Using Data Mining to Evaluate Patient-oriented Medical Services for Chronic Senility Outpatients

The problem with previous research of health care service were failed to isolate the study objects. Therefore, the purpose of this study is using data mining to analysis disease clusters of chronic senility to enhance quality of health care service. This study used cluster and association analysis of data mining to analyze the health insurance data of outpatients suffering from chronic senility in a hospital in Taiwan, over the period from January to December 2002 (N = 5836). According to analysis of revisit frequency, and disease correlation, the patients were grouped into different clusters, after which expert interviews discovered target clusters with abnormal numbers of revisits. This information was assist planning service strategy for difference groups of patients. Through analysis, two target clusters were isolated, Clusters 4 and 7. Cluster 4 (n=114), had excessive return visit times, and had 13 chronic diseases on average, with 27.2 revisits per year. Cluster 7 (n = 426), had in frequent return visits, and had 4 chronic diseases on average, with 2.68 return visit times per year. After expert interviews, the goal for Cluster 4 was to effectively control chronic diseases, to enhance the patient health and to raise satisfaction levels. The goal of Cluster 7 was to promote patient loyalty.