Wednesday, February 10, 2010

Comments on the Interim Final Rule

The comment period for the Interim Final Rule on Standards runs from January 15 to March 15. I've spoken to many stakeholders in industry, hospital IT departments, and government. Here are the five general themes I'm hearing from the comments they'll be submitting:

1. Need for additional guidance on content standards
As I've noted in earlier posts, the IFR is a regulation and ONC did not want to provide too much specificity that would require changing regulation to support evolving implementation guidance. Although this was the right thing to do, the lack of constraints on HL7 2.51 and CCD creates ambiguity that could result in a tower of Babel - thousands of incompatible implementations of these standards. ONC needs to rapidly provide HL7 2.51 and CCD implementation guides (such as the HITSP Lab and HITSP C32 specifications)

2. Need for additional guidance on vocabulary standards
Do the vocabularies required - RxNorm, SNOMED-CT and LOINC apply to EHRs themselves or just the transactions sent between EHRs? What if a proprietary, self built medication vocabulary is integrated to the EHR but is mapped to RxNorm for all health information exchange? What about a proprietary problem list vocabulary that is mapped to SNOMED-CT? If mapping is allowed, what subset of these vocabularies should be used i.e. NLM's SNOMED-CT Core problem list subset, LOINC Lab ordering compendium etc.

3. Need for additional guidