Informatics 99sep2

Journal of Informatics in Primary Care 1999 (September):2-9


Papers


Implications of WWW technologies for exchanging medical records

Maurice Dixon BSc PhD+, Stephen Cook BA MSc$*, Brian Read BSc DPhil CEng MBCS$

+ Computing Information Systems and Mathematics, London Guildhall University, London EC3N 1JY, UK; email M.Dixon@rl.ac.uk
$
Information Technology Department, CLRC Rutherford Appleton Laboratory, Chilton, Didcot, Oxon. OX11 0QX, UK; email b.j.read@rl.ac.uk
* now at: Science Systems (Industrial) Ltd., Fountain House, 2 Queens Walk, Reading, RG1 7QF, UK; email Stephen.Cook@scisys.co.uk


Abstract

This article addresses some of the implications for medical record exchange of very recent developments in technology and tools that support the World Wide Web. It argues that XML (Extensible Mark-up Language) is a very good enabling technology for medical record exchange. XML provides a much cheaper way of executing the exchange of medical information that circumvents the need for proprietary software. Use of XML can also simplify solutions to the problems associated with coping with the evolution of medical systems in time. However XML on its own does not resolve all the semantic heterogeneities.


Introduction

Expectations

The rapid development of the World Wide Web (WWW) has transformed the expectations and the technology of information systems1. The purpose of this article is to present some of the implications that arise for medical record exchange due to very recent developments in Web technology.

The Web is now a dominant and ubiquitous communication tool, leading members of the public as well as professionals to enjoy immediate (or at least, uniform) access to information across the world. With international travel becoming ubiquitous, it is reasonable to expect that medical records could be rapidly available anywhere in an emergency. However there are different medical systems in different countries.

Record exchange models

Recently we reported an architecture that had been developed for exchanging medical records between medical centres in the Czech and Slovak Republics2,3. A central feature of that project was the insistence that medical information systems were rapidly evolving. Any successful system for the exchange of medical records has to build into its architecture the capacity to handle change as the norm rather than as an exceptional circumstance. We used some Web-based protocols such as HTTP (HyperText Transport Protocol) and MIME (Multipurpose Internet Mail Extensions) encoding to overcome the heterogeneous nature of the different medical information systems. Hypermedata used an exchange file constructed according to rules imposed on suppliers. A complex graph modelling language was used to translate between different medical information systems4.

Hypermedia’s dynamic approach to the schemas of exchange documents was proposed as an advance on the fixed file definition approach implicit in UN/EDIFACT, which was set up originally as the standard for exchange of structured data such as purchase orders, invoices, and payments between businesses5.

Impact of new www technology

Although the central principle of designing for system change remains, it is fair to say that one aspect has been overtaken by the rapid evolution of web-based technologies. Specifically, the construction of the record exchange package could now be actualised using emerging WWW standards such as XML (Extensible Mark-up Language)6 and RDF (Resource Description Framework)7. The latest versions of standard Web browsers such as Internet Explorer 5 are beginning to include support for these standards (particularly XML), raising the possibility of using these tools as the Hypermedata client, rather than a specialised client for the customised graphical mark-up language that we devised when developing the Hypermedata prototype. This opens up new possibilities for exploiting the Hypermedata approach to information exchange.

These developments have important consequences even within the UK. For example, draft proposals have been published on the WWW by the Edinburgh-based Information Systems Support Group8 for XML-based versions of a patient discharge document called an Immediate Discharge Document (IDD) and Electronic Health Care Record Request and Provide messages9. Already there is a drive within the European Electronic Messaging Association’s EDI Working Group to create a UN Repository for XML tags based upon UN/EDIFACT to overcome EDI’s acknowledged weaknesses10.

We should not assume within the UK that the NHS has or can impose a monopoly on the format of medical information exchanges. Independently of the NHS there is a substantial private sector. Also different local medical information systems will evolve at different rates. XML capability will be available through cheap or free desktop browsers so local users will have the tools to exchange medical information on their own initiative. Furthermore there is a need to exchange medical records with medical centres outside the UK.

