Convert your FHIR JSON -> XML and back here. The CDA Book is sometimes listed for Kindle here and it is also SHIPPING from Amazon! See here for Errata.

Thursday, August 4, 2011

On Comments disparaging of ISO 21090 Healthcare Datatypes

Recently, someone shared with me some rather disparaging comments made about ISO 21090 Healthcare Data Types, otherwise known as HL7 Data Types Release 2.0.  In those comments several assertions were made:
  1. TC-215 is trying to create something completely new (or acquiring something completely new created by HL7), rather than standardizing something that has had substantial implementation and development.
  2. Competencies in modeling and software development are under-represented in ISO TC-215.
  3. People have voted in blocs without reading the materials, without showing that they have the competency to understand them.
  4. The existing work violates principals of Object-Oriented Design.
  5. The notion of profiles of standards does not exist anywhere else in informatics.  
  6. The process did not show that the standard was implementable, implemented, or fit for the purpose by which it was created.
  7. Claims that certain organizations (that the commenter is not affiliated with) do not support the standard because they are using something else.
These assertions are largely unsupported by facts, evidence or any other references to support the claims.  I would like to present my counter-claims:

#1: IS0 21090 is a submission of HL7 Data Types Release 2.0.  That work is based upon HL7 Data Types Release 1.0, which is one of the foundation components of HL7 Version 3, and thus CDA, Care Record, and Patient Administration standards or DSTUs.  Substantial implementation and development has been done using these standards worldwide.  As evidence1, I demonstrate the XDS map, which also shows where CDA is in use, and where PIX/PDQ V3 (based on the HL7 Version 3 Patient Administration standards) are used, and the IHE product registry, which shows more than two dozen products implementing V3, and a greater number implementing CDA.

#2: While I cannot speak for the whole of ISO TC-215, I am quite familiar with many of the members of ISO TC-215 from Germany, Canada, the US, Japan, Australia and a few other countries. I don't find modeling and development experience under-represented in TC-215.  But don't trust my opinion, try this Linked-in search and see what you think (your results may vary depending upon your use of Linked-in).

#3: With regard to blocs, having read the material, and competency: I have seen quite a bit of discussion on HL7 lists showing evidence that ISO TC-215 representatives from several countries have both read the material, understand it, and have offered significant suggestions for improvement.  These are competent individuals that have demonstrated their abilities numerous times.  That traffic also represents a pool of quite experienced modelers and developers, further addressing point #2.  The Ad Hominem attack on the competency of these individuals is unwarranted.

#4: This is the one claim that I cannot refute. Disproving a general negative is nearly impossible.  I'd like to see what principles of OO design were violated, but no such information has been provided in the comments.  Until such information is provided, this is merely an unsupported opinion.
#5: The principle of profiling is the foundation of two organizations quite well involved in Healthcare informatics:  IHE and Continua (click the links to see the connection of the organization to the concept "profile").  There are also General IT standards organizations that Profile standards (e.g., see OASIS profiles of web services).

#6: See #1.  The data types are in large part identical to HL7 Data Types release 1, which have been implemented worldwide, and were deemed suitable for purpose by nations, regions, and healthcare provider organizations of all sizes.

#7: With regard to claims that organizations are using something other than ISO 21090 for other purposes, the commenter is correct.  But, they also happen to be using the forerunner of the 21090 data types in existing work, and will be using the 21090 data types when they become available in the HL7 V3 Normative editions of certain standards and drafts already in use (e.g., in IHE Profiles).  If you care to see the opinions of those organizations on ISO Datatypes, you can view the publicly available HL7/ISO Joint ballot results in September and in May of 2008.  I voted affirmatively, as did almost all others.  Cross check the voters in this pool with the comments posted on HL7 lists (see #3 above) and you will find that they voted on this material after both reading and understanding it.

   -- Keith
1 I find it interesting that I have to use IHE resources to show HL7 implementation. I wish the HL7 product registry project would get off the ground.