健保資料庫儲存了自健保開辦以來全台投保者的就醫資料，2000年起健保局委託國家衛生研究院建置「全民健康保險研究資料庫」，期望利用健保資料以累積實證的基礎。但研究人員在面對健保資料庫的巨量資料時，花費非常龐大的精力在處理資料以及等待資料的篩選，而這項資料處理也往往占去研究者大半時間，成為研究健保資料庫的門檻，令許多有意想研究的研究者裹足不前。

而目前所使用的關聯式資料庫自從二十多年前發展以來，一直被視為是資料儲存的主流，但在現今面對大量的資料讀寫以及應用，傳統關聯式資料庫面臨到效能以及擴展不易的問題。因此如何使用新型態的資料模型取代部分舊有的資料儲存方式是目前須解決的問題。

根據2010年向國衛院申請健保資料庫的件數顯示，探討健保資源使用在特定族群上的醫療耗用約占了總申請件的百分之十五。由此可見醫療費用的上漲及減少醫療資源的濫用，不僅是健保欲解決的當務之急，同樣也是研究人員感興趣的主題。

關於檢視健保資源耗用方面，健保局僅止於申請金額資料的提供，而非實際有用資訊的產出，且對於特定族群及疾病之健保資源耗用缺乏一通用的分析比較平台，因此本論文的目標為建立一個容易使用之健保資料篩選入口平台讓使用者從不同維度去深入健保資料庫，並將其相關資料統計後以圖像化方式呈現資源耗用狀況，幫助研究者從中得到有意義的資料，進而做更進一步的分析。

Since 2000, Bureau of National Health Insurance of Taiwan commissioned the National Health Research Institutes (NHRI) to build up the "National Health Insurance Research Database(NHIRD)". NHIRD contains all insured medical record and provides excellent material for medical researcher. But researchers are difficult to handle the data amount. To those researchers, such huge amount of data has become a problem.

The relational database was developed at over twenty years ago; it has been seen as the mainstream of data storage technique. But relational database is not suitable for working with Big Data. Traditional relational database can’t working well in the approach with scale up. The new type of data storage models are commonly subsumed under the term NoSQL(Not-Only SQL). The key point is we may need to apply a new type of data storage model in NHIRD.

Focus on NHI Resource Utilization. Main purpose of present research is to set up a system based on non-relational database, try to provide a platform for analyzing NHI resource utilization systemically and to help researchers to derive significant information from NHIRD with data visualization tools.

誌謝 I
中文摘要 II
Abstract III
目錄 IV
圖目錄 VI
表目錄 VIII
第一章緒論 1
1.1研究背景 1
1.2研究動機及目的 1
1.3論文架構 1
第二章相關研究 3
2.1非關聯式資料庫 3
2.1.1 Key-Value Database 5
2.1.2 Column-Oriented Database 6
2.1.3 Document-Oriented 7
2.2 MapReduce 8
2.3 Data Visualization 10
2.4病例組合指標(CMI, Case Mix Index) 11
第三章材料與方法 12
3.1研究材料 12
3.2系統設計 13
3.2.1資料前處理 14
3.2.2資料儲存元件(DataStorage) 17
3.2.3計算元件(MapReduce Component) 18
3.2.4使用者介面及控制元件(User Interface/Controller) 19
第四章研究結果 21
4.1健保資源耗用概觀 21
4.2健保資源耗用排名 23
4.2.1疾病 23
4.2.2醫令 24
4.2.3醫療院所 24
4.3疾病資源耗用呈現 26
4.4 Timeline 28
4.4.1患者 28
4.4.2醫師 29
4.5疾病族群耗用比較 31
4.6病歷組合指標計算 32
4.7資料預篩匯出 33
4.8平台展示與討論 34
第五章結論與未來展望 37
5.1結論 37
5.2未來展望 38
參考文獻 39

[1] National Health Research Institutes:[Internet], Available fromhttp://nhird.nhri.org.tw/file_talk/apply101.xls(2012)
[2]連賢明,“如何使用健保資料進行經濟研究”經濟論文叢刊 36:1, pp115-143(2008)
[3] E. F. Codd,A Relational Model of Data for Large Shared Data Banks. Commun. ACM 13(6): 377-387(1970)
[4]M. Stonebraker, “Sql databases v. nosql databases,” Communications of the ACM, vol. 53, no. 4, pp.10–11, 2010.
[5]F Chang,Bigtable: A Distributed Storage System for Structured Data(2006)
[6]G DeCandia, Dynamo: Amazon’s Highly Available Key-value Store(2007)
[7]W. Vogels, “Eventually consistent,” Communications of the ACM, vol. 52, no. 1, pp. 40–44, 2009.
[8]J. Han, E. Haihong, G. Le, and J. Du, “Survey on nosql database,” in Pervasive Computing andApplications (ICPCA), 2011 6th International Conference on. IEEE, 2011, pp. 363–366.
[9]C. Strauch, U. Sites, and W. Kriha, “Nosql databases,” Lecture Notes, Stuttgart Media University,2011.
[10]RishabhAgrawal:[Internet], Available fromhttp://nosql.rishabhagrawal.com/2012_07_01_archive.html
[11]Wei-ping Z. and Ming-xin L. “Using MongoDB to Implement Textbook Management System instead of MySQL”, 3rd IEEE International Conference on Communication Software and Networks (ICCSN), May 27-29, 2011, Xi’an, China.
[12]Jeffrey Dean and Sanjay Ghemawat,MapReduce: Simplified Data Processing on Large Clusters,6th Symposium on Operating Systems Design and Implementation,2004
[13]http://blog.trifork.com//wp-content/uploads/2009/08/MapReduceWordCountOverview1.png
[14]M Memedi,A web application for follow-up of results from a mobile device test battery for Parkinson’s disease patients,COMPUTER METHODS AND PROGRAMS IN BIOMEDICINE 104 (2011)
[15]Wikipedia:[Internet], Available fromhttp://en.wikipedia.org/wiki/Case_mix_index
[16]National Health Insurance Research Database:[Internet], Available fromhttp://nhird.nhri.org.tw/date_02.htm
[17]EddDumbill,The SMAQ stack for big data,2010
[18]LingPipe,[Internet], Available fromhttp://alias-i.com/lingpipe/
[19]中華民國監察院:健保濫用調查報告-2013[Internet], Available from
http://www.cy.gov.tw/AP_HOME/Op_Upload/eDoc/調查報告/102/102000084健保濫用調查報告.pdf
[20]Baltimore Sun Graphic,2009
[21] Seth Flam DO,The Longitudinal Patient Record — A virtual “MRI” of your patient’s medical record,2011
[22]游斯雯,Characteristics and Medical Expenditures of Children Aged 0-19 Hospitalized for Injuries in Taiwan, 2006-2008,The Journal of Health Sciences,14:1,2012
[23]Chao-Hsiun Tang,One-year post-hospital medical costs and relapse rates of bipolar disorder patients in Taiwan: a population-based study,Bipolar Disorders 2010: 12: 859–865,2010
[24]Tien-Chiung Hung,Trend Analysis of Hospital Resource Utilization for Prolonged Mechanical Ventilation Patients in Taiwan: A Population-Based Study,RESPIRATORY CARE ‧ APRIL 2013 VOL 58 NO 4,2013
[25]Po-Huang Chiang et al,Healthcareutilizationandexpenditureanalysisbetween individualswithintellectualdisabilitiesandthegeneral populationinTaiwan:Apopulation-basednationwide child and adolescentstudy,Research inDevelopmentalDisabilities34,2013
[26]Cattell, R., “Scalable SQL and NoSQL data stores” SIGMOD Rec. 39, 4, May 2011, 12-27.
[27]Pokorny, J., “NoSQL databases: a step to database scalability in web environment.” In Proceedingsof the 13th International Conference on Information Integration and Web-based Applications andServices (iiWAS ''11). ACM, 2011, 278-283.
[28]Cogswell, Jeff., “SQL vs. NoSQL: Which Is Better?” [Internet], July 2012, Available fromhttp://slashdot.org/topic/bi/sql-vs-nosql-which-is-better/
[29]Gu, Genqiang,An overview of newly open-source cloud storage platforms,Granular Computing (GrC), 2012 IEEE International Conference
[30]MongoDB White Papers, "A Brief Introduction to MongoDB", In http://www.IOgen.comiwhite-papers.