Plenaries   Table of contents   Indexes   Why Your Document Management System Should Care About Hyperlinks
 
 

SGML/XML in healthcare information exchange standards
 
Liora   Alschuler
  consultant
  The Word Electric
Route 5 & Sanborn Rd.
POBox 155
East Thetford   Vermont  05043  USA
Phone: 802/785-2623
Email: liora@the-word-electric.com
 
Biographical notice:
 
Liora Alschuler
 
Liora Alschuler is a consultant and writer specializing in the application of SGML  (Standard Generalized Markup Language) and XML  (eXtensible Markup Language) to healthcare information systems. She is the author of ABCD... SGML: A User's Guide to Structured Information, International Thomson Computer Press, 1995, a Contributing Editor for the Seybold Report on Internet Publishing and Senior Analyst at xml.com.
 Brennan, Sean 
 England  
 Leeds  
National Health Service
 
Liora is active in the HL7 SGML SIG which is developing standards for the application of SGML to healthcare information exchange. She was the Project Manager for Operation Jumpstart which created the Kona Architecture, a proposal for scaleable exchange of SGML-encoded clinical records and she is Chair of the Kona Editorial Group, recently chartered by theSIG to create a ballotable draft of the specification. She was Program Chair of the HL7 SGML Mixer, held in August, 1997, and worked with eight vendors to create a prototype application for SGML and XML Claims Attachment Processing presented at that meeting. She has spoken on hypertext and SGML at local, regional, national, and international conferences. She is a member of the W3C XML Special Interest Group and of OASIS (formerly SGML Open).
 
Sean   Brennan
  Electronic Patient Record Project Manager
  National Health Service
Room 1N 35C
National Health Executive Department of Health
Quarry House
Quarry Hill
Leeds   LS2 7UE  England
Phone: +44 113 254 6247
Fax: +44 113 254 6278
Email: epr@doh.gov.uk
 
Biographical notice:
 
Sean Brennan
 
Originally trained in Pathology specialising in Immunology, as part of his general management training, Sean was seconded to establish a Medical Audit support unit at a Trust where he Information Manager at Huddersfield being responsible for the Resource Management, Information and IM&T staff in addition to those in the newly created Medical Audit Unit.
 
Sean was seconded (part-time) to the NHS Executive as Clinical Audit Advisor. His residual duties at the Trust included managing the pilot project for the development of computerised links between the Trust and four GP surgeries, resulting in the exchange of Casualty, Admission, Discharge and Outpatient letters, together with Pathology and X-Ray reports.
 
Sean became full-time Clinical Audit advisor to the NHS Executive in 1995 and more recently Project Manager for the Electronic Patient Record Project.
Istituto Tecnologie Biomediche, Consiglio Nazionale delle Ricerche
 Italy 
 Rome 
 Rossi Mori, Angelo 
 
The project has concluded that the main objective of EPR is NOT simply to replicate the paper record on computer but to generate an EPR as a by-product of supporting the clinical process with IT where appropriate. This approach will require traditional approaches to IT alongside the opportunities offered by the ‘emerging technologies’ including the use of browsers, mark-up languages etc.
 
Angelo   Rossi Mori
  Head of the Unit of Medical Informatics, Reparto Informatica Medica
  Istituto Tecnologie Biomediche, Consiglio Nazionale delle Ricerche
viale Marx 15
Rome   I-00137  Italy
Phone: + 39 6 - 827 71 01
Email: rossi@color.irmkant.rm.cnr.it Web: http://win.irmkant.rm.cnr.it/termhome.htm
 
Biographical notice:
 
Angelo Rossi Mori
 
Angelo Rossi Mori is Project Leader for the European standard CEN ENV 12265 "Medical Informatics--Categorical structure of system of concepts--Model of semantics and co-chair of CEN/TC251/WGII  (Working Group II on "Terminology, semantics and knowledge bases" (in CEN/TC251)) on Medical Informatics--Terminology and knowledge bases. He is working on a systematic compositional approach to medical phrases, based on predefined semantic sturctures (ontologies). He is developin principles, methods, and tools for extraction, integration, maintenance, and usage of high-level schemata for healthcare terminologies and continuity of representation from medical record to clinical content in SGML .
 Arlington 
 Sokolowski, Rachael 
 USA  
iTRUST
 
