Evaluating Content Management Systems Based on Information Chunk Size   Table of contents   Indexes   From Publishing to Interaction: How to Gain Competitive Advantage Through XML with Dynamic, Interactive and Personalized Content
 Alschuler, Liora 
 East Thetford 
 The Word Electric 
 		Liora Alschuler
 consultant, writer
The Word Electric
 Route 5 East Thetford (Vermont) (05043)
Email: mailto:liora@the-word-electric.com Web site:http://www.HL7.org
 Biography
 Liora Alschuler is a consultant and writer specializing in the application of SGML and XML to healthcare information systems. She is a Senior Analyst at XML.com, the author of ABCD... SGML: A User's Guide to Structured Information, ITCP, 1995, and a Contributing Editor for the Seybold Report on Internet Publishing. Ms. Alschuler and Dr. John Spinosa are forming a partnership to be known as alschuler.spinosa which will specialize in the application of structured markup to healthcare information.
 Liora is Co-chair of the HL7 SGML/XML SIG which is developing standards for the application of SGML to healthcare information exchange. She was Project Manager in charge of clinical documents for the HL7 HIMSS '99 demo which coordinated 10 applications in a simulated exchange network using the next generation of HL7 XML-based specifications. 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, chartered by the SIG to create a ballotable draft of the specification, now called the Patient Record Architecture (PRA). 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, XML, 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).
Beeler, Jr. Ph.D., George
Mayo Clinic
 		George W. Beeler, Jr. Ph.D.
 HL7 President
Mayo Clinic
 Biography
 George Beeler is a Division Chair in Information Services at Mayo Clinic, Rochester, Minnesota. His is involved advanced technology planning, systems planning, data administration, development technology support, and standards. He received a BSEE degree from Princeton University, and M.S. and Ph.D. degrees from the California Institute of Technology. He joined Mayo in 1967 to pursue research interests in the application of computers to biomedical research. He assumed responsibility for the Section of Information Processing and Systems of Mayo Clinic in 1980, and moved to his current responsibilities in 1987.
 Dr. Beeler is a senior member of the Health Information and Management Systems Society, a member of the Institute of Electrical and Electronics Engineers and of the American Medical Informatics Association. In addition to his responsibilities at Mayo, Dr. Beeler is Chair of the Board of Directors of Health Level Seven (HL7), an ANSI-accredited developer of health information interchange standards. Within HL7 he has had a role in developing methodologies for HL7's next generation standards.
 Boyer, Sandy 
XML consultant
 		Sandy Boyer
 Pharmacist
XML consultant
 Biography
 Sandy Boyer is a pharmacist/drug information specialist and consultant. She worked for many years with the U.S. Pharmacopeia developing content for the USP DI drug information compendium and played a major role in the conversion of their database into SGML. She has been working for over 2 years with the AusDI project in Sydney, Australia which is publishing an SGML-encoded compendium of Australian drug information in print and on CD-ROM.
 Sandy has been involved with the HL7 SGML/XML SIG for 2-1/2 years in its efforts to develop standards for the application of XML to healthcare information. She is one of the original members of the KEG (Kona Editorial Group), which was assigned primary responsibility by the SIG for development of the PRA Architecture for ballot within HL7. She has given a presentation on use of XML in healthcare at XML Europe in Paris and co-presented The Basics of XML at the HL7 Plenary Meeting in September 1999 in Atlanta. She is a member of HL7 and the American Medical Informatics Association.
Owens Technical Research
Owens, Fred
 		Fred Owens
 President
Owens Technical Research
 Biography
 Fred Owens is President of Owens Technical Research of Melbourne, Florida offering information technology consulting and businessservices. A career engineer, technologist, and businessman, Mr. Owens has over 20 years experience in the fields of communications and information systems. Mr. Owens has a BS in electrical engineering from Virginia Tech and an MBA from Florida Tech. He has designed a variety of large and small systems for clients ranging from the United States DoD and NASA, to large tertiary hospital systems, and new startup e-commerce businesses. Since 1994, Mr.Owens has been working to create better designs for healthcare information systems.
Lincoln, MD, Tom
Univ. of IL; Rand Corporation
 		Tom Lincoln, MD
 Doctor
Univ. of IL; Rand Corporation
 Biography
 Additional Contributor
Gartner Group
Rishel, Wes
 		Wes Rishel
 HL7 Vice-technical chair, Board of Directors, Co-chair SIGOBT
Gartner Group
 Biography
 Additional Contributor
Scripps Health; CAP representative to ANSI HISB
Spinosa, MD, John
 		John Spinosa, MD
 Pathologist
Scripps Health; CAP representative to ANSI HISB
 Biography
 Additional Contributor
IBM XML Strategy and Technology
Sutor, Ph.D, Robert
 		Robert S. Sutor, Ph.D
 XML Industry Standards Liaison
IBM XML Strategy and Technology
 Biography
 Additional Contributor
 

