Research on Data-Driven Two-Sided Matching between Doctors and Patients

摘要

Objective: Based on the theory of two-sided matching, using data-driven statistical methods to provide methodology for two-sided matching of doctors and patients. Methods: Firstly, Calculate characteristic statistics of doctors and patients based on HIS system data, and the difference characteristics of patients contacted by different doctors were identified as the evaluation index of matching satisfaction. Then, Fisher discriminant is used to calculate the weight of matching satisfaction index.Construct two-sided matching function with discriminant function.According to the paste progress of matching function value, the two-sided matching between doctors and patients is realized. Results: Using 13748 HIS data of two doctors in thoracic medicine department of a large hospital, 22 matching satisfaction evaluation indexes were extracted by the matching method given in this study. And a matching test was carried out with the actual patient data, which proved the validity of the matching results of this method. Conclusion: The matching method presented in this study can realize the process of doctor-patient matching from patients' characteristics to doctors' labels, from doctors' labels to patients' matching and sticking progress, and provide methodology for data-driven doctor-patient two-sided matching. The research of this paper is one to many matching, and the matching satisfaction index is limited. In the future, we can further expand the research perspective and carry out the research of intelligent matching system for doctors and patients on Internet platform.