Data Maintenance on the Web Made Easy with XML Templates   Table of contents   Indexes   An Object/Relational Approach to Content Management
Almonte
 Canada 
Killdara Corporation
Ogilvie, John
 		John Aloysius Ogilvie
 President
Killdara Corporation
 73-77 Mill Street Almonte (Ontario)  Canada (K0A 1A0)
Email: mailto:john@killdara.com Web site:http://killdara.com
 Biography
 John Ogilvie is a systems engineer who currently runs Killdara, a software products and consulting company. Over the past seventeen years, John has executed dozens of software projects in many fields. Healthcare customers have included the Veteran's Administration research group at Loma Linda, CA, and radiology vendor ISIS (now IS2 Research), in Canada. The most recent healthcare project was a system for ICUSA in Baltimore, MD for the remote surpervision of multiple intensive care units from a networked command center. Killdara is developing the Paraphrase Engine, an XML-based product with applications in healthcare.
 

Introduction

 "The health care industry is particularly ripe for reform through Internet technology because it involves a long paper trail among fragmented players, including manufacturing, pharmaceutical, health providers, and insurance companies. " (Red Herring Magazine)
 "Health care accounts for one-seventh of the economy, and it's hugely inefficient, ripe for the efficiency gains of e-commerce." (William Whyman, Internet strategist at Legg Mason's Precursor Group)
 In May 1999, the Ontario government threatened physicians with legal sanctions if they didn't start writing legible prescriptions: illegible handwriting has cost the life of at least one patient. There cannot be another industry where handwriting is still permitted for critical documents..
 This paper addresses the practical issues in building commercial products which implement HL/7 and standard healthcare vocabularies in XML.
 

The Problem

 It's well known that healthcare is still largely paper-based. The industry has been slow to benefit from productivity increases made possible by computer technology over the past fifteen years. Many reasons have been suggested.
 There have been numerous industry organizations formed to address these issues. One area which has seen a lot of attention is standards for healthcare informatics. Without standards, there can be no electronic exchange of documents, and therefore clinical documents such as lab results are manually retyped into a series of computer systems, introducing errors and delaying care.
 

The Solution is XML

 Healthcare provider organizations,typically spends several million dollars and several years to build Computerized Patient Records (CPRs). There is finally an elegant and inexpensive way to share these records with other organizations, using industry standards for document exchange : HL/7 and XML.
 HL/7 ("Health Level 7") is an accepted standard for messaging between healthcare information systems (HIS), but it has been held back by the lack of software which understands this specialized standard. XML is the digital document format which will replace HTML on the web, especially for business-to-business communication. When the two standards converge in the coming year, we feel that the electronic walls which separate different healthcare organizations will be torn down rapidly.
 HL/7 has had some success to date, but nothing as overwhelming as say, the adoption of the web. The reason: it takes man-months to write interfaces to send/receive HL/7 messages. However, this past winter, a progressive committee at HL/7 (the "Kona" group) has quickly laid the foundations for HL/7 expressed in XML. They demonstrated ten vendors exchanging messages at the HIMSS show this Spring.
 

What will we need to make this work?

 The healthcare industry is fragmented and conservative. Therefore, we need a solution which is easy to install and maintain and which doesn't disrupt the existing workflow. Perhaps a "back-room" network device which will not be visible to users, so that they won't need retraining. And it must be compatible with existing HIS applications: it must extends these application, rather than replacing them.
 Another part of the solution is the use of standard vocabularies, standard 'nouns' and 'verbs'. Is this a "CXR" or "chest x-ray"? There are many synonyms for diagnoses, medications, tests, observations. Standardized vocabularies are emerging, but there is no means of enforcing them. Two healthcare organizations may have radically different formats for the same document.
 

HL/7

 HL/7 ("Health Level 7") is ten-year-old non-profit org trying to establish standards for document exchange in healthcare. They have had some success to date, but nothing dramatic. The reason: it takes man-months to write interfaces to send/receive HL/7 messages.
 

XML

 "Extensible Markup Language" is a descendant of SGML, but simpler. XML resembles HTML, but it's better suited to complex documents, since it describes the meaning, not the appearance of the document. It is designed to be extended for new applications. XML has broad industry support: there are no competitors, and Microsoft is promoting XML heavily. In the web industry there is a widespread acceptance that 1999 will see the emergence of XML as the language of a next-generation web.
 A progressive committee at HL/7 (the "Kona" group) has quickly laid the foundations for HL/7 expressed in XML. This is leading towards clinical document exchange using standard templates (DTDs).
 Who is defining the DTDs?
 Standard vocabularies
 There are multiple, sometimes overlapping coding schemes for diagnosis, treatments, drugs, labs:
 
  •  NDC
  •  LOINC
  •  ICD9/10
  •  CCI (Canada)
 Which of these are suitable for use in XML-based clinical document exchange?
 Who do we need to Convince?
 We estimate that there are about 6,600 hospitals in the U.S. and Canada. In addition to these hospitals there are many private laboratories, multi-specialty clinics and other facilities.
 The U.S. Health Care Financing Administration did detailed Active Provider Facility Counts in a number of states in 1996. On analysing their results we determined that
 If we extrapolate this to the entire U.S. population in 1996 (226M), we would expect there to be approximately 226,000 provider facilities in the entire U.S. (A confirmation of this estimate is the statement by HCFA that they oversee 158,000 laboratories (70% of 226,000).)
 We estimate that there are therefore at least 226,000 potential users in the U.S. alone, and probably an addition 20-25,000 in Canada (the infamous divide-by-ten rule).
 

What Issues will they Have?

 
  •  Reliability
  •  Availability
  •  Data Integrity
  •  Data Security
  •  Auditability
  •  Usability
  •  Legal Characteristics
  •  Confidentiality
  •  Authentication
  •  Retention
Data Maintenance on the Web Made Easy with XML Templates   Table of contents   Indexes   An Object/Relational Approach to Content Management