Rachael   Sokolowski
  Chief Scientist and Vice President of Research
  iTRUST
8 Central St
Arlington   Massachusetts  02174  USA
Phone: 781 646 8877
Email: rsokolowski@itrust.net
 
Biographical notice:
 
Rachael Sokolowski
 Dudeck, Joachim 
 Germany 
 Giessen 
Institut fuer Medizinische Informatik, Justus-Liebig-Universitaet Giessen
 
Rachael Sokolowski is Chief Scientist and Vice President of Research for iTRUST. Prior to joining iTRUST, she was with Kurzweil Applied Intelligence, where she was Principal Investigator and System Architect on two multi-million dollar research grants from the National Institute of Standards and Technology (NIST), developing a voice-enabled medical reporting system and an electronic medical transcription system. Prior to Kurzweil, she was Lead Engineer at Houghton Mifflin, where she helped develop the CorrecText grammar and spell checker used in Microsoft Word and electronic representations of dictionaries, almanacs and college textbooks in SGML . She is co-chair of the HL7-SGML/XML Special Interest Group. She holds a B.A. in mathematics from Smith College, and is pursuing her Master's degree in Public Health from Boston University.
 
Joachim   Dudeck
  Professor Doctor
  Institut fuer Medizinische Informatik, Justus-Liebig-Universitaet Giessen
Heinrich-Buff-Ring 44
Giessen   35392  Germany
Phone: 49.641.994.1350
Email: Joachim.W.Dudeck@informatik.med.uni-giessen.de
 
ABSTRACT:
 
This panel presents recent work usingSGML andXML as the basis for medical information exchange standards. Alschuler and Sokolowski will present the Kona Architecture, a project of the Health Level 7 SGML / XML SIG , which uses architectural forms to express the relationship between exchange DTDs of varying levels of granularity and specificity and between DTDs modeled for different clinical specialties including encounter notes and DICOM-conformant image reporting. Rossi Mori will report on use of SGML to meet stringent requirements for flexibility, non-ambiguity, and coherence in healthcare terminologies within a standards context. His approach uses a systematic compositional approach to medical phrases, based on predefined semantic structures (ontologies). Dudeck will report on the application of SGML / XML to transactional messaging including the resolution of ambiguities in current messaging standards. In addition, we will have a special report on the SGML clinical record project directed by the British National Health Service from Sean Brennan.
 
 

The HL7 SGML/XML SIG
 
Health Level 7 ( HL7  (Health Level 7) ) is a health-care specific, application level protocol which has been developed by an ad hoc organization consisting of hospital, vendor and consultant volunteers. Initial efforts began in March of 1987 and continue to date The current version addresses interfaces for admissions, discharges, and transfers (ADT), clinical instructions, lab results, patient observation records, charges and other billing data. Of the major healthcare data exchange standards, HL7 has the widest scope and is the most widely used data exchange standard. HL7 has emerged as a major factor in standards for clinical information. Although it primarily has been focused on intra-provider (especially hospitals) communications, the standards can be modified to work in Electronic Data Interchange (EDI) for inter-entity exchange.
 
The HL7 SGML / XML SIG  (Special Interest Group) is a Special Interest Group of HL7 . The SIG formed to create a national standard for the use of SGML in all domains of health care. The SIG is interested in working within the existing health care standards community and one of its objectives is to develop and coordinate Document Type Definitions (DTDs) for health care. The HL7 SIG has three areas of work: SGML / XML representations of HL7 messages, a proposal for representations called the Kona Proposal and education and training.
 
The Patient Identification (PID) segments is used in all applications as the primary means of communicating patient identification information. Part of the focus of the messaging group of the SIG is to research replacing the HL& syntax with SGML / XML representations. Data elements combined with the syntax of an HL7 message constitute HL7 message suitable for transmission. Below is a sample HL7 message for patient information and a possible XML representation.
 
 

The Document-Centered Approach to Clinical Information
 
Last year in Barcelona, a panel of member of the HL7 SGML / XML SIG reported on the work done to introduce SGML / XML as an alternate syntax for transmission of messages within a clinical context. We also had a preliminary report from the SGML clinical records project of the British National Health Service. This year, we have a followup report from the NHS project in Oswestry and we will have up-to-date reports on the work of the SIG and related standards efforts underway within CEN  (European Committee for Standardisation) , ISO  (International Standard Organisation) , and ASTM (American Society for Testing and Materials). This year we will also report on the development of the Kona architecture for exchange of clnical documents.
 