The Role of HL7 in Healthcare

 HL7, Health Level 7 
Health Level-Seven (HL7)
 healthcare 
 		 Health Level-Seven (HL7) is an ANSI-certified Standards Development Organization founded in 1987 to establish messaging standards for interoperability among healthcare delivery and administration applications. Since that time, it has published several versions, each expanding the functional coverage. Version 2.3.1 was certified by ANSI in March 1998. It is supported by virtually all vendors of healthcare-specific applications, and is used at 70-90% of all U.S. secondary and tertiary institutions. Its usage is increasing among primary care as they begin to employ systems that support clinical functions. In addition to EDI-like messaging, HL7 publishes ANSI standards for portable expressions of clinical rules and visual application integration. Although founded in the United States, HL7 has chapters in Europe and the Pacific Rim and is beginning to be used in Latin America.
 

Opportunities for Improvement

 HL7 has recognized several opportunities for improvement. Despite its widespread use, the current HL7 standard requires adaptation at each site where it is used. This creates substantial per-interface installation cost. While some variability stems from inescapable differences in healthcare information practices, some can be laid to the dated methodology which describes fielded data independently of the underlying information structure.
clinical documents
 		Although the standard has some facility for sending narrative and transcribed reports in a structured, coded format, most such clinical documents are sent as monolithic blocks of text. This limits the ability of the receiving application to search, analyze or take programmatic actions based on the specifics of the report.
X12 (EDI)
 messaging 
 		The HL7 messaging syntax was designed in 1987, following the direction of X12 (EDI) , and several other messaging standards. The chief virtue of the syntax is that it is quite compact. This economy of bandwidth has become increasingly less important in the intervening twelve years. At the same time, HL7 continues to suffer the negative consequences of having a parochial syntax. Analysts are hard to recruit and require more training, and applications developers must write their own software for parsing and creating messages. Were the developers able to rely on publicly-available software tools, driven by syntax specifications not specific to healthcare, development costs would be reduced and per-site testing time would be reduced.
 

Version 3, the Reference Information Model (RIM), and Medical Vocabularies

 Object Management Group  
Reference Information Model
 Unified Modeling Language 
 		The organization launched a major initiative to address these problems in 1997. In HL7, the label "Version 3" stands for a complete re-thinking of the basis for information exchange. Under this general rubric, HL7 has built a Reference Information Model (RIM) which will be the basis for information exchange under the new standard. The RIM is a comprehensive, information model for clinical healthcare, constructed using the Unified Modeling Language , the standard modeling approach of the Object Management Group . Today, it includes the work of hundreds of experts. It comprises 120 classes, 782 attributes, and 160 associations. Messages are thought of as the information necessary to convey events from one application to another. The combination of an event and a specific sending and receiving application is called an interaction. The Interaction Model is a second lynchpin that helps to rigorously establish the semantics of Version 3.
National Library of Medicine
Unified Medical Language System
 		There is also a formal vocabulary model, crafted in a manner that is consistent with the Unified Medical Language System (UMLS) propagated by the National Library of Medicine . There is a well-defined interface between the vocabulary model and the information model. The Version 3 approach is not to re-invent medical terminologies but to permit the various terminologies to be employed in a rigorous fashion.
 In V3, the structure of messages will be derived strictly from the information structures expressed in the RIM and certain predefined data types to deal with the semantics of describing definite and indefinite quantities, complete or partially complete coding. The specific methodology for deriving messages from the RIM permits a far more rigorous description of the semantics of the data fields than was previously possible. This will reduce the need for site-specific analysis and decrease implementation costs.
 The structure and semantics of Version 3 messages is described using the Hierarchical Message Description (HMD), which is entirely independent of the syntax or "wire format" of the message. HL7 has determined that the XML syntax has the necessary structural features to express the structural information of the HMD. XML will be one of the syntaxes that can be used to construct HL7 messages. Indeed, it may finally be the only syntax. The message for each interaction can be described by a DTD that is algorithmically derived from the HMD.
 These DTDs, however, will not represent fully the semantics contained in the HMD. HL7 is defining a set of extensions of the Domain Object Model to support the development of tools that can post-process the output of an XML parser, further verifying semantic correctness. As a W3C member, HL7 is also participating in the work on XML Schemas, expecting a much more expressive metalanguage. Schema-driven parsers could greatly reduce the amount of post processing required to validate and use the information in Version 3/XML messages.
 

The HL7 XML SIG

HL7 XML SIG
 		In a small, independent start-up effort, a group of physicians and medical informaticists began meeting in early 1996 to explore the application of Standard Generalized Markup Language (SGML) and later Extensible Markup Language (XML) to the needs of healthcare. Among this group were several individuals active in HL7 including HL7 founder and the author of the original exchange spec, Don Simborg. The core question this group wanted to explore was whether structured markup, with its ability to specify, encode and validate the structure and content of free-form text and narrative documents might open up the hither-to impenetrable area of clinical documents to open exchange machine processing and facile re-use. Despite 30 years of effort, most of theclinical information required for patient care, public health monitoring, and the management and oversight of healthcare delivery remains paper-based or as excised bits or blobs stored in isolated repositories often without the original document context. In December 1996, this group made its first public presentation and announced its intention to join HL7. This occurred during the same meeting, GCA's SGML '96, at which XML was announced.
 Patient Record Architecture 
 		This meeting, XML '99, will mark three years of experience and experimentation with structured markup in the healthcare information exchange arena. We expect that the September HL7 Plenary meeting will see a proposal to ballot a recommended method for XML-encoding the current generation of HL7 messages, as well as further work on the proposal to use XML as a primary syntax for Version 3. In addition, we expect to introduce as a proposed Ballot Package a collection of material specifying, documenting, and supporting the first level of the HL7 Patient Record Architecture (PRA), a comprehensive framework for exchange of XML-encoded clinical documents.
 

The PRA

 HyTime 
Kona Architecture
 		The PRA is a framework for exchange of clinical documents that is built on the privately-funded Kona Architecture (all rights to the Kona work were given to HL7 at the conclusion of the project and it has since been renamed the HL7 Patient Record Architecture). XML DTDs written for specific applications and environments can be mapped to the PRA. Using a transformation script written in XSLT or perl, for example or using ISO 10744 (known as HyTime ), senders can transform documents with their locally-defined tags into documents carrying the industry-standard HL7 PRA markup. While it might seem simpler to write common DTDs that everyone can use, in practice, applications and providers need to express their own clinical practice and business rules in their DTDs. This ability is key to the success of any proposed information standard in healthcare where everyone knows that imposition of a single model for clinical documents will not fly. The PRA is a mechanism for automating exchange while making the minimum imposition on the expressivity or flexibility of local implementations. This attribute of the PRA makes a compelling case to providers and vendors. The benefit for exchange lies in the power of the XML PRA to express relationships between local DTDs used for document creation and the RIM using the architectural exchange DTDs.
 The PRA is based on a set of design principles that include keeping the barrier to entry low, while providing a migration path to sophisticated electronic medical records for implementers and for the standard itself. The basic level of the architecture supports display and simple processing. Level One PRA documents have a standard framework for including coded vocabulary terms and control terminology and can carry the semantic of the local markup as an optional feature. One benefit of the multi-level architecture is that implementation and standards development can inform and assist each other in a cooperative, iterative process.
 The hierarchical structure of the architectural exchange DTDs will be derived from document analysis. Like Version 3 messages, they will be consistent with the information structures in the Reference Information Model. The principles of vocabulary developed within HL7 will be used to specify the domains of code value associated with specific utterances in a standard encoding across all levels of the architecture.
 

HL7 XML: The HIMSS Demo

Health Information Management and Systems Society
 		Perhaps most exciting for those of us who have worked on the RIM, the PRA, and the XML ITS is the manner in which the integration of the SGML group into HL7 has led to new thinking on both sides about the explosion in possibility rendered by combining the "traditional" HL7 messaging approach with the new XML-based approach to narrative documentation. The synergy of HL7 XML messages and documents is encapsulated in a prototype exchange network based on these emerging standards which we designed, built and demonstrated on the floor of the Health Information Management and System Society trade show in February, 1999 and our planning for the upcoming HIMSS 2000.
 The HIMSS interoperability demo shows how even simple applications provide tremendous utility because basic patient, practitioner, event and document information is synchronized and compatible between the PRA and the HL7 V3 messages. Ten vendors and several consultants contributed time, software, and integration skills to make the demo a reality. While the actual hours spent in design and implementation are not recorded, the preliminary staging held at the offices of Wizdom Systems in Naperville, IL, three weeks before the conference consumed at least 500 person-hours, not to mention cases of soft drinks, gallons of coffee, and bushels-full of patience. This investment reflected the realistic nature of the demo that produced a ten-vendor, ten-application hospital network registering patients, ordering labs, recording hospital visits, archiving and retrieving records, all in real-time using frozen drafts of our standards in-development: the RIM, V3 messaging and the PRA.
 The resulting demo included several "firsts" in XML RIM-based messaging, PRA document exchange and message/document interaction. The demo used these scenarios:
 
  1.  Enhanced Clinical Care Scenario.
     A patient is registered in one system, seen by a physician as recorded in an Electronic Medical Record system, has lab work done by an external lab, and the combined information is available in the EMR, a data repository, and a medical Intranet. The technologies involved include a mixture of legacy technologies, client/server, and Web-based applications, and both document- and relational-technology databases. Indeed, some of the systems are operating on HL7 version 2.3, demonstrating the evolution that will be necessary to introduce the benefits of version 3 to the marketplace.
  2.  Repository scenario.
     A sophisticated medical records system and a transcription-to-XML system send PRA-compliant documents to an HL7 clinical data repository (CDR) and to an XML document repository. Each repository can respond to queries for documents from both sources–the CDR by indexing the V3 message envelope that carried the document and the document repository by parsing the document itself. Data fields in the V3 envelope message are automatically generated from the contents of the PRA documents. While the documents are created with vastly different technology, common processing methods apply to all, regardless of source. Both types of repository also store and index V3 XML messages including patient registrations, lab orders and results, so they can build a complete picture of patient care.
  3.  Lab scenario.
     A PRA History & Physical document triggers a V3 lab order message by including the appropriate LOINC code under "Labs" in the Plan section of the document. The HL7 interface engine translates the V3 message to V2.3 for the lab system. Then, going the other way, it translates the lab results from V2.3 to V3. The V3 XML results message is used to auto-populate the results section of a Progress Note. The progress note is written with a local markup language and transformed by an architectural engine to a PRA-compliant document before it is picked up by the HL7 interface engine for distribution.
 The demo showed how the HL7 PRA and RIM can use XML to break down the compartmentalization of healthcare information systems and the response among show-goers was strong and positive. Traffic in the 10x30 booth was steady throughout the 3 days of the conference and there was interest from the press and industry analysts. Most notably, the booth remained busy throughout the when the exhibits were officially closed. This is because representatives of vendor organizations would use their precious break time away from their own booths to look at ours.
 A related benefit, and one that was central to HL7's planning and budgeting for the event, was that it provided a compelling test arena for our standards development work. The unforgiving deadline of a tradeshow meant that the work had to be done and implemented and tested, even though it had not been perfected. This kind of demonstration is tremendously useful in a standards-development environment where waiting for tomorrow always brings a more perfect product, but not a published specification.
 

Presenting HL7 XML at XML '99

 Our presentation to XML '99 will show the process of deriving DTDs from the Reference Information Model, and scenarios and transformations that use them. This will be illustrated with the DTDs and documents used in the '99 demo and will report on the planning and design of the '00 demo.
Evaluating Content Management Systems Based on Information Chunk Size   Table of contents   Indexes   From Publishing to Interaction: How to Gain Competitive Advantage Through XML with Dynamic, Interactive and Personalized Content