Hospital Audit Data Analysis Method Based On Data Mining Technology

The hospital audit mainly focuses on examining whether the hospital has excessive medical treatment,illegal charging or falsely reporting medical expenses to obtain medical insurance funds.The common audit method is to obtain the data in the hospital information system according to the relevant provisions of the policies and regulations,compare and analyze the actual behavior with the regulations to find the audit doubts.However,due to the lack of professional medical knowledge,it is difficult for auditors to further analyze the specific diagnosis and treatment process and medication details in the doctor’s orders and prescriptions.According to this problem,this thesis proposed a method for analyzing the compliance of hospital charges based on a clustering algorithm and a method for analyzing the rationality of hospital diagnosis and treatment based on the Apriori algorithm.Besides that,to directly and quickly analyze the problem of medical insurance settlement,this thesis also proposed a medical insurance settlement analysis method based on artificial neural network.Aiming to facilitate ordinary auditors to use data mining algorithms to analyze,a software system is designed and implemented in this thesis,which mainly included the following aspects:(1)Preliminary preparation for data analysis,including determining data analysis ideas and objectives according to audit objectives;investigating and understanding the hospital’s routine information system and data;designing audit standard tables;collecting standard table data and verification data;(2)A solution based on clustering algorithm is proposed for the problem of charging compliance.First of all,it combined the derivation process of the clustering algorithm with real problems to get the solution to the problem;secondly,it used data mining tools to find out the categories with a small number of elements and few connections with other categories as suspicious points;and thirdly it combined the actual cases to confirm the suspicious points;finally,compared with the traditional analysis method,it showed the advantages of data mining algorithm.(3)Aiming at the rationality of diagnosis and treatment,a solution based on Apriori algorithm was proposed.Firstly,it combined the derivation process of the Apriori algorithm with real problems to get the solution to the problem,then used data mining tools to get the association rules between diagnosis and treatment items,in order to find out the diagnosis and treatment items that should be used but not as doubt points,and then combined the confirmed suspicious points in actual cases,and finally reflected the advantages of data mining algorithms,comparing with traditional analysis methods.(4)In order to analyze the medical insurance settlement problem directly and quickly,a solution based on artificial neural network method is proposed.Firstly,its derivation process is combined with the actual problem to obtain the solution.Secondly,using the data mining tool,the factors affecting the settlement amount are used as input signals,and the settlement amount is predicted and compared with the actual amount,with large difference.As a doubtful point,combined with the actual case to confirm the doubtful point,it finally mixed the traditional analysis method in order to reflect the advantages of the data mining method.(5)Since most auditors do not have the ability to use data mining tools,in order to facilitate their data analysis work,this thesis built a software system based on the above methods.In the system,administrators are responsible for data collection,data authorization and configuration of data mining algorithms,as a result,auditors can easily and quickly use the system to complete data analysis.