病歷電子化讓病歷資料更容易被保存,龐大的資料庫可做為學術研究與業務改善的參考資料。本研究以台灣某實證醫院病歷資料庫作為資料來源,以病歷號為關聯鍵整合病患住院與門診病歷,並選取病患主診斷、次診斷及基本資料做為研究對象。為了找出容易引起病患住院的門診疾病,本文提出資料編碼排序法(Data Encoding and Sorting Method, DESM)建立病歷資料關聯規則,經實例驗證,資料編碼排序法除了可以正確得搜尋關聯規則外,執行速度也優於傳統的Apriori演算法,當支持度為2%的情況下,資料編碼排序法較Apriori提高了93.77%的運算效率。最後,根據本研究搜尋的關聯規則進行文獻驗證,符合文獻驗證的規則可作為病患健康管理與醫護人員追蹤治療之參考,尚無文獻驗證的規則可成為未來醫學研究的方向。
The medical records electronic medical records are more likely to be saved, a large database can be used as a reference for academic research and business improvement. In this study, a Taiwan empirical hospital medical records database as a data source, medical record number for the associated integration of inpatient and outpatient medical records, and select patients with primary diagnosis, sub-diagnosis and basic information as the research object. Order to find out easily lead to inpatient, outpatient disease, this paper presents the sort of data encoding (Data Encoding and the Sorting Method DESM) the establishment of the medical records of association rules, instance validation, data encoding, sorting method, in addition to correctly search for association rules, perform The speed is superior to the traditional Apriori algorithm, support for 2%, the sort of data encoding Apriori improve computational efficiency of 93.77%. Search for association rules, according to this study document verification, and literature validation rules can be used as tracking a reference to the treatment of patients with health management and medical staff, there is no literature validation rules may be the future direction of medical research.