Building a two-stage integrated model for outpatient appointment service

Abstract

There are many medical departments in Taiwan's hospitals which make the patients not easily selecting the appropriate outpatient appointment. Patients often make their selection based on their own experience. This may lead incorrect appointment, delay of treatment, and waste of time. To help patients in making appropriate choice of outpatient appointment, this study has built a two-stage integrated model for decision support applications. artificial neural networks (ANN), support vector machine (SVM), and classification and regression tree (CART) are used in combination for modeling of the system. The first is to input the symptoms into the classifier and generate the category of outpatient department. The second stage is for continued classification of input symptoms and generates the recommended medical specialty/division for appointment. The results show ANN-SVM model has achieved the highest overall yield of 93.94%. ANN-CART and SVM-CART models achieved 80.75 and 78.57%, respectively. This two-stage model is a cost-effective tool in medical decision support and helps providing efficient medical service.