The Arden Syntax: A sleeping HL7 giant

The Arden Syntax: A sleeping HL7 giant

Although the subject of frequent presentations and tutorials at HL7 meetings, the Arden syntax is one of those oft-overlooked components of the HL7 standard. In fact, that is one of the biggest problems with this syntax: that it is not very well known. Furthermore, its name does not indicate clearly what it is or what it can do.

So what is Arden? A comprehensive attempt to represent medical knowledge in a standardized and transferable format and thereby promote knowledge sharing between medical institutions. It captures medical knowledge in small “medical logic modules” (MLMs) that include triggering events, information about data that can be accessed from a database, and some logic representing medical rules and potential alerts. Arden first appeared about 10 years ago under the ASTM standards organization, but it is now part of the HL7 standard (it still needs to be harmonized with version 3.0 of HL7, however).

How might Arden syntax be used? Say there is a relationship between the output of a specific lab test and a potential risk, as with glucose content outside a normal range. The trigger event might be an HL7 message with the test result, the rule being a normal to abnormal range; the alert might be an e-mail to a physician.

Instead of every institution having to develop its own rules regarding this process, a standard set of MLMs could be developed, exchanged, and reused by other institutions. Major healthcare organizations such as Kaiser Permanente could develop these rules based on research or from their large patient bases. A community hospital could then take these existing MLMs and customize them as appropriate for their practices.

With the increasing emphasis on workflow managers, intelligent routers, and the like, there is pressing need for this type of standard. But despite its long tenure in the HL7 world, Arden syntax is just beginning to be implemented in medical imaging.

The reasons for this lag are many. In addition to the lack of knowledge and understanding about the potential of this standard, the development of decision support systems is still very much in its infancy, primarily because of physician resistance. However, most rules-based engines are typically used as “push technology.” That is, a physician will receive an alert or message to support the treatment process, not override it.

Early experiments have demonstrated that rules-based engines relying on the Arden syntax have a positive impact on patient care because they facilitate more timely scheduling of tests, which can prevent potential complications.

I urge anyone who is considering implementing a rules-based engine to look carefully at the Arden specification and learn about the power of its constructs. It could save a lot of time and effort.

Comments/questions: Herman Oosterwijk at herman@otechimg.com

• Search Again
• Return To HNN Homepage