Disease Cost Estimation Model Based on Improved Grey Relational Analysis and Case-based Reasoning

Disease cost estimation is an important and challenging issue. It serves as the key factor of hospital cost and cost control. It has been a challenge to estimate the disease cost because of the complexity of the disease pathology, individual differences of treatment process, adverse drug reaction, and complications. Especially in some high risk and high noise diseases there is another challenge from small sample and poor data. Grey relational analysis is a kind of effective method to study small sample, and poor data. In this paper, an improved model based on grey relational analysis and case-based reasoning is proposed to estimate the disease cost. Firstly, the similarity on cost attributes is calculated through the grey relational analysis method. And then the optimal attribute weights is found with the cuckoo search algorithm and there are a few cases matching to the target case according to the similarity in the historical case database to estimate the cost of target case. At last the estimation accuracy using the improved model in experimental study of the cost estimation of simple appendicitis disease, cesarean section, and heart bypass surgery is compared with the three models respectively with Euclidean distance, Angle cosine, and Mahalanobis distance. The improved model is best on cost estimation precision, especially in small sample and disease with high risk and high noise.