Research On Grid Data Loading Agent Based On HL7

Abstract With the continuous development of medical informationization, a variety of medical information systems with numerous medical clinical data have been established by many hospitals. The effective utilization of the data could serve for early-warning, forecast and tracking of acute infectious disease, chronic disease and endemic disease, which could provide strong support for doctors to make correct predications, preventions and prescriptions.However, medical information systems between hospitals even between different departments in a hospital are isolated with each other. The data within different systems are various in organization and storage. All of these bring great inconvenience to data collection. In order to solve the problems above, Grid Data Loading Agent (GDLA) based on HL7 integrates medical information systems from different hospitals through grid middleware, and collects heterogeneous medical information uniformly to provide data for medical application. GDLA is composed of data collection agent and data loading central server. Data collection agents are distributed in hospitals and responsible for collecting data from medical information system. Data loading central server is responsible for managing data collection agents, and uploading the data which is collected by data collection agents to the digital health data warehouse. GDLA adopts HL7 as data exchange protocol, and heterogeneous data of different hospitals could be transformed into the HL7 format through HL7 message middleware. Data loading central server collects HL7 format data. GDLA encapsulates virtual client to grid service and takes virtual client to access the original query interface of medical information system, which enables grid service to interact with medical information system as physical client. With visiting permission provided by medical information system, GDLA could conduct restricted querying operation to ensure security of medical data. GDLA utilizes idle time to execute data loading to reduce the burden of medical information system.Based on the CGSP gird middleware, GDLA is implemented by using programming languages such as Java, JSP, and the implementation has been tested. As the test result indicated, GDLA is able to load the clinical data to digital health data warehouse from heterogeneous medical information systems in a timely fashion. In the process of data loading, the conversion of HL7 message format accounts for about 10% of the overall processing time.