Development of decision support system model for hypertension management using sequential patterns

보건정보관리학과/석사[한글] [영문]Data Mart was established in order to develop reminder system for managing hypertension risk group. Hypertension management reminder system was developed using hypertension risk factor prediction prototype presented by applying data mining. Consequently, 92,354 people who had no symptoms of hypertension from 1994- to 1998 who either had or did not have hypertension in 2000 were selected as researchsubjects out of a 4 year period medical record database of the National Health Insurance Corporation. The current research determined patients at risk of hypertension through logistic regression analysis and decision tree analysis and compared the effectiveness of the two methods. As a result of determining patients at risk of hypertension through logistic regression analysis, males who were older, had a longer period of obesity, had a longer period of binge drinking, and had smoked for a longer period of time had a more statistically significant and higher probability of suffering from hypertension. Main factors explaining risk factors that affect hypertension through decision making tree analysis were smoking, drinking, obesity when gender variable was excluded, and when gender was taken into account, obesity and drinking were risk factors for hypertension in men, and age and obesity for women. Decision tree analysis had lower accuracy (75.17%) and sensitivity (73.38%) than logistic regression analysis while having higher particularity (74.35%). Also, the current research has found a pattern where many risk factors appear by using sequential patterns and calculating the degree of influence on all combinations of each risk factor that affect hypertension with the passage of time. As a result, a sequential pattern able to predict potential hypertension patient with risk factors was proven. In this manner, reminder system for hypertension management was developed using rules deduced from data mining. The system organization is made up of Physical exam Review, Prediction, and Recommendation. It is the belief of the current research that its findings can greatly contribute in reducing the occurrences of hypertension patients if efforts for their early detection and primary preventive measures are taken by expanding and applying reminder system for hypertension managements.ope