Esprit’s W3C-LA project

A recent EC Esprit project has been dedicated to ensuring that businesses are aware of the potential benefits of using Web technologies. This is the "World Wide Web Consortium – Leveraging Action" (W3C-LA) project, which was undertaken jointly by the Rutherford Appleton Laboratory and INRIA, the French national research institute for information technology11. The project was successful both in encouraging more European enterprises to join the World Wide Web Consortium (W3C) and in publicising the standards, including XML and RDF, which W3C has developed. The NHS is a member of W3C and a number of projects are already looking at the application of W3C standards and technologies to health informatics. Use of WWW approaches could also tie in with some of the thinking behind the UK National Electronic Library for Health Project12.

Reasons for applying XML approach to medical record exchange

This article argues how WWW-based XML technology provides advantages for patient record exchange over both UN/EDIFACT and over our recently implemented Hypermedata approaches to record exchange. It does so by removing key elements of the exchange mechanism from proprietary software and file management into freely and universally available Web browsers and other tools based on open standards. However it is important to remember that underlying the exchanged document structure there is a hierarchical data model. This hierarchical structure means that care must be exercised if different exchange packages are assembled and used for another purpose from the original intention. Careless use of an XML-based approach could increase the risk of incompatible semantics.


The challenges of medical record exchange

Introduction

In our recent paper we discussed in considerable detail the issues associated with medical record exchange2,3. The fundamental difficulty with exchanging medical information is integrity, particularly the preservation of semantics. Traditionally, solutions to this problem relied very heavily on direct exchanges of information between medical practitioners who shared a common understanding of the matters in hand and often knew each other personally. This problem becomes much harder to handle reliably if automation is required, particularly when systems are changing rapidly. Thus the exchange package and the target system must be able to respond to changes in the source system of the medical record. Ad hoc exchanges between unfamiliar systems present an even greater challenge to ensure that the exchanged data mean the same thing at the receiving system as they do at the sending system. Finally, the confidentiality of the information transferred needs to be guaranteed.

Resilience to change

An exchange mechanism needs to be able to adapt to change over time. Ideally the restructuring of the exchange package should be transparent to the requester who would simply identify what was being requested. Database research has already shown that an approach to inter-operating information systems based upon globally defined schemas cannot work for non-centralised information systems13.

Changes can arise in the underlying medical computer systems when it becomes necessary to record a new medical procedure, or there is a change in legislation, or the organisational structure, or the business practice. Ideally it should be possible to respond to an urgent one-off request from another country for a patient who needs emergency treatment. There will be changes in types of data exchanged: already many British GPs wish to have computerised access to X-rays, CT scans and MRI scans14. An exchange system needs to be robust with respect to enhancements of hardware and operating system2,3.

It is assumed that there is multi-way record exchange between collaborating medical centres taking place in addition to the centralised distribution from a hospital of structured discharge information at the point of discharge. These decentralised exchanges may need tuning frequently.

Semantic heterogeneity

The key feature of compatible source and target systems is that the exchanged data are interpreted in the same way at both ends of the exchange15. The differences that could arise are called semantic heterogeneities and are regarded as the major obstacle to exchanging data between systems. This is a subtle and difficult problem and should never be underestimated because incorrect results can be produced without any warning of error15.

One classic example is the case of an unconscious patient receiving emergency hospital treatment: how can one be sure about the specific identity of a patient after his or her discharge from the Accident and Emergency department? Quite sophisticated techniques are used in multi-database systems to clean up identification problems, since names are relatively unreliable ways of identifying people16. Colomb & Orlowska exposed the occurrence of the identity problem even in the joining of two simple relational databases: they showed that additional information was needed to be able to assert that a record in one system corresponded to the same person in another system. It is a general problem that cannot be avoided17.

Another example of semantic heterogeneity arises when the same term means different things. Within the university sector, vacation address could be used to mean the address at which a staff member or student lives out of term time. It becomes the effective address of contact at certain specified periods during the year. Outside the university sector, the vacation has the connotation of a short-term holiday. Within one health centre this is unlikely to be problematic since everyone at the centre knows the context. The problem arises when another centre needs to share and act on that information; both centres need to reconcile their interpretations of the meaning. It was a misunderstanding over "last addresses" of service personnel within hospital provision that Heiler used to illustrate how failure to reconcile semantics can lead to completely wrong conclusions and action15.

Different systems structure their data in different ways. For example, at the field level many systems conflate the fields of dosage quantity and dosage units, whereas others have explicitly separate fields. If a maximum dose constraint is applied to the dosage quantity field it is essential that the correct units are used. Similar considerations apply to dose frequency. Note that the medically significant constraint is more likely to involve the product of dosage and frequency. An example at the table or object level is that one system may record the consultant in charge of the patient whereas another may record other significant clinicians as well. Mapping the latter onto the former would result in significant information loss.

Security

Any exchanged information must have its content secured so that only users with the appropriate level of authorisation can access and interpret it..


World wide web changes the record exchange problem

The first strength of the World Wide Web approach to collaboration between information systems is that it provides a route which can be regarded as universally available. The protocols that define interconnection allow almost all platforms to intercommunicate in a standard way. This eliminates most of the system heterogeneity problems. The browser tools are distributed either free or at marginal cost and so are universally available on new systems and most older systems. This means that a system built using WWW browser facilities is not dependent for interconnectivity on a proprietary medical information system supplier. Even the poorest parts of Europe have Web access available for emergency use. The WWW applications are dependable because they are so extensively used that most foibles will have been detected and reported before they could seriously impact on a medical record exchange. Although the current releases of the most popular browsers show some incompatibilities, these are primarily concerned with the detailed rendering of information, rather than its encoding or transmission. Compared with most other kinds of software, the development of the Web is remarkable for its strong trend of convergence towards open standards, namely those sponsored by W3C.

Information security can be achieved in several ways. Encryption of the exchanged record through public key encryption would enable the integrity of the data to be maintained on either an intranet or over the World Wide Web.

The World Wide Web and the internet are often treated as synonymous, but the use of Web-based tools does not, of course, require that the internet is used to transport the exchanged record package. The European Union through Esprit projects like W3C-LA is actively promoting the deployment of Web tools in information systems development11. Increasingly, these information systems will not be confined to using the internet for transport. In particular, the mobile telephony industry is actively exploring how to add WWW access to its capabilities18. This could be invaluable in several health care settings, particularly "scene of accident" situations.

Part of the expectation is that there will continue to be cheap and compatible upgrade paths for Web browser-based tools that will be universally available. Future versions of browsers will be adapted to particular hardware devices, not just PCs, and the Web Content Accessibility Guidelines19 developed by W3C will guide browsers in how they should render different kinds of content for different capabilities of both hardware and user.


XML for the mark-up of record exchange packages

Record exchange

XML has been introduced as a way of transferring structured information across the WWW in a way that will be recognised by standard tools at the receiving end and be readable by humans (because XML encoding uses only plain text). XML provides a way of labelling the structure of information and, through a Document Type Definition (DTD), of defining the structure of particular kinds of well-formed information. It is an Extensible Mark-up Language because it allows a user to define the labels, which provide the structure, and separate this from the presentation format, defined by style sheets (which may themselves be written in XML)20. Now these are what we require for the definition of an exchange package for patient record exchange2,3. A user can define the mark-up elements and their sequential and nested relationships according to the semantics of the specific domain. In particular, the user can dynamically reconstruct the exchange package according to differing circumstances by redefining, or defining extensions to, the DTD. The Web tools interpreting XML automatically adapt to the changes in the exchange package. Also it is possible for the exchange package to be marked with any special processing instructions; these instructions would not be executed by the web tool but passed to the application sitting on the web tool20.

For illustrative purposes, we show in the Appendix an example marked up record based on a simplified DTD. The data are fabricated. The illustration was based on a fragment from the working papers for the Hypermedata project2,3,4. The example shows how the exchange package can be tuned fairly easily to meet the specific needs of the target site.

Preliminary work on an Immediate Discharge Document8 and Electronic Health Care Record Request9 is being carried out in Scotland.

An important feature of XML is that standard libraries of tags can be published on the WWW. XML can include these by using a URL reference. This allows for the sharing and informal standardisation of exchange data schemas.

Automatic transfer of record information

XML enables channels to be defined at web sites; these channels are combined with scheduling facilities21. This means that a central medical centre could act as a "publisher" and automatically notify another centre (playing the role of "subscriber") of any relevant changes to patient records (for instance, a new biopsy report). This notification could be controlled in different ways. With pull technology the receiving site would poll publishing sites and receive records on demand. Alternatively the publishing site could enable push technology and notify its subscribers, either by email or, when the appropriate standards have been defined, by pushing the updated record20.

