Study On Clinical Decision Support Based On Electronic Health Records Data

The medical industry of China has been developing at a high speed to a huge volume,and demands for high quality medical services are growing.The application of electronic medical records(EHRs)and clinical decision support systems(CDSSs)is considered to be an effective way to improve medical service quality.At present,the hospital information construction has delivered some good results,taking patients as the center of diagnosis and treatment patterns,and data produced by whole process of clinical activities can be recorded.At the same time,the traditional CDSSs based on knowledge have difficulties in keeping and updating knowledge with the pace of medical development,and cannot be seamlessly integrated into the clinical workflow,which limits the development of CDSSs.This paper researched the clinical decision support based on EHR data in three aspects,including the comprehensive hospital management,doctor-patient communication and optimization of electronic medical record system work process,by building a data-driven knowledge discovery process to serve as a new source of knowledge for making clinical decisions and connecting EHR and CDSS,to improve the work efficiency of physicians and the quality of medical services.The main innovation points are as follows:According to the situation of hospital information construction,a real-time medical data analysis platform was proposed,with overall architecture and detailed design scheme.Data warehouse technology,online analysis processing(OLAP)are introduced to archive real-time data import and multi-dimensional,multi-level data analysis without affecting the other hospital information systems.Both the clinical and administrative information are transformed into knowledge and provided to hospital administrators by proper visualization technologies for clinical decision support.A novel self-learning EHR system framework was proposed by constructing a direct pathway between the EHR workflow and EHR data reuse.Using this framework,a prototype of this framework was implemented based on patient similarity learning.A recommended order menu is provided by the EHR's self-learning regarding patient similarity and average behavior of physicians,adding intelligence to the EHR.The proposed EHR enhances the working efficiency of the physicians,also can serve as education or references for junior physicians.Real-world EHR data were substituted to verify the availability and effectiveness of the system.Shared decision making(SDM)was studied,where physicians are encouraged to involve patients into the clinical decision making process.According to SDM,this paper put forward a decision aid system for patients with type 2 diabetes medication choice connecting to EHR,which combines clinical data and clinical guidelines,jointly,as the basis of decision support.The decision aid could make a recommendation between 8 classes of hypoglycemic medications based on patients'EHR and help tailor the knowledge for physicians and patients the decision making process.In this paper,data warehouse,OLAP and data mining technology were introduced into the process of building CDSSs,for perform the specific research of data-driven CDSS in three different clinical application scenarios.In order to provide the right information at the right time for the right people,these CDSS applications are capable of assisting the hospital administrators,optimizing the physician-patient communication,and improving physicians' work efficiency,thus achieving the goal of improving medical service quality.