Conformance Drift in Consolidated CDA R2; Part Three of Three
PART THREE: Tying in With C-CDA R2 Ballot Reconciliation & Conclusion
Tying in with C-CDA R2 ballot reconciliation
In determining the resolution of ballot comments grouped as “tighten constraints,” the C-CDA IG development team settled on five heuristics as the basis of their dispositions. Here are my thoughts on the applicability of those heuristics:
- [Heuristic 1] Further constraints aren’t needed, given CDA R2 inheritance rules—I agree with this heuristic as used in the dispositions.
- [Heuristic 2] In cases where we pulled in optional CDA R2 components so as to include narrative guidance, the intent isn’t to further constrain, particularly in the case of uncommonly used attributes—I agree with this heuristic as used in the dispositions.
- [Heuristic 3] Do not change an R1.1 MAY to an R2 SHALL (but okay to go from MAY to SHOULD)—This heuristic could benefit from explanation of why a change to SHALL would not be appropriate. The comments here provide such a reason.
- [Heuristic 4] Don’t constrain to SHALL where NULLing is hard (on participants)—I agree with its use in the ballot not because nulling is hard but rather because “SHALLing” is not beneficial enough to warrant difficult NULLing in those instances.
- [Heuristic 5] Focus on clinically meaningful. Where we really want to encourage the data, and where not too onerous to NULL, then further constrain—I agree that essential subcomponents of a concept or essential business process pieces should be SHALLs.
As an industry, we need to give more thought to the rationale behind why and how conformance rules are applied to implementation guides. Good spec design is not an end in itself; however, avoiding pitfalls such as conformance drift in IGs would make for better specs and better health care.