Collating of a Distributed XML-based Medical Records into a Relational Database

We are currently developing the Personal Internetworked Notary and Guardian (PING) whose main goal is to implement a secure, distributed and patientcontrolled repository of life-long medical information on the Web. The PING record differs substantially for institutionally-oriented medical information systems. First, it uses XML as a basic data format and virtual directory structure as data storage and does not employ a relational database management system. Second, it is a highly secure system under a patients’ own control on the bases of authentication, authorization, and encryption technology. Third, PING records are distributed over the Internet and not confined to a single institution’s database. Fourth, PING records may have multiple data sources including patients as well as several health care providers. PING records could therefore provide very extensive, comprehensive and valuable information that other medical information systems may not encompass. However the essential distributed nature of PING records imposes several functional limitations in their use as a data warehouse for public health research. Consequently, we have addressed the construction of an efficient data management system required to generate useful public health information from PING records. We have developed a PING Poller, which retrieves PING records via a PING server and parses, reorganizes these data and inserts them into a relational database management system. This collation system, across widely distributed health records, will allow health care professionals and public health researchers to access PING records for clinical, epidemiological and public health researches, if given appropriate authorization by patients.