The Kona architecture represents a document-centered approach to the exchange of clinical information. (Document is used here in contrast to messaging and to records whose canonical form is database storage.) This architecture promotes exchange of structured information while building in flexibility in several key modalities. First, the architecture used three axis of derivation according to level of granularity, medical domain, and localization. Document types created under business rules for a local specialty practice are mapped declaratively, using ISO 10744 (HyTime) architectural forms to the closest Kona meta-DTD. When documents validated against the local DTD are sent to a Kona-compliant repository or recipient, standard architectural forms processors, such as the popular SP and JADE from James Clark, automatically transform the locally-variant document into a compliant document, retaining the source identification and, in a predictable fashion, a high degree of the local semantic markup. We will demonstrate this type of exchange processing and report on the status of current implementations, as well as the status of the standardization effort.
 
 

Opportunities for XML in healthcare
 
 

The view from Italy
 
(This section on the general outlook for XML in healthcare was contributed by Angelo Rossi Mori with the assistance of Fabrizio Consorti.)
 
We feel that a dramatic revolution will come from the active management of clinical information.
 
Some major obstacles to management and transmission of clinical information are: variable nature, extreme flexibility, unpredictable levels of granularity, purposive views (to face re-use of information according to the various tasks performed by the same or other professionals).
 
Key features of XML on this respect are:
  • the original text can be combined with structured tags, allowing the simultaneous processing of the same document by humans and by computer;
  • pre-coordinate tags can co-exist with elementary tags, allowing to browse and reorganise data with arbitrary level of granularity and multiple views;
  • a system of tags can be transmitted as a DTD (and standardised in meta-DTDs), so that users can effectively use general-purpose browsers and applications;
  • information can be extracted from a document, preserving and purposively rearranging the structure of the original document. The resulting information element can be then included into another explicit structure. Automatic transformations can be performed among coherent DTDs, e.g. for the construction of flexible messages from structured documents (with a unified representation mechanism);
  • documents have the same nature than messages, i.e. we can send documents as they are, or we can have a collection of documents and messages processed by the same browser. The same software environment combines tasks of information retrieval, user interfaces and communication;
  • a collection of documents and messages is still a document, i.e. documents can be nested within documents, allowing the user to select for each task the most appropriate level at which the corresponding document could be considered as the "operative context" for the data (see below).
    •  
    XML is not the magic panacea for all potential issues in the clinical field. Nevertheless, it appears to be a good vehicle to introduce advanced features in the patient record, specially if combined with a powerful representation of semantics (i.e. a semantically based mark-up language).
     
    We envisage that XML can eventually provide a crucial added value: the support for "interoperability" of clinical applications [1, 33]. In fact, the patient record is basically a collection of documents, not a set of records in a database. Therefore the metaphor based on documents (as in XML ) is very near to the original nature of the patient record.
     
    Furthermore, if a collection of documents is still a document, then sections of a document can be considered themselves as documents. The metaphor of nested documents can cover every discrete amount of information, from the most simple record item value to the most complete collection of data about all the encounters of a patient. By using advanced database techniques on a collection of XML documents, we can effectively manage user's views and retrieve information in any set of XML -structured patient records.
     
    Let us discuss some opportunities for XML in the immediate future, based on the above principles.
     
    We assume that plug-ins for Web browsers, XML -oriented database systems, and developer's toolkits will be widely available soon at reasonable prices, and that security problems will be appropriately solved.
     
     

    Represent and transmit the organization of a healthcare record
     
    A fixed, all-purpose set of headings for patient record (labels for high-level sections) is hardly conceivable. In fact, the organisation of the patient record and the corresponding headings depend on various factors:
     
    • different views on the same data according to tasks.
    • the needs of different healthcare professionals in various environments (e.g. clinicians vs. nurses in hospitals, general practitioners);
    • local and National regulations and customs;
    • different views on the same data according to tasks.
    • different views on the same data according to tasks.
     
    We define here a "structural context" for a record element as the context - within the structure of a given record system - actually determined by the labels of higher rank in the chain nested above that element. For example, elements in a section involving "history of ..." have a different context than the ones under a heading involving a "plan". And context is different if a heading modifies the subject of information from "patient" (the default) to a "relative" (e.g. in "family history") or to a "foetus" (e.g. in echographic data). Interpretation of data heavily depends on the structural context.
     
    The "Domain Termlist" standard, under development by CEN/TC251  (Technical Committee for Health Informatics (of CEN)) (WI1.8, part 2), is the opportunity to fix a limited number of tags for the characterisation of record elements, in order to preserve information about the structure of the original record and the context of these elements. Provisional classes of tags are:
    C0 Identification
     
    C0 the tags that identify the nature of data (focus or associated topic): e.g. it is an action, a device, a substance, a condition, an event.
    C1 Safety
     
    basic details (e.g. subject of information, negation, modality, knowing mode, ...) that are crucial for correct management of information, to convey the main context of data.
    C2 Interpretation
     
    relevant details about interpretation of data in the original context by the original user (e.g. roles: diagnosis, side effect, complication, ...).
    C3 Intention
     
    additional details that make explicit the links to activity-related information items (e.g. goals), and thus reveal the intention of the original user. They complete the picture on the context.
    C4 Organisation
     
    further details that show the original organisation of the original record, dealing with terminological knowledge that could be partially derived from the data themselves (abstractions).
     
    Relevance of tags - and thus their need - decreases from class C0 to C4.
     
     

    Encapsulate structured data elements within richer coded entries
     
    The scenario #1 focuses on the "envelop" of data (information about the context). This scenario focuses on the "inside" (the content of an information element). Both are needed to assure a correct interpretation. A coded entry can be defined as a field that contains a code and/or text, ie. terminological data. Within a coded entry, various segments or subfields can be defined, according to a template. In turn, each subfield can be organised again as a template, and so on, with a unpredictable level of nesting (granularity). We should represent various degree of complexity, from a single terminological phrase (ie. simple semantic networks as dissections and canonical forms in a Description Logic, e.g. GRAIL [13, 18]), to a set of sentences in a whole report or discharge letter (e.g. SNOMED-DICOM Microglossary, SDM, for structured imaging report [2, 3, 32]).
     
    The synergy with the compositional approach and the concept-based representation in CEN/TC251/WGII [36] could lead to a significant impact. In fact, the use of CEN standards on categorial structures (according to [7], i.e. templates to mark categories, links, time [12]) could be a base to develop a powerful set of semantic tags. Available categorial structures in CEN are about surgical procedures [11], laboratory properties [10], medical devices [9, 30]. Nursing phenomena [17], vital signs, and imaging procedures, as well as a meta-standard on semantic links, are under development. The CEN standard about identification of medicinal products could be another opportunity.
     
     

    Activity-based patient record
     
    The most attractive application of the above scenarios is in the organisation of patient records by activity. Each activity could correspond to a structured document. A document could contain "sub-documents", with the same structure, related to each sub-activity, and so on. The chain of nested activities therefore corresponds to a chain of nested documents, up to an unpredictable level.
     
    At each level, an activity faces clinical or administrative problems, intends to achieve goals, is performed according to predefined guidelines, and produces clinical and managerial data, that can be the support for audit and quality assurance. These links can be represented explicitly by XLM in a document of any level.
     
    Moreover, we noted that XML can preserve the "structural context" when storing or transmitting elements in a patient record. XML can also manage another kind of context. The "operative context" (or co-text) of a given element can be defined as the set of the other elements produced by the same activity. Given a record element, the peculiar user's purpose determines which level in the chain of the nested documents should be considered to describe its operative context.
     
    Relevance of this scenario is in the rationalisation of healthcare. In fact, it allows to relate data (and costs) to clinical actions and their goals and to use "clinical profiles" of the patients to view and organise the data.
     
     

    Unified access of multiple independent sites in the Web
     
    Many sites on the Web are based on collections of documents with relatively similar structures. With HTML, even if the content was natively structured (e.g. coming from a database), there is no mechanism to record explicitly that structure in the Web site. With XML , the structure can be downloaded in a processable format by a DTD.
     
    Standardisation bodies can facilitate integration of DTDs, so that structures from different sites can be semi-automatically integrated into a unique DTD. Therefore the unified access to collections of structured documents in multiple sites (Intranet or Internet) can be performed. This opportunity is the basis for distributed healthcare record, i.e. for the direct access to healthcare documents about a patient in multiple sites.
     
