應用關聯法則分析不同科別高血壓處置的差異

Apply Association Rules to Analyze Hypertension Treatment Difference Among Physicians in Different Departments

徐啟勝(Chi-Sheng Hsu)

指導教授 ： 詹前隆

元智大學/資訊學院/資訊管理學系/碩士(2005年)

摘要

本研究以高血壓患者為研究對象，分析高血壓患者在門診的就醫紀錄中，是否會因為就醫科別的不同，會有不同的處置（Treatment）偏好，在比較醫師的處方偏好時，需要在病患的病徵接近相同的情況下，較為客觀，依據JNC 7 REPORT的高血壓治療指導原則，當病患為罹患高血壓就醫時，其是否罹患其它的慢性疾病，將是影響醫師開立處方及檢驗項目的重要因子，本研究共納入心臟血管疾病、糖尿病、腦血管疾病、腎臟疾病及肺疾病等五種慢性病，視為相同病徵的一個依據，分別探討各科在開立處方及檢驗項目時的傾向，本研究使用卡方獨立性檢定（Chi-Sqaure Test for Independence），檢定科別與用藥、科別與檢驗項目是否相關，進而採用資料探勘技術中的關聯法則（Association Rule）進行分析，本研究初步發現，雖然在高血壓患者具有相同的病徵下，各科開立降血壓用藥有所差異，結果可提供醫師了解各科用藥的差異及藥師審查處方用藥合理性時的參考依據之一，亦可提供醫院管理者在管控藥品的參考。

關鍵字

卡方檢定 ； 關聯法則 ； 糖尿病 ； 獨立性 ； 腦血管 ； 高血壓 ； 慢性病 ； 高血壓

並列摘要

The objective of this study was to apply association rules to analyze hypertension treatment difference among physicians in different departments. The focus is on the hypertension clinic. According to the JNC VII REPORT, Physician’s administrations are influenced by different corbidity. This study investigate five systematic disease coming along with hypertension, such as cardiovascular disease, diabetes mellitus, cerebro-vascular disease, renal disease and lung disease. Chi-square test is a widely used method for testing the independence or correlation among attributes, we use it to test whether the physician and administration are correlated. In this study, association rule mining is applied to explain the relationship of attributes. Experiment results show that, although there is same systematic disease, there is different prescription between the physicians. Result can offer physician realize the different administration in the peer, it can also offer reference for pharmacist to check the accuracy of prescription. The finally, hospital’s manager control the medicines can reference the result.

並列關鍵字

Chi-Sqaure test ； Association rule ； Hypertension

參考文獻

1.Aram V. Chobanian, George L. Bakris, Henry R. Black,William C. Cushman, Lee A. Green,MPH Joseph L. Izzo,Daniel W. Jones,et al.（2003）.The Seventh Report of the Joint National Committee on Prevention,Detection，Evaluation，and Treatment of High Blood Pressure.JAMMA, 289,2560-2572.

2.Barbara Starfield, Is Us Health Really the Best in the Word ？, JAMA （2002），vol 284, 483-485.

5.Harry Road , San Jose , Mining generalized association rules , Future Generation Computer Systems 13 （1997）161-180

6.Margaret H.Dunham（2003）. Data Mining introductory and Advanced Topics. Kentucky：Prentice Hall.

7.Mehmed Kantardzic（2002）. Data Mining Concepts- Models, Methods and Algorithms. US：John Wiley & Sons Inc.