運用全民健保資料庫建立中醫藥與疾病之關聯知識庫

近年來，替代及輔助醫學在全球之使用率漸增，而中國傳統醫學為其中之一。根據衛生福利部之統計資料，在臺灣約25%的人口曾透過中醫門診治療疾病，故我們應需逐漸注意中醫之醫療品質及病人安全。投藥治療為治療疾病的方法之一，若醫師開立不適當的藥物給病患可能影響病人之健康狀況，甚至導致病情惡化。因臺灣中醫之疾病診斷代碼為國際疾病分類第九版(ICD-9-CM)與西醫相同，此為全世界之首例，在其他國家中醫的疾病診斷是利用文字敘述的模式，相對之下，使用國際疾病分類第九版紀錄疾病的診斷較便於整理及分析資料。本研究以2000~2011年之全民健康保險研究資料庫百萬歸人檔、2007 ~ 2011年的台北醫學大學三院的中醫門診資料庫為樣本，建立中藥與疾病、中藥與中藥之關聯知識庫(Knowledge Base)。本研究方法包含三個步驟，首先運用資料探勘的關聯規則尋找出中藥與疾病、中藥與中藥各種組合及關聯性，並將關聯性強度量化，建立出中醫藥與疾病之關聯知識庫，再運用相同方法以某醫院的中醫門診資料庫為樣本建立另一知識庫。接著，運用皮爾森相關係數 (Pearson correlation coefficient)分析這兩個知識庫。最後，透過敏感度分析尋找出適合用於偵測不適當處方籤之閥值。本研究樣本來源為健保資料庫之百萬歸人檔，為全臺灣的抽樣樣本，不僅具代表性亦可供全國醫療院所參考與使用。故本研究預期結果為建立具有優勢，且可普及適用於任何一家醫院的知識庫，透過敏感度分析(Sensitivity analysis)找出適合的閾值，提升偵測不適當處方的靈敏度。期望在未來可與中醫門診醫令系統整合，降低用藥疏失，提升中醫醫療品質。

關鍵字

中醫；科學中藥；健保資料庫；巨量資料；決策支援系統；關聯性資料探勘

並列摘要

In recent years, the utilization rate of alternative and complementary medicine in the world is rising, and Chinese traditional medicine is one of alternative and complementary medicine. According to statistics of the Ministry of Health and Welfare in Taiwan, about 25 % of the population had to treat diseases through Chinese traditional medicine, so we should need to pay much attention to patient safety and quality of care of Chinese traditional medicine. In Taiwan, Traditional Chinese medicine use the International Classification of Diseases, Ninth Revision (ICD-9-CM) as the code of disease diagnosis. Western medicine is also using the same way,and this is the world's first case. In other countries, Traditional Chinese medicine diagnosis is to use text description. By comparison, using this way to record is more easy to organize and analyze data. The sample database in this study was using the millions of people file of the National Health Insurance Research Database is recorded from 2000 to 2011,and the Taipei Medical University Hospitals of Chinese medicine outpatient database is recorded from 2007 to 2011. Using thess database established the knowledge base of association of disease - Scientific Chinese Medication and Scientific Chinese Medication - Medication. The research method consists of three steps, the first use of data mining association rules to find out the various combinations of disease-Scientific Chinese Medication, and Scientific Chinese Medication- Medication.And to quantify the strength of association to establish a knowledge base in disease-Scientific Chinese Medication, and Scientific Chinese Medication- Medication, then use in the same way to a hospital outpatient of traditional Chinese medicine database to establish another knowledge base. Then, using Pearson correlation coefficients analysis these two knowledge base. Finally, through a sensitivity analysis to find out cut-off value(Q) for detection of inappropriate prescription. The study sample source for millions people of NHIRD, This sample is a national sample, not only representative is also available for reference and use of medical institutions nationwide. Therefore, the expected results of this study for the establishment of an advantage, and can be applied to any one hospital spread of knowledge, through sensitivity analysis to identify the appropriate cut-off value to improve the sensitivity of detection of inappropriate prescribing. In the future, expecting this knowledge base can be integrated in Chinese medicine outpatient physician order systems, to reduce medication errors, and improve quality of care medicine.

並列關鍵字

Traditional Chinese Medicine ； Science and medicine ； Health Insurance Database ； Big Data ； Decision Support System ； Association Rule Mining

參考文獻

石崇良, 林仲志, 廖熏香, 楊漢湶, & 翁惠瑛. (2007). 台灣病人安全通報系統三年經驗. 臺灣醫學, 11, 298-305.

周珮琪, & 林昭庚. (2010). 臺灣早期中醫發展簡史. 北市中醫會刊, 16, 23-33.

林昭庚, 陳俊明, 施純全, & 陳欣欣. (2001). 中醫師人力現況與規劃. 中醫藥雜誌, 民 90 年, 12, 231-242.

施純全. (2011). 台灣中醫醫療利用之研究: 醫療資源, 人口地理社經因子及健康行為之影響. 中國醫藥大學中醫學系博士班學位論文, 1-253.

許明暉. (2011). 臺灣健康資訊科技之現況與未來.

國際替代計量

運用全民健保資料庫建立中醫藥與疾病之關聯知識庫

全文下載

主題瀏覽