Padhraic Smyth. (1996). From Data Mining to Knowledge Discovery in Databases. American Association for Artificial Intelligence, 17.
W. Philip T. James, M. D., D.Sc., Ian D. Caterson, M.D., Ph.D. (2010). Effect of Sibutramine on Cardiovascular Outcomes in Overweight and Obese Subjects. The new england journal of established in 1812 medicine, vol. 363 no. 10, 905-917.
Walter, S. R., Day, R. O., Gallego, B., & Westbrook, J. I. (2017). The impact of serious adverse drug reactions: a population-based study of a decade of hospital admissions in New South Wales, Australia. Br J Clin Pharmacol, 83(2), 416-426. doi:10.1111/bcp.13124
WHO. (2002). Safety of medicines: a guide to detecting and reporting adverse drug reactions-why health professionals need to take action.pdf>.
WHO. (2016). ATC/DDD Index 2016. Retrieved from https://www.whocc.no/atc_ddd_index/
WHO. (2018). The use of the WHO-UMC system
for standardised case causality assessment.
Wilke, R. A., Lin, D. W., Roden, D. M., Watkins, P. B., Flockhart, D., Zineh, I., . . . Krauss, R. M. (2007). Identifying genetic risk factors for serious adverse drug reactions: current progress and challenges. Nat Rev Drug Discov, 6(11), 904-916. doi:10.1038/nrd2423
Wu, L., Ingle, T., Liu, Z., Zhao-Wong, A., Harris, S., Thakkar, S., . . . Fang, H. (2019). Study of serious adverse drug reactions using FDA-approved drug labeling and MedDRA. BMC Bioinformatics, 20(Suppl 2), 97. doi:10.1186/s12859-019-2628-5
Yao, S. H., Tsai, H. T., Lin, W. L., Chen, Y. C., Chou, C., & Lin, H. W. (2019). Predicting the serum digoxin concentrations of infants in the neonatal intensive care unit through an artificial neural network. BMC Pediatr, 19(1), 517. doi:10.1186/s12887-019-1895-7
Zhou, L., & Rupa, A. P. (2018). Categorization and association analysis of risk factors for adverse drug events. Eur J Clin Pharmacol, 74(4), 389-404. doi:10.1007/s00228-017-2373-5