現今醫學愈來愈講求實證，凡事皆以經驗去解決照護病人所遇的臨床問題已經不適合了，而是必須應用最佳的臨床研究成果於病人照護上，落實實證醫學，才能提升照護病人的品質。目前有許多醫學文獻資料庫可供臨床醫師搜尋所需的相關實證，其中以Medline為提供最完整且豐富的醫學文獻。但面對如此繁雜且龐大的資料庫，臨床醫師在找到所需的資訊前，常常在瀏覽和過濾這些文獻花費不少時間，雖然已有實證醫學資料庫被發展來讓臨床醫師快速獲得實證資料，但由於是人工建置，文獻的更新速度比一般的原始資料庫慢，無法像原始資料庫擁有較多較完整的資訊。所以本論文發展一套由臨床醫學文件中萃取PICO之文件摘要系統，萃取文章中的PICO資訊，即病人描述資訊、給予的治療、比較的治療、及治療結果。以期臨床醫師在搜尋最佳實證的文獻資料時能更快速的了解每篇文章的要旨，進而幫助臨床醫師更快速的找到最佳實證。
本論文的方法著重在分析摘要文章的結構和語意的剖析，根據所得到的摘要文章結構和詞彙語意，利用關鍵詞句、樣版、上下文等資訊來進行PICO萃取，最後以條列方式呈現PICO。最後實驗結果顯示，在系統萃取PICO的精確率評估上，有相當不錯的效能。

Evidence is more and more important in the medical domain. It is not suitable to solve problems by experience all the time for patient care. Instead of this, the clinician has to apply the best clinical study results to patient care and promote the quality of patient care by fulfilling the Evidence-Based Medicine (EBM). There are many medical literature databases to provide search the related medical evidence these days. Among these, the Medline provides the most complete and rich medical literature. Yet, the clinician often spends much time reading and filtering the literature in face of the miscellaneous database. The clinician can acquire the evidence quickly from the EBM databases, but the EBM databases can’t hold the complete information like the primary databases and the update speed of the EBM database is slower than the primary databases. In this thesis, we developed a document summarization system by extracting the PICO from the clinical medical articles. The P stands for Population. It means the information regarding patients; the I stands for Intervention. It means offering the agents or the clinician’s acts of dealing with the patient’s problem; the C stands for Comparison. It means the alternative intervention; the O stands for Outcome. It means the effects of the intervention. Hope let the clinician realize the main ideas in the citation more quickly. The strategy proposed in this thesis is to analyze the structure of the abstract and parser citation semantically at first. And based on this, exploit the key phrases, patterns, and local contexts to extract the PICO.

[1]Tracy CS, Dantas GC, Upshur RE. Evidence-based medicine in primary care: qualitative study of family physicians. BMC Fam Pract 2003;4: 6.
[2]Haynes RB. What kind of evidence is it that Evidence-Based Medicine advocates want health care providers and consumers to pay attention to? BMC Health Serv Res 2002;2: 3.
[3]Bradt P, Moyer V. How to teach evidence-based medicine. Clin Perinatol 2003;30: 419-33.
[4]ACP Journal Club. Available at http://www.acpjc.org/
[5]Cochrane Library. Available at http://www3.interscience.wiley.com/cgi-bin/mrwhome/106568753/HOME?CRETRY=1&SRETRY=0
[6]Evidence Matters. Available at http://www.evidencematters.com/emweb/gotoSimpleQuestionWizard.do;jsessionid=BE5B794F880A688FDD8A009CE4E7C7E3
[7]Bandolier. Available at http://www.medicine.ox.ac.uk/bandolier/
[8]PubMed. Available at http://www.ncbi.nlm.nih.gov/sites/entrez?db=pubmed
[9]PubMed Help Available at http://www.ncbi.nlm.nih.gov/books/bv.fcgi?rid=helppubmed.chapter.pubmedhelp
[10]PubMed Clinical Query. Available at http://www.ncbi.nlm.nih.gov/entrez/query/static/clinical.shtml
[11]Corrao S, Colomba D, Arnone S, Argano C, Di Chiara T, Scaglione R, Licata G. Improving efficacy of PubMed Clinical Queries for retrieving scientifically strong studies on treatment. J Am Med Inform Assoc 2006;13: 485-7.
