The Research And Implementation Of MPI Servers In Regional Health Information

Medical information sharing and collaboration between different medical institutions within the region has become an inevitable trend of development of medical and health information, and it is also a research focus. The unique identification and integration of the patient’s identity is the basis for the realization of regional health information sharing. The master patient index is the unique encoding which is used to uniquely identify each patient instance within the specific region.To unify identification of patients in different institutions in regional range, and to solve unique master index of patient and index service of electronic health records problems, we designs and implements the master patient index server. This master patient index server is used in regional health data exchange and sharing platform. On one side, it provides the service that is used to query the master index of specified patient and the health record indexes associated with the master patient index for the application layer, on the other side, it provides the service that is used to establish or match the master patient index for the medical data collected by front-end processor, and to implement the function like making health data associate with patient entity.Firstly, personal identification and integration technologies are investigated;this paper adopts the the identifier match and merger method by comparison and analysis;The Patient Identifier Cross-referencing Integration Profile as the mainstream technology to solve the problem of the ID match and merger is particularly analysed, and the cross-domain MPI collaborative mode is particularly analysed. Secondly, this paper introduces the patient master index server application environment:regional health data exchange and sharing platform; this paper designs and implements these function modules of the master patient index server, such as index query, index register, MPI merging, MPI splitting.In addition, as the difference and complexity of patient identifiers from different medical institutions, and large volumes of the data collected by front-end processor, and the characteristic that data exchange platform requires to frequently access the master index server in a short time, this paper applys a series of strategies and technologies for enhancing the performance of the master patient index server. By analyzing the data items for recognizing patient identity, this paper proposes one effective patient identity matching strategy that can be flexibly configured with different priority, to achieve flexible and efficient patient identity matching; This paper adopts the method of building index in memory to shield access of the database as patient identity is matching, directly to complete patient identity match in memory; In addition, by adopting asynchronous processing of multithread and epoll+thread pool technologies, it can greatly improve the speed of operation, the utilization of CPU, and the ability that the system handle concurrency; Further more, by system configuration of XML, log support of log4cpp, real-time monitoring of watchdog, it can improve the flexibility and integrity of the system. By testing, it is found that the system can achieve the desired functionality. Finally, the summarization of the whole research work is generalized, and the inadequacies of the present work are pointed out.