    Other applications regard access to clinical guidelines, retrieval of similar cases, comparison of statistical indicators. A particularly meaningful one could be in the retrieval of images by description of their content or by features of the procedure (a standard DTD could be derived from the Structured Report supplement of DICOM). Different imaging departments worldwide could make available their collections of images by the Web. The National Library of Medicine (US) could provide the images from the Visible Human project, with multilingual labelling, in a suitable XML format developed by the corresponding G7 project.
     
     

    Report from Oswestry
     
    Since their report at last year's conference in Barcelona, the team in Oswestry has been extremely busy making their vision a working reality and have made major advances in their thinking and approach to EPR and the application of SGML / XML . We received this advance report from Tony Sharer of Graphnet:
     
    Funding from the NHS Executive enabled us to study the problems facing clinicians and investigate how EPR, in SGML / XML format, would be of benefit.
     
    Extensive discussions with clinicians in three Acute Trusts, (John Radcliffe Oxford, United Leeds Teaching Hospitals, now the largest in Europe and Milton Keynes General) gave us a fairly representative sample of clinicians within UK Hospitals..
     
    We also explored, with the Royal Colleges, how the SGML / XML approach could be implemented throughout the NHS and across several clinical disciplines. A strategy paper is about to be published.
     
    As a result of the feedback from these studies we have now developed an interactive browser that allows clinicians to view any part of the record and also if required request/receive lab tests, X Rays etc using the same user interface..
     
    We are also currently evaluating how to make the browser bilingual (the capability to read both SGML and XML ) We are adopting this approach because we believe, at present, both technologies have their place.
     
    Whilst SGML is the established ISO Standard, XML still evolving and relatively immature. There are still elements in SGML which make it more suitable than XML for Patient Records but the market noise about XML means that we cannot ignore it. So for now we'll take both routes.
     
     

    References
     
    • Andover Working Group. Accelerating the Movement Toward Standards-Based Interoperability in Healthcare http://www.dmo.hp.com/mpginf/whitepaper.html
    • Bidgood WD Jr., et al. Controlled Terminology for Clinically-Relevant Indexing and Selective Retrieval of Biomedical Images. International Journal of Digital Libraries. 1997.(Accepted for publication).
    • Bidgood WD Jr., et al. Image-Acquisition Context: Procedure-Description Attributes for Clinically-Relevant Indexing and Selective Retrieval of Biomedical Images. (Submitted to JAMIA). 1997
    • Bob Dolin (ed) SGML as an Interchange Format for HL7 V2.3 Messages. In the Web site of HL7 SGML Special Interest Group, DRAFT 5/1/97, http://www.mcis.duke.edu/standards/HL7/committees/sgml/hl7sgml2.rtf
    • Bob Dolin (ed) SGML as an Interchange Format for HL7 V2.3 Messages. In the Web site of HL7 SGML Special Interest Group, DRAFT 6/1/97, http://www.mcis.duke.edu/standards/HL7/committees/sgml/hl7sgml3.rtf
    • CEN / TC251 "Health Informatics", at http://www.centc251.org
    • CEN ENV 12264:1997. Medical Informatics - Categorial structure of systems of concepts - Model for representation of semantics (MoSe). Brussels, 1997 (prepared by CEN/TC251 Project Team PT003, leader: Rossi-Mori A)
    • CEN ENV 12265: 1995. Medical Informatics - Electronic Healthcare Record Architecture (HCRA). Brussels: CEN, 1995
    • CEN ENV 12611:1996. Medical informatics - Categorial structure of systems of concepts - Medical devices. Brussels: CEN, 1996
    • CEN ENV 1614:1994. Health care informatics - System of concepts for systematic names, classification, and coding for properties, including quantities, in laboratory medicine. Brussels: CEN, 1994
    • CEN ENV 1828:1995. Health care informatics - Structure for classification and coding of surgical procedures. Brussels: CEN, 1995
    • Ceusters W, Steurs F, Zanstra P, Haring E V D and Rogers J (1997). From a time standard for medical informatics to a controlled language for health. Conference on Standardisation in Medical Informatics: Towards international consensus and cooperation, Bermuda.
    • GALEN and GALEN-IN-USE. documentation (1992-97), available from the main contractor AL Rector, Medical Informatics Group, Dept. Computer Science, Univ. Manchester, Manchester M13 9 PL, UK e-mail galen@cs.ac.man.uk ; see also home page at http://www.cs.man.ac.uk/mig/galen and at http://www.cs.man.ac.uk/mig/giu
    • HL7 SGML/XML Special Interest Group. The Kona architecture proposal. http://www.mcis.duke.edu/standards/HL7/committees/sgml/kona.htm
    • HL7 SGML/XML Special Interest Group. Web site http://www.mcis.duke.edu/standards/HL7/committees/sgml/
    • Kirby J and Rector AL. The PEN&PAD Data Entry System: From prototype to practical system. AMIA Fall Symposium, Washington DC, Hanley and Belfus, Inc., 1996
    • Mortensen RA (ed). The International Classification for Nursing Practice ICNP with TELENURSE introduction. Copenhagen: The Danish Institute for Health and Nursing Research, 1996
    • Rector AL, Bechhofer S, Goble C, Horrocks I, Nowlan W and Solomon W. The GRAIL concept modelling language for medical terminology. Artificial Intelligence in Medicine 9: 139-171. 1997
    • Rector AL, Gangemi A, Galeazzi E, Glowinski AJ, Rossi Mori A The GALEN CORE Model Schemata for Anatomy: Towards Re-usable Application-Independent Model of Medical Concepts. in: P. Barahona et al (eds.) Proceedings MIE 94, pp. 229-233
    • Rector AL, Glowinski A, Nowlan W, Rossi Mori A. Medical concept models and medical records: An approach based on GALEN and PEN&PAD. Journal of the American Medical Informatics Association 1995;2(1):19- 35.
    • Rector AL, Rossi Mori A, Consorti F. Practical Development of Re-usable Terminologies: GALEN-IN-USE and the GALEN Organisation Conference on Standardisation in Medical Informatics: Towards international consensus and cooperation, Bermuda.
    • Rodrigues J M, Trombert-Paviot B, Baud R, Wagner J, Rusch P and Meusnier F. Galen-In-Use: An EU Project applied to the development of a new National coding system for surgical procedures: NCAM. Medical Informatics Europe '97, Porto Carras, Greece, IOS Press, 1997
    • Rogers J and Rector AL. The GALEN ontology. In: Medical Informatics Europe (MIE 96), Copenhagen, IOS Press: 174-178. 1996
    • Rossi Mori (ed) Preliminary material for the CEN Project Team on Domain Termlist, WI1.8 dt of CEN/TC251 (available on request at rossi@color.irmkant.rm.cnr.it )
    • Rossi Mori A, Consorti F, Galeazzi E and Merialdo P. A second generation of terminological systems is coming. Medical Informatics Europe '97, Porto Carras, Greece, IOS Press. 1997
    • Rossi Mori A, Consorti F, Galeazzi E. Standards to support development of terminological systems for healthcare telematics. Invited paper at the Working Conference of IMIA-WG6, Jacksonville, FL, Jan 97 (to be published in Meth Inform Med, 1997)
    • Rossi Mori A, Galeazzi E, Consorti F, Bidgood WD Conceptual schemata for terminology: a continuum from headings to values in patient records and messages (suppl. JAMIA, 1997)
    • Rossi Mori A, Galeazzi E, Consorti F. An ontological perspective on surgical procedures. Journal of the American Medical Informatics Society, Symposium Supplement 1996: 115-119.
    • Rossi Mori A, Galeazzi E. Issues about terminology in healthcare, in the Web site of ITBM-CNR at http://win.irmkant.rm.cnr.it/termhome.htm#defin
    • Rossi Mori A. Towards a generalised system of descriptors on medical devices. presented at "Health Telematics 95", Ischia, I, 2-6 July 1995
    • Rossi Mori A., Consorti F. Exploiting the terminological approach from CEN/TC251 and GALEN to support semantic interoperability of healthcare record systems, presented at the IMIA WG16 Working Conference on Standardisation in Medical Informatics: "Towards international consensus and co-operation", Bermuda, Sept 1997 (to be printed in Int.J. Medical Informatics, 1998)
    • SNOMED DICOM Microglossary Information, at http://www.snomed.org/sdm/sdm.htm
    • Sokolowski R (ed). Integration services for clinical data exchange White Paper draft, 1997, in: http://www.omg.org/corbamed/cprcdr.htm
    • Tim Benson. "Application of the Standard Generalized Mark-up Language (SGML) in Electronic Patient Records" prepared by, 9/96, commissioned by The NHS Executive's EPR Project Board, http://www.mcis.duke.edu/standards/HL7/committees/sgml/sgmlepr.rtf
    • Tomelo. (an European Project) see documentation at http://www.ehm.kun.nl/tomelo/WS1/ws1.htm
    • WGII "Terminology and knowledge bases" of CEN/TC251, at http://www.imc.exec.nhs.uk:8000/tc251/wgii
     
     

    Glossary
    ISO
     
    International Standard Organisation
    CEN
     
    European Committee for Standardisation
    CEN/TC251
     
    Technical Committee for Health Informatics (of CEN)
    CEN/TC251/WGII
     
    Working Group II on "Terminology, semantics and knowledge bases" (in CEN/TC251)
    record item ENV12265
     
    "architecture component for the representation of basic information elements in electronic healthcare records" (CEN ENV12265, 3.14) "record items" are organised into "record item complexes" (at least one record item and one record item complex in a record) Terminologically, is made of name, content, context
    record item complex ENV12265
     
    "architecture component for the representation of the logical structure and the content within this structure of healthcare records" (CEN ENV12265, 3.15) (record item complexes can be nested in other record item complexes; a record item complex could be the record itself) Terminologically, is made of name, content (subordinate record item complexes or record items), context
    name of record item
     
    a label at the lowest level (leafs) in the record structure
    name of record item complex
     
    a label used in the record structure, not at the lowest level
    heading
     
    name of a topmost-level record item complex
    value domain (syn: value set, answer list, domain)
     
    the list of admitted values for a record item (i.e. given a name and a context)
    semantic structure of the record
     
    record items, record item complexes and their organisation
    semantic context
     
    structural context or dynamic context of a record item or a record item complex
    structural context
     
    chains of record item complexes above a record item or a record item complex
    dynamic context
     
    actual values of other record items (e.g. on sex, age, pregnancy) influencing the value domain of a record item or the components of a record item complex
    operative context (syn: co-text)
     
    actual values of other record items related to the same activity (i.e. included in the same document).
     
    if documents are nested in higher-rank documents, the scope of the operative context is arbitrary, i.e. it depends on the level of document considered
    name+content+context
     
    triple introduced in [Rossi Mori, Galeazzi et al 1997], to indicate the terminological aspects of a record item or a record item complex and the semantic continuum among them
    unit of information
     
    the complete meaning behind a record item or a record item complex (corresponding to name + content + context)
    paraphrase
     
    expression in a controlled language to explicitly represent without ambiguities all the details of a unit of information
    controlled language
     
    language restricted by explicit rules for the usage of words and constructs, to avoid ambiguities and imprecisions
    categorial structure
     
    schema that presents semantic categories and their relations in a given subject field
    cross-thesaurus
     
    set of descriptors, arranged by semantic category, derived from the comparative analysis of multiple corpora, to be used for uniform representations of entries across those corpora
    dissection
     
    structured expression, made by descriptors and semantic links according to a categorial structure, to represent systematically a phrase or a paraphrase
    three generations of terminological systems
     
    the first generation
     
    is made of enumerated lists, optionally with defined hierarchies and modifiers;
    the second generation
     
    is made of enumerated lists, with corresponding systematic dissections made according to pre-defined categorial structure and thesaurus of descriptors;
    the third generation
     
    is made of formal systems, with a model based on a description logic and with a corresponding engine.
     
    CEN/TC251/WGII is producing standards on categorial structures and thesauri of descriptors.
    QA
     
    Quality Assurance
    HL7
     
    Health Level 7
    LOINC
     
    Logical Observation Identifiers, Names and Codes
    XML
     
    eXtensible Markup Language
    SGML
     
    Standard Generalized Markup Language
    DTD
     
    Document Type Definition
    HTML
     
    HyperText Markup Language
     
    Acknowledgments
      This work was partially funded by the Italian Project "Sistemi Informativi Ospedalieri Integrati ed Integrabili" from MURST and by the European projects "GALEN-IN-USE", "TOMELO" and "HANSA". It is also a contribution to the G7 project on "Multilingual labelling of the Visible Human". Although some of the Authors are involved in the activities of CEN/TC251/WGII , HL7 SGML / XML SIG , and HL7 Vocabulary SIG , the present report is not on behalf of these organisations.
      The authors would like to acknowledge the contribution-at-large of all members of the HL7 SGML / XML SIG while at the same time asserting that the views contained here belong to the authors and do not necessarily reflect a concensus of the SIG membership.
    Plenaries   Table of contents   Indexes   Why Your Document Management System Should Care About Hyperlinks