[12]Efficient Medline search filters for clinical queries.
[13]West S, King V, Carey TS, Lohr KN, McKoy N, Sutton SF, Lux L. Systems to rate the strength of scientific evidence. Evid Rep Technol Assess (Summ) 2002: 1-11.
[14]Richardson WS, Wilson MC, Nishikawa J, Hayward RS. The well-built clinical question: a key to evidence-based decisions. ACP J Club 1995;123: A12-3.
[15]Huang X, Lin J, Demner-Fushman D. Evaluation of PICO as a knowledge representation for clinical questions. AMIA Annu Symp Proc 2006: 359-63.
[16]Niu Y, Hirst G. Analysis of Semantic Classes in Medical Text for Question Answering. In Proceedings of the 42nd Annual Meeting of the Association for Computational Linguistics, Workshop on Question Answering in Restricted Domains 2004.
[17]UMLS. Available at http://www.nlm.nih.gov/research/umls/
[18]Bodenreider O. The Unified Medical Language System (UMLS): integrating biomedical terminology. Nucleic Acids Res 2004;32: D267-70.
[19]UMLSKS. Available at http://umlsks.nlm.nih.gov/
[20]Pratt W, Yetisgen-Yildiz M. A study of biomedical concept identification: MetaMap vs. people. AMIA Annu Symp Proc 2003: 529-33.
[21]Aronson AR. Effective mapping of biomedical text to the UMLS Metathesaurus: the MetaMap program. Proc AMIA Symp 2001: 17-21.
[22]Divita G, Tse T, Roth L. Failure analysis of MetaMap Transfer (MMTx). Stud Health Technol Inform 2004;107: 763-7.
[23]CATAL? N, CASTELL N, MART?N M. A portable method for acquiring information extraction patterns without annotated corpora. Natural Language Engineering 2003.
[24]Proceedings of the Third Message Understanding Conference (MUC-3). 1991.
[25]Document Summarization. Available at http://www.nsysu.edu.tw/TANET99/DownLoad/TANET073/TANET073.DOC
[26]Fiszman M, Rindflesch TC, Kilicoglu H. Summarizing drug information in Medline citations. AMIA Annu Symp Proc 2006: 254-8.
[27]Fushman DD, Lin J. Answer extraction, semantic clustering, and extractive summarization for clinical question answering. In Proceedings of the 21st International Conference on Computational Linguistics and 44th Annual Meeting of the ACL 2006.
[28]Rindflesch TC, Pakhomov SV, Fiszman M, Kilicoglu H, Sanchez VR. Medical facts to support inferencing in natural language processing. AMIA Annu Symp Proc 2005.
[29]Fushman DD. Complex question answering based on a semantic domain model of clinical medicine (PhD thesis). In Maryland 2006.
[30]Fushman DD, Lin J. Knowledge extraction for clinical question answering: preliminary results. In Proceedings of the AAAI-05 Workshop on Question Answering in Restricted Domains 2005.
[31]Demner-Fushman D, Few B, Hauser SE, Thoma G. Automatically identifying health outcome information in MEDLINE records. J Am Med Inform Assoc 2006;13: 52-60.
[32]Niu Y, Hirst G, McArthur G, Rodriguez-Gianolli P. Answering clinical questions with role identification. In Proceedings of 41st Annual Meeting of the Association for Computational Linguistics, Workshop on Natural Language Processing in Biomedicine 2003.
[33]Niu Y, Zhu X, Hirst G. Using outcome polarity in sentence extraction for medical question-answering. AMIA Annu Symp Proc 2006: 599-603.