Limitations of XML for record transfer

A key feature of XML is that it does not support data typing in the language; data are assumed to be readable character data or pointed to as a resource. This leads to inelegant expansions such as dates being expanded to separate strings representing days, months, years. The situation is even more unsatisfactory for non-character types such as sound, images, scans, and video. (However, we can note that XML, through its SMIL application22, does provide a way of synchronising the rendering of disparate and distributed multimedia resources, which could be relevant to some kinds of medical documents.)

The precise mechanism that XML will use to add data typing and other semantics is still under discussion in W3C. Both RDF (Resource Description Framework)7 and XML Schemas23 have their proponents. However, the main outlines are already clear: "vanilla" XML will remain untyped, and metadata, such as typing, will be layered on top. This semantic layer might initially be rather limited in its expressiveness but it will have several distinct advantages:

  • it will be more powerful than the DTD mechanism and able to express a richer set of constraints;
  • it will itself be an XML application, and therefore accessible to the same processors as XML (this alone will be a major advance on DTDs);
  • it will be inherently distributed and reusable;
  • authors will be able to choose whether to embed the metadata in a document or to reference it by a URL (Uniform Resource Locator), thus creating the possibility of documents with relatively rich, shared semantics for only a modest increase in the size of the transmitted file.

The ease of use of XML will encourage the danger of it being introduced in an arbitrary way by local centres to exchange records. The result could be uncontrolled and unreliable datasets (similar to the way spreadsheets have become mistrusted for reliability and accuracy) unless the lessons from database research are taken on board. Database experience has much to contribute to the development of complex documents, in particular:

  • the understanding of data integrity through the specification and enforcement of constraints;
  • the achievement of data consistency through the development of data models and the clear separation of schema (extension) from data (intension);
  • the provision of standard interfaces through database tools and management services.

The underlying data model for an XML exchange package has a hierarchical structuring of relationships. Hogan24 identified a correspondence between the hierarchical model and the hierarchies representable by the object model in object databases. However we anticipate that the Object-Relational approach is more likely to dominate in practice over a pure Object Oriented one25.

One of the conclusions of the Hypermedata project was that a good ontology is needed for combining different record exchange packages schemas in a semantics-preserving way. One early use of XML has been in hyperglossaries, notably in the biochemistry domain26. This kind of activity could provide a starting point for the development of the much richer ontology that the Hypermedata approach expects. It also illustrates convincingly that XML is well-suited to this kind of structured, cross-referenced, machine-processed combination of data and metadata.


UN/EDIFACT as an Exchange Procedure

Origin of UN/EDIFACT

EDI dates back to the 1970s, when it was introduced by the Transportation Data Co-ordinating Committee; it was based upon fixed transaction sets24. UN/EDIFACT arose in the early days of the electronic exchange of data and represented the response to differing standards in North America and in Europe. In 1986 the United Nations Economic Commission for Europe, UN/ECE, approved the simple concept of a single international standard that was flexible enough to meet the needs of government and industry.

EDI Experience

The claimed benefits of EDI in a business context5 were the reduction or elimination of:

  • repeated keying of information and associated error correction;
  • manual reconciliation of different documents;
  • document mailing and telephoning of information.

The procedure was based upon dedicated proprietary software sometimes provided by a third party. An example is the exchange of patient examination information between hospital departments and a GP’s surgery using Medilink’s Postroom and GPViewer. There were common standards, but different proprietary systems generate different outputs. One of the weaknesses of EDI was that it was quite invisible and only worked well when it was application-to-application. The result was that dummy applications had to be created to view or create the contents of an EDI message10.

European EDI Initiative

The Commission of the European Communities launched the TEDIS (Trade EDI Systems) Programme in 1990. TEDIS aimed to promote the use of electronic data exchange for the automatic exchange of structured documents. The research consortium included researchers from Italy, Denmark, Germany, Spain, France, The Netherlands and Switzerland. As part of that programme there was to be a survey of the adoption and implementation of EDI in Europe. It was found that the diffusion of EDI in Europe was slower than anticipated27. Similarly experience in North America shows EDI to have flagged because it remains too expensive and the software developed does not make it easy to leverage implementations across different trading partners28.

Evaluation of EDI within a Regional Health Care System

An evaluation has already been reported of the use of EDIFACT for a Regional Health Care Network, RHCNET, in The Netherlands29. The following EDI medical messages were applied in RHCNET:

  • radiological analysis
  • laboratory analysis for clinical-chemical and haematological laboratories
  • patient admission, discharge, and movement between hospitals
  • diagnostic or treatment report of medical specialists
  • survey of patients admitted by a general practitioner

The key benefit was identified as data quality rather than cost reduction or increased speed of communication. Indeed for a general practitioner, participation cost rather than saved money, but this was outweighed by the advantages of better data quality. It was important for the added value of EDI that the message transfer was automatically connected to the internal information systems by message transfer modules on both sides of the message exchange to avoid transcription errors. There was no impact of EDI on the working methods of the hospitals and very little on those of the general practitioners. The weakest link in the RHCNET EDI implementation was seen to be the general practitioners’ transformation modules, both in terms of availability and reliability. Mono-directional transfer from hospitals to general practitioners but not the other way round was seen as an important barrier.


Comparison of approaches

Introduction

In this comparison we will distinguish between UN/EDIFACT27 and the proposals for XML based EDI/EDIFACT10,28 by using the term XML/EDI for the latter. For the Hypermedata approach, use of net protocols was an explicit objective, while the deployment of the proprietary graphical language was made necessary by the need to structure and reason about the record exchange package.

Adaptability to change

Adoption of an XML defined mark-up of a medical record exchange means that the exchange package itself carries the structure and content of the exchange explicitly. This implies that the receiving site can interpret and reorganise the information according to local needs. UN/EDIFACT expects direct write capability on the receiving site. That is, there needs to be a tight, trusted coupling between the different centres. This tight coupling results in a centralised system that is very difficult to change.

Software provider dependence

The integration of systems via the UN/EDIFACT approach requires customised software that is maintained by the initial network vendor. The use of proprietary software binds a network into vendors, their development policies and pricing strategies. XML-based solutions can be independent of proprietary software.

Cost of basic software

UN/EDIFACT networks are constructed as customised products by vendors who are geared to the recovery of the initial outlay through pricing tariffs. The cost benefits of the UN/EDIFACT network approach tend to accrue mainly to the central sites while the distributed sites incur extra costs. Revision of the agreed set of transactions requires customised rewriting of sophisticated software. The original Hypermedata employed its own proprietary graphical mark-up language; this was intended to be licensed to users of the architecture. The basic XML capability will be available cheaply or freely on the standard PC desktop tools. This means the entry cost could be low for a new partner using simple load and unload facilities. The decoupling of the exchange package technology from the local medical information systems through XML tools means that a major element of risk and cost is removed from the pricing equation.

Underlying ontology

Each of the exchange mechanisms assumes consistent semantics by reference to an ontology. With UN/EDIFACT it is assumed that there is a direct matching of ontologies between different sites. The XML approaches allow a negotiated matching of semantics: if the medical record exchange package is received using a transition file then the local system could be adapted retrospectively to deal with variation of the data model. There is a danger with XML that the ease and flexibility of handling record exchanges could lead to the construction of semantically inconsistent datasets when these are used for purposes other than what determined the original transfer.

It is essential to remember that semantic reconciliation needs to be negotiated and none of the exchange mechanisms automates those negotiations. The use of attributes on tags can facilitate the negotiation by making some of the semantics more explicit (for example, default units). This can be taken further by including or referencing metadata to provide more semantic hints about the correct interpretation of tags. RDF and XML Schemas provide alternative approaches to this issue. However identifying the correspondence between a person in one system and the same person in another system requires real world confirmatory evidence. Automatic negotiation was one of the unfulfilled aims of Hypermedata.

Security during transmission

There are now cheaply available public key encryption facilities for the web (such as Pretty Good Privacy). These can be used to maintain the confidentiality of any medical information that is transmitted. This will enable medical centres handling one-off emergencies to receive in confidence medical records over the internet and interpret them satisfactorily.


Conclusions

The slow take-up of EDI by the business community28 in both North America and Europe arose from:

  • the expense of the underlying proprietary software and network needed to support it;
  • the rate of change of business processes making adaptation of EDI networks difficult;
  • the reluctance of small businesses to adopt different EDI systems for different partnerships.

