| DrawML and SVG | Table of contents | Indexes | Information and Content Exchange | |||
Electronic health care record transfer |
| Andrew J Hinchley |
| Managing Consultant |
| Communications Planning
49 Abbeygate Street Bury St Edmunds Suffolk United Kingdom IP33 1LB Phone: +44 1284 750658 Fax: +44 1284 750620 Email: ahinchley@communicationsplanning.ltd.uk Web: www.communicationsplanning.ltd.uk/ |
Biographical notice: |
Markwell, David C Reading ![]() The Clinical Information Consultancy United Kingdom ![]() |
Andrew Hinchley runs his own consultancy company, specialising in communications, particularly in healthcare. He has played a very active role in the developments of methods for specification of clinical messages. The message design methodology developed in the European healthcare standards group CEN TC251 offers for the first time the ability to provide detailed conformance and quality checks on clinical message implementation. He has been actively reviewing XML for this and other healthcare applications over the last 18 months |
| David C Markwell |
| Principal Consultant |
| The Clinical Information Consultancy
93 Wantage Road Reading Berkshire United Kingdom RG30 2SN Phone: +44 118 9584954 Fax: +44 118 9504464 Email: david@clinical-info.co.uk Web: www.clinical-info.co.uk |
Biographical notice: |
Messages to transfer the electronic patient record |
|
Provide EHCR Message >
|
|||||||||||||
The diagram below illustrates a shared care model, where the Provide EHCR is one of several messages needed to support the flows shown. |
|
Shared Care Model >
|
|||||||||||||
Decision to Use XML |
There are some specific conclusions so far: |
Proposals that a "cut-down" XML should be used for EDI messaging are however believed to be mis-guided, since experience has shown that access to all the standard tools has been invaluable and creating a separate different specifications for messaging would be unfortunate and against the spirit of the W3C recommendations. |
Work on XML for messaging in XML/EDI and TC251 |
In Europe, two of the major thrusts on investigating use of XML for messaging have been the CEN ISSS Electronic Commerce workshop and the CEN healthcare standards group already mentioned. |
CEN TC251 started looking at these issues at the end of 1997 and in mid-1998 a technical report on the use of XML was published www.centc251.org/ . This study resulted in the setting up of a task force under Professor Joachim Dudeck. In autumn of 1998, a first draft DTD was produced, corresponding to the EPR transfer message standard www.communicationsplanning.ltd.uk/ahinchley/ |
The XML/EDI work in Europe has been co-ordinated by the CEN ISSS Electronic Commerce workshop http://www.cenorm.be/isss/workshop/ec/xmledi/isss-xml.html |
XML/EDI and XMLEPR projects |
The European XML/EDI project established at the beginning of 1999 is sponsoring message development for a number of sectors and the collective result of this experience is part of the overall XML/EDI message development methodology. The objective of the European project is to feed back to the standards arena from the practical work carried out in 1999. This work is supported by the European Commission under the DGIII ISIS Programme. XMLEPR is part of the European XML/EDI project. |
Work in these projects is based on substantial experience in design and implementation of structured messages. This has shown that the data structures need to be fully defined and shared between the co-operating systems if the information is to be as fully available in the receiving system as it was in the sending system. |
The logical conclusion from this in XML terms is to use valid and well formed DTDs, requiring that XML messages correspond to the DTD in full, i.e. no undefined tags. Means to extend these standard DTDs to meet local requirements remains of course a requirement |
Automated mapping from UML model/object database |
One of the great steps forward resulting from this work is the ability to automatically generate XML DTDs from the UML models and object databases used to describe the data structure to be transferred - a generalised patient record structure. The tools used are specific for the project, but the conclusions are general. The XML mapping rules defined below are a set which can potentially, support any of the healthcare messages that have been defined by TC251 over the last 5 years. |
Refinement of specific mapping rules |
A set of about 20 mapping rules have been found adequate to automatically produce an XML DTD corresponding to the message represented as a UML model with supporting objects/attributes. |
The following 6 rules listed below show the general approach of these rules. A further 12 rules are needed to pick up particular cases arising out of the models used. |
|
Clinical validation of the XML-based EPR transfer message |
Experience, particularly in the UK, has shown that clinical validation is a challenging, but essential part of the mass introduction of new electronic methods into every day patient care. Some of the major benefits of using XML are expected to come from achieving a better quality of validation. |
The challenges come from devising a method of validation that:- |
|
The process of clinical validation involves the production of XML messages (each being a patient record extracts) by exporting several patient records from a variety of clinical information systems. The information is exported in the form of an XML document that complies to the DTD for the Provide EHCR message. These messages are then viewed through one or more of a selection of the available XML browsers or viewers,or a specially developed viewer. The available viewers used have included freely distributed products such as Microsoft's XML Notepad and Internet Explo |
One of the main software deliverables of the project has been a special XML viewer developed. The viewer is designed to ensure that those parts of the message of most interest to clinicians can be displayed in a way which directly supports the evaluation process. This viewing activity is purely a function of validation. Once implemented, these process-process messages do not require to be displayed. |
|
XML Viewer >
|
|||||||||||
Clinician compare views of the record within the original application with the content and structure of the message. Comments on discrepancies or inadequacy can then be added (using the project specific viewer) and these are fed back for analysis. |
Particular features of the specially developed XML viewer are:- |
|
At the time of writing this paper, solid results are already becoming available from the project. Of the 4 suppliers of GP computer systems involved, two have been able to extract fairly comprehensive structured records from their own proprietary record systems. The companies involved, EMIS and In-Practice Systems have achieved excellent results within the tight time constraints that the XMLEPR project has set itself. |
Issues-looking to the future |
The key next stage once the work described is complete will be to evaluate the receipt of EPR transfer messages and assimilation of the information to the patient record at the receiving system. These challenges are related to the different database structures used indifferent systems and the different way in which clinical aspects of the patient record have been recorded. It is critical that the EPR message structure and its XML representation are powerful enough to achieve the desired results. |
Once at that point, there are then a number of further XML issues that will need to have been solved in order to move to real implementations in day-to-day healthcare. |
Validation |
Sending structured information from one legal entity to another is problematic if the receiver is to take full responsibility for the semantics of what has been sent, without full scrutiny by an expert of every last tagged item. |
Implementing satisfactory validation has however been a huge problem in messaging over the last decade. This is for a number of reasons:- |
A first requirement on any solution is to be able to fully automate whatever checking mechanisms are to be put in place. |
In a XML context, there are a number of ways in which we can seek to meet these requirements:- |
|
It will be part of the European XML/EDI project to investigate these alternatives once the requirement is better understood. |
Conclusion |
The work achieved so far is promising in providing high quality XML-based solutions at a much lower effort than with previous messaging methods. |
It will be important to inject into the standards activity, enough of the results for this work to accelerate take-up, but not over limiting the way that solutions are implemented.This is the mandate of the European XML/EDI project. |
On-going work on many fronts in XML development will be promising if additional facilities can be harnessed in the way that XML 1.0 has been used as described in this paper, but negative for messaging applications, if it leads to too much diversity in facilities used and hence inoperable systems. |
Acknowledgments |
The authors wish to express their thanks to the European Commission DGIII Programme, the UK National Health Service and the UK Royal College of General Practitioners |
|
Bibliography
|
| DrawML and SVG | Table of contents | Indexes | Information and Content Exchange | |||