Thursday, January 16, 2014

The Rules for Displaying what isn't there in MeaningfulUse are Invisible

The following question showed up in my inbox since I subscribe to the Transport Testing Tool Google Group:

We have come across an issue with the human-readable display of C-CDA sections. It is our understanding that systems receiving C-CDA documents must render all human-readable content from the received C-CDA. However, our certifying body is expecting certain sections of the C-CDA to display when that section is not included in the xml file that we receive from the TTT.

Is it appropriate to display sections of the C-CDA in the human-readable when they are not included anywhere in the xml? If so, can someone point us to the documentation that explains how to best accomplish that?

Here's how I responded:

From the CDA standard and C-CDA implementation guide perspective, there is NOTHING that says you MUST make content not in the human readable narrative visible on the display.

What the CDA standard suggests can be found in sections 1.2.3 (Human Readability and Rendering CDA Documents), and in 1.3.3 Recipient Responsibilities.

What the MU Certification requirements say for view (the critical component of the query above) for 2014 is (see 45 CFR §170.314(e)) found under Patient engagement requirements:
(e) Patient engagement.
(1) View, download, and transmit to 3rd party.
(i) EHR technology must provide patients (and their authorized representatives) with an online means to view, download, and transmit to a 3rd party the data specified below. Access to these capabilities must be through a secure channel that ensures all content is encrypted and integrity-protected in accordance with the standard for encryption and hashing algorithms specified at § 170.210(f).
(A) View. Electronically view in accordance with the standard adopted at § 170.204(a), at a minimum, the following data:
(1) The Common MU Data Set (which should be in their English (i.e., non-coded) representation if they associate with a vocabulary/code set).
(2) Ambulatory setting only. Provider’s name and office contact information.
(3) Inpatient setting only. Admission and discharge dates and locations; discharge instructions; and reason(s) for hospitalization.

The NIST test procedures must be followed by testing labs, you can find the approved procedure for VDT here.  That does not specifically state that data not present be shown, only that the MU Common Data set be shown. Items 1-6 and 16 in their list actually come from the CDA Header, and so would have to be shown according to the test procedure and Meaningful use requirements.

So, if the document doesn't contain data from the MU data set, what should you do? That's really an implementation decision. That strikes me as being a sub-regulatory implementation decision, but one that should be different across certifiers. It sounds as if your certifier has determined that the right behavior is to do something that shows that there is missing data from that data set.

For items 1-6 and 16 in the list, if the item isn't present, best practice (my opinion, not an ONC rule or NIST requirement) would be to indicate that it wasn't there, since you wouldn't expect a user to remember what those items are all of the time. Items 7-15 in that data set all show up as part of the narrative portion of the document, and there should be way in the document to say "unknown", although admittedly that can be challenging sometimes (as for labs). So if it's a compliant document, displaying the narrative section content should address the display issues for items 7-15 in the common data set.

You can find some guidance on how to address these issues in the S&I Framework Transfers of Care Companion Guide. While this guidance does not have official status anywhere yet (other than S&I Framework consensus), it is presently going through HL7 Ballot, and I fully expect this to show up in the 2015 Optional Certification criteria (in fact this guide makes much more sense than trying to include C-CDA 2.0 for 2015, because it is compatible with the existing requirements for 2014). However, you won't find much on dealing with display of missing data in that other than what I've already said above.

I'd love to hear what some of the authors (especially David Tao) of that guide have to say.

Basically, the rules about what to do when the data isn't there, aren't there.


1 comment:

  1. As the technical editor of the S&I Framework TOC Companion Guide, I'd say the question of "what to display when data you expect to receive is simply missing" was outside of the scope of what the Guide covers. What we set out to do was inform creators of C-CDA how to do so properly, and it seems that the circumstance you're dealing with relates to receiving C-CDA documents that arguably don't comply with the ONC-mandated data requirements.

    Looking just at the Transitions of care certification requirements, specifically the Data Incorporation criteria - 170.314(b)(1)(iii), I would think it would be reasonable for an implementer, for their "human-readable display" to render a received C-CDA that always shows info on Medications, Problems, and Allergies, regardless of whether or not the corresponding sections are in the inbound document. With that said, I'm not sure from a certification perspective how EHR technology can be faulted if they don't do this. As I read the display requirements, it simply says you need to 1) Display "the data included in transition of care/referral summaries" and 2) Display "each additional section or sections [...] that were included" - there's nothing in there that addresses what to do with info that's missing.

    I feel like the heart of this question relates to category of EHR capabilities entirely ignored by the Certification Criteria but important in the real world - elegant exception handling.

    Russell Ott
    Deloitte Consulting LLP