XML will emerge as a key enabling tool for UN/EDIFACT. This will enable us to circumvent the problems associated with global schemas in tightly coupled systems that permit updates. These systems degrade and become unworkable because of the difficulty in setting up consistent global schemas for non-centralised organisations and their inflexibility to system change. The use of tag repositories based on UN/EDIFACT10 and Common Business Libraries28 could expedite the adoption of the XML approach to EDI. Taking a web based approach overall allows medical centres to avoid being locked into a particular vendor of proprietary systems.

Record exchange using web-based technologies, especially XML, will enable medical centres to:

  • benefit from what will be almost universal cheap or freely available web tools;
  • adapt gracefully to development of new medical procedures, change in legislation, change in the organisational structure, or change in the business practice;
  • negotiate the content of their medical record exchange according to the specific medical needs at the time;
  • respond gracefully to system upgrades;
  • allow local users to tune their own medical systems to their local medical and business needs;
  • have a high level of security because of the encryption facilities available.

It is crucial that the dangers arising from unreconciled semantics are recognised and addressed. The ease of exchange of medical records that will be enabled by XML could lead to substantial errors if not subject to clearly identified good practice guidelines.


References

1 Akass C. The XML-files. Personal Computer World, 1998 (November):40
2 Dixon M, Kohoutkov J, Cook SC, Jeffery KG, Read BJ. Exchanging Medical Records: the Hypermedata Solution. Journal of Informatics in Primary Care 1998 (November):8–12
3 Dixon M, Kohoutkov J, Cook SC, Jeffery KG, Read BJ. Managing Heterogeneity in Medical Information Systems. Proceedings of the 10th ERCIM Database Research Group Workshop on Heterogeneous Information Management, Prague, Czech Republic, 1996; ERCIM-96-W003:7
4 Cook S, Kohoutkov J, Jeffery KG. Hypermedata: Meta-structures for Exchanging Hypermedia Documents. Proceedings of 9th ERCIM Database Research Group Workshop on Multimedia Database Systems (ERCIM-96-W001 GMD), Darmstadt, Germany, 18-19 March 1996:135–144
5 O’Callaghan R, Turner JA. Electronic Data Interchange – Concepts and Issues. In: Krcmar H, Bjorn-Andersen N, O‘Callaghan R. EDI in Europe: How it Works in Practice. Wiley Series in Information Systems, 1995, Copyright 1995 European Communities pp1–19
6 World Wide Web Consortium. Extensible Mark-up Language (XML). http://www.w3.org/XML/
7 World Wide Web Consortium. Resource Description Framework (RDF). http://www.w3.org/RDF/
8 http://scotland-xml.uk.eu.org/
9 http://scotland-xml.uk.eu.org/, CEN TC251/PT29/T-039
10 Raman D. Proposal for a UN Repository for XML Tags Based upon UN/EDIFACT. European Electronic Messaging Association Working Document TRADE/CEFACT/1998/CRP.25, 31 August 1998
11 http://www.dci.clrc.ac.uk/Activity/W3C-LA
12 Jeffery KG (Private Communication). Inaugural Meeting of National Electronic Library for Health and Eclinik Project, 2 December 1998. See also http://libsun1.jr2.ox.ac.uk/ocihi/ukihi.html for a general pointer to activity in the area.
13 Bright M, Hurson A, Pakzad S. A Taxonomy and Current Issues in Multi-database Systems. IEEE Computer 1992 (March):50–59
14 Ahmed MA, Berlin A. Information Technology in General Practice: current use and views on future development. Journal of Informatics in Primary Care 1997(November):5–8
15 Heiler S. Semantic Interoperability. ACM Computing Surveys 1995; 27(2):271
16 Hernandez M, Stolfo SJ. Real World Data is Dirty: Data Cleansing and the Merge/Purge Problem. Data Mining and Knowledge Discovery 1998; 2:1–31
17 Colomb RM, Orlowska ME. Interoperability in Information Systems. Information Systems 1994; 5(1):37–50
18 World Wide Web Consortium. Mobile Access. http://www.w3 .org/Mobile/
19 World Wide Web Consortium. Web Content Accessibility Guidelines 1.0. http://www.w3.org/TR/1999/WAI-WEBCONTENT-19990505/
20 Harold ER. Structuring Complex Content for the Web: XML – Extensible Mark-up Language. IDG Books, 1998
21 Azeez LO. The XML Contribution to Internet Push Technology. MSc Project in Information Systems Development, London Guildhall University, 1998
22 World Wide Web Consortium. Synchronised Multimedia. http://www.w3.org/AudioVideo/
23 World Wide Web Consortium. XML Schema Part 1: Structures. http://www.w3.org/1999/05/06-xmlschema-1/; and XML Schema Part 2: Datatypes. http://www.w3.org/1999/05/06-xmlschema-2/
24 Hogan M. XML and the Internet: Driving the Future of EDI, Poet Software – XML EDI White Paper, February 1998
25 Stonebraker M, Moore D. Object-Relational Database Management Systems: Tracking the Next Great Wave. Academic Press/Morgan Kaufmann, 1998
26 Murray-Rust P, West L. Knowledge, Language and Semantics: XML and VHG. ASLIB Managing Information 1998; 5(4):34–36
27 Krcmar H, Bjorn-Andersen N, O‘Callaghan R. EDI in Europe: How it Works in Practice. Wiley Series in Information Systems, 1995, Copyright 1995 European Communities
28 Meltzer B, Glushko R. XML and Electronic Commerce: Enabling the Network Economy, SIGMOD Record 1998; 27(4)
29 Ribbers PM. EDI in Regional Health Care – RHCNET in The Netherlands. In: Krcmar H, Bjorn-Andersen N, O‘Callaghan R. EDI in Europe: How it Works in Practice. Wiley Series in Information Systems, 1995, Copyright 1995 European Communities pp259-276

