A Two-Stage Medical Expenses Estimation Model for Inpatients During Diagnosis Process Under Artificial Intelligence Environment

Abstract

Under artificial intelligence (AI) environment, medical staff and patients benefit tremendously from sophisticated AI algorithms in various healthcare activities. In order to enhance diagnosis performance for coronary heart disease (CHD) inpatients with surgery (e.g. stent), this research presented a two-stage medical expenses estimation model. Two intelligent modules were integrated into this model, SVM-based module and SOM-based module, and they were developed to estimate total medical expenses and detailed medical expenses. The model was also compared with classic AI techniques, back propagation neural networks and random forests. The data from a real hospital was introduced. For the target disease CHD, several attributes were extracted for inputs/outputs. Based on experimental results, the proposed model not only achieved excellent performance for total medical expenses estimation, but also a powerful tool for detail expenses estimation. The related managerial insights would assist medical staff and patients in reliable decision-making.