Letting Machines Read
Maybe the easiest way to explain why the XML approach is important is to start by thinking about what PDF does and doesn't
do. PDF is designed to be easily readable by human beings; the format makes it easy to display a document on a screen or to
print it out. That's useful for many purposes, but not for others. If you need to search through large volumes of information,
for instance, human readability is less important than the structure of information. If you want to transfer data into a database,
you need to have a way to categorize and label the different types of information.
In a presentation last June, Randy Levin, MD, director for health and regulatory data standards and associate director for
medical informatics for CDER, mapped out some of the shortcomings of today's labeling. Paper labels are difficult to read
and copy, he said. There are problems of accessibility. Distribution is limited and the information cannot be accessed directly
by computer systems. Drug terminology and code sets are not standardized and up to date. As a result, healthcare systems
generally draw their information from third-party suppliers. And what they get does not come directly from FDA, and it is
frequently out of date. Curing those problems is one of the main objectives of SPL.
"Machine readability" issues are crucial in creating a system for electronically distributing prescribing information. At
the same time, it is quite desirable to keep prescribing information in "human readable" form. XML is a simple, very flexible
text format that was originally developed to meet the needs of large-scale electronic publishing. XML is playing an increasingly
important role in the exchange of a wide variety of data on the web and elsewhere.
XML has numerous advantages:
It preserves human readability and separates content from presentation. This allows authors to focus on content and improves
consistency of the presentation.
- XML also enables a usable format. Information can be easily stored and access to specific parts can be provided to various
consumers of labeling: including reviewers, customers, and other interested parties. Data are maintained in one place but
can be published in many formats.
- XML promotes reuse within a document and across multiple output documents. It delivers information to users on any platform,
such as Windows, Macintosh, or Unix, and it is easily transmitted across the internet.
- XML is not proprietary so it is tool-and platform-independent in an accepted archive format.
- XML is fully internationalized.
- Semantic markup of documents in XML facilitates machine readability, e.g., extraction, storage in a database, re-use, automated
assembly and re-ordering.
- XML will never become outdated. Version upgrades will make data "backward compatible."
Accurate Data—Revised Daily
Technically, XML is not a computer language, but what programmers call a "metalanguage"—something used to create a variety
of new languages. In the case of SPL, there was a two-step process. First a general standard for healthcare information was
developed, and then SPL was developed to work within that standard.
The broader, XML-based healthcare standard is known as the Clinical Document Architecture (CDA). It was developed by the standards
group Health Level Seven (HL7) and is expected to serve as the foundation for a nationwide e-health information system. A
small group within the HL7 Regulated Clinical Research Information Management (RCRIM) Technical Committee did the groundwork
for SPL, basing it on CDA. The PhRMA HL7 Task Group formed the SPL Working Group in January 2004 to further the work of the
initial development team. The SPL working group was composed of experts from HL7, industry, technology solution providers,
and FDA, with a goal of addressing key challenges of design and implementation of the SPL standard, including document and
requirements analysis, model definition and construction, management of the ballot process, and development of an implementation
guide. In May 2004, SPL passed the HL7 Committee Ballot process and has now officially become an American National Standards
Institute standard (ANSI/HL7 SPL, R1.0-2004).