Appendix

Example of a Record Fragment Marked-up according to the DTD based on part of the Hypermedata Class Structure

<discharge-notes>      
<hospital> Royal Reading Hospital </hospital>    
<patient>        
  <p-name>      
    <p-surname>  Richterova </p-surname>
    <p-forename>  Alena </p-forename>
    <p-maiden>  Zelniek </p-maiden>
  </p-name>      
  <p-address Type="CURRENT">    
    <p-address-number>  12 </p-address-number>
    <p-address-street> Gila Road </p-address-street>
    <!-- note that p-address-district is not mandatory-->
    <p-address-city>  Chandler </p-address-city>
    <p-address-postcode>  RG32 1RJ </p-address-postcode>
  </p-address>      
  <p-dob>      
    <p-dob-day> 19 </p-dob-day>
    <p-dob-month> June </p-dob-month>
    <p-dob-year> 1987 </p-dob-year>
  </p-dob>      
  <p-id> K99/150 </p-id>    
</patient>        
<prescription>        
  <prescription_item>      
    <item> Becotide </item>
    <quantity> 200 </quantity>
    <units> micrograms </units>
    <frequency> 4 times daily </frequency>
    <notes> Self administered by patient using rotahaler </notes>
  </prescription_item>      
  <prescription_item>      
    <item> Serevent </item>
    <quantity> 50 </quantity>
    <units> micrograms </units>
    <frequency> twice daily </frequency>
    <!-- there are no notes recorded-->  
  </prescription_item>      
</prescription>        
</discharge-notes>      

Notes

1 The corresponding Document Type Definition includes declarations for:

  • optional elements, such as <p-maiden>;
  • element values restricted to a specified list, such as for <p-address Type>;
  • default values, such as "per Day" for <frequency>.

2 The absence of data typing means that dates, such as <p-dob> are expanded into character elements.

Refbacks

  • There are currently no refbacks.


This is an open access journal, which means that all content is freely available without charge to the user or their institution. Users are allowed to read, download, copy, distribute, print, search, or link to the full texts of the articles in this journal starting from Volume 21 without asking prior permission from the publisher or the author. This is in accordance with the BOAI definition of open accessFor permission regarding papers published in previous volumes, please contact us.

Privacy statement: The names and email addresses entered in this journal site will be used exclusively for the stated purposes of this journal and will not be made available for any other purpose or to any other party.

Online ISSN 2058-4563 - Print ISSN 2058-4555. Published by BCS, The Chartered Institute for IT