xml as standard for communicating in a document-based electronic patient record: a 3 years experiment

Abstract

During the past few years, the eXtensible Markup Language (xml) has progressively become a gold standard for accessing, representing and exchanging information, especially in the health care environment. This paper presents an implementation of the use of xml for the electronic patient record (EPR) and discusses more specifically its growing use in two areas of the EPR: first, as a format for the exchange of structured messages, and second, as a comprehensible way of representing patient documents. These statements rely on a 3 years experiment conducted at the Geneva University Hospital as part of its document-centered EPR.

Introduction

The eXtensible Markup Language (xml), completed in early 1998 by the World Wide Web Consortium (usually called the W3C) [1], is emerging as a universal format for describing the content and structure of electronic documents on the Web. Its growing use is promoted by the fact that xml is independent of any vendor, platform or application, together with the availability of useful tools for processing and browsing it. This text-based and human-readable language, which is a simplified form of the Standard Generalized Markup Language (sgml), allows users to define their own markup or tags for data description. Unlike formatting tags, as strictly used in the HyperText Markup Language (html) and that only reference the surface aspect of information, xml allows the use of content tags that are more expressive and thus likely to carry semantic interpretation. Therefore, xml documents can convey information both on the content of the document and on the format of the content, thus encouraging demarcation between presentation and content. Styles are then matched to tags in xml documents using the standard stylesheet language called xsl, an acronym for eXtensible Stylesheet Language. As healthcare requires information structures that are highly flexible and evolving, xml-based applications are gaining attention in the medical field [2], [3], [4]. However, these research fields are currently limited, as no well-accepted rules that describe how to structure xml documents and messages exist. In order to ensure the interoperability and consistent representation of medical data and clinical information embedded into patient records, researchers [5], [6], as well as several working groups and technical committees related to CEN, HL7 or ASTM to name a few [7], [8], [9], [10], are currently addressing both the formal aspects of xml standards for data interchange and the semantic aspects for textual data. All these organizations working within the healthcare arena strive to comply with the xml standards defined by the W3C.

At the Geneva University Hospitals (HUG), a dedicated software tool called domed (a French acronym for “DOssier MEDical”) has been first developed in DELPHI 5, for rapid access to patients’ electronic documents. This application, in use since 1999, is currently replaced by a new version renamed DPI (a French acronym for “Dossier Patient Intégré” i.e. “integrated patient record”). This release broadens the functionalities of the electronic patient records (EPR) in the HUG as it allows, from now on, the managing of images, laboratory results, questionnaires, and medical orders, as well as the edition of new documents. This application is installed on more than 1200 client computers at the five campuses of the HUG, and has already been run by more than 1500 different users. The basic architecture of DPI relies on a multi-tier architecture. Basically, three independent layers share the various tasks: first, a storage layer allows for the storage and retrieval of both data from a structured relational database, and of full text narratives from a file oriented database; second, a middleware layer, which is a business logic layer, carries out most of the functionalities of the application; and third, a presentation layer deals with the user interface on the client-side application. This separation simplifies the overall developments and allows the best technology to be selected for each layer. As xml offers a good level of acceptance within the HUG applications teams and also presents a good capability to evolve, it has been integrated into all layers, more specifically for the exchange of messages between layers and for the representation of patient documents. Therefore, shareable services within the HUG were adapted for accepting xml as the syntactic structure for supporting inter-components messages and the HyperText Transfer Protocol (HTTP) [11] was chosen for supplying the exchange of information. Besides, an xml-compliant paragraph oriented structure was developed to formalize and normalize the narrative documentation produced within the HUG. These various aspects, leading to a better-integrated EPR that clearly separates knowledge from applications, are specified and discussed in detail in the following sections. The overall strategy and experience in integrated xml/HTTP technologies within the clinical information system of the HUG is precisely described in another paper [12].