Starting a healthy dialogue
- By Ron Miller
- Jun 02, 2003
Long before the Sept. 11, 2001, terrorist attacks, the Centers for Disease Control and Prevention recognized the need for a single, integrated disease reporting system — for naturally occurring outbreaks or bioterrorism incidents.
Bringing together health data nationwide required a common data transmission system that would allow the different entities to share data from a variety of databases and computer systems. That system required a common language, and CDC officials determined early on that Extensible Markup Language provided the best avenue to achieve smooth data exchange.
Using XML, CDC's integrated system — known as the National Electronic Disease Surveillance System (NEDSS) — will provide a way for states to enter and transmit disease data to the federal agency using industry-standard technologies.
"It is a major collaborative initiative involving public and private partners intended to streamline movement of public health data by taking advantage of advances in information technology," said Daniel Pollock, a medical epidemiologist at CDC.
Efforts thus far on NEDSS illustrate the critical groundwork required to adapt the malleable XML for a specific project.
XML consists of a set of user-defined tags that provides a common definition of data elements in order to facilitate data exchange across heterogeneous computer systems. Unlike HTML, which uses static tags to describe how information should be presented on a Web page, such as in bold or in a chart, XML is used to describe data types and how computers can process them.
"Being able to define custom tags within the XML document, which is obviously in contrast to HTML static tags, is a key element of this technology," said Jim Kauflin, NEDSS technical architect with Computer Sciences Corp., a consulting firm working closely with CDC on the project.
For disparate computer systems to communicate data to one another, they need a common set of XML tags. "XML is just a tool, and what we need to do to make that tool work for NEDSS, or any other interoperability project, is to agree on a standard set of tags so that the [system] is truly interoperable between organizations," said Dr. John Halamka, chief information officer at Boston-based CareGroup Healthcare Systems.
Once parties agree on a set of tags, XML provides a way to take advantage of a group of other standards collectively called Web services. NEDSS participants will be able to use Web services to build systems that can exchange data over the Internet very efficiently. "The real win of XML is fast interoperability using standards that can be done quickly and at low cost," Halamka said.
With NEDSS, the parties — in this case, 50 states, seven municipalities and CDC — needed not only to come to an agreement around a common set of data tags, but also to develop a way to allow each computer within the system to understand and interpret those tags. Enter the XML schema, a document that defines the set of tags and checks to be sure they were applied
correctly in a process known as validation.
"The XML schema expresses the shared vocabulary that allows the machine to carry out rules made by people," CSC's Kauflin said. "So, we can use the schema to define the structure, content and semantics within an XML document."
What's more, the project required a standard format to help define its data. In this case, project leaders chose the Health Level 7 (HL7) standard, which has been used in the medical industry for many years to allow systems within the same medical organization to communicate with one another.
Basically, HL7 defines the semantics of the message being communicated, said Mead Walker, an independent consultant working with CDC on the implementation of HL7 for NEDSS. In this case, the semantics provide a way to express various types of health care-related data. XML provides the syntax for the message, or the format for how the information will be organized.
In the upcoming release of HL7, Version 3, project leaders are moving to expand the format to allow the transfer of information, not only among different systems within the same institution, but across systems in a variety of locations (a NEDSS requirement).
"The tie-in with HL7 is that HL7, Version 3, is being designed with the intention that it would work well with an XML communication," Walker said.
Abundance of Tools
The use of XML and industry standards was important to CDC officials because they wanted to make it as simple as possible for states to link to NEDSS using
"The other strong point for XML is that there are a great many tools becoming available to allow you to process XML messages, so the availability of commercial tools that are of general purpose allows us to craft systems much more rapidly then if we were using a proprietary building [method]," said Joseph Reid, associate director for science at CDC and chief technology officer for NEDSS.
Project leaders also believed that using standards would provide a way for states to adopt NEDSS at a local level. To that end, CDC has developed what it calls the NEDSS base system, a starter schema of sorts. Developed by CDC and CSC, it allows states to quickly get NEDSS up and running.
"We have roughly, at this point, over 20 states that have requested the base system," Reid said.
CDC is offering grants to other states to develop their own NEDSS-compliant systems. "The remainder of those are presumably interconnecting [existing health information systems], or they are building their NEDSS software in whatever development technologies that they choose to use," he said.
As individual states apply for grants and begin their development projects, there is an effort to share information about that process among states and CDC.
"A lot of states do similar types of things. Then there are conferences where people discuss what they've done and see what works and what doesn't," said Bob O'Doherty, CIO at the Colorado Department of Public Health and Environment. "Each state is trying to do something that works for that state then share knowledge with other states."
Ultimately, the real test is actually sending HL7 messages in XML from the states to CDC. "The proof of the pudding then is whether or not you can send the messages that the CDC will be interested in," he said. "That's how you really know the schemas are compatible. Can I send you messages that we both agree are what you want?"
Although NEDSS is still in development, its core building blocks are complete, and CDC and state health departments continue to work together to refine the process.
"I think the No. 1 challenge is communication — understanding what we need to communicate in public health, understanding how to express that within the HL7 message development framework, understanding how we want to give shape to the specification using XML in such a way that we make sure the content that we're intending to send from states and local partners to the CDC remains as we specified to begin with, without losing meaning," CDC's Pollock said. l
Miller is a freelance writer based in Amherst, Mass. He can be reached at firstname.lastname@example.org.
Extensible Markup Language schema: A document that defines a customized set of XML tags for a specific project. It is used in a validation process to make sure the tags are applied correctly.
Participants: Centers for Disease Control and Prevention, all 50 states, seven municipalities and outside consultants (including Mead Walker, an independent expert in Health Level 7 data standard messaging, and Computer Sciences Corp.).
Nature of information exchange: CDC wants to provide a way for states to transmit disease data over the Internet on a regular basis from a variety of computer systems.
IT solution: Using Extensible Markup Language, Health Level 7 and industry-standard development tools, states can transmit data to CDC from many databases and computer systems in a common messaging format. CDC has developed a base system (general XML schema) for states to use, or they can develop their own systems using grant money supplied by CDC.
Cost: To date, $37.5 million in federal money has been funded/obligated for state grants to support the National Electronic Disease Surveillance System (NEDSS) initiative. Approximately $6 million in additional federal funding is scheduled to be awarded to states in fiscal 2003 for a total of $43.5 million from fiscal 2000 to 2003.
Program costs: To date, $40.8 million in federal money has been funded/earmarked for NEDSS program activities. This cost includes funds used by contractors to work on the actual NEDSS Base Systems and program area modules, software/hardware and support contracts to support the initiative. A request for additional funds to support program costs is estimated at $10 million for the remainder of fiscal 2003 for a total of $50.8 million from fiscal 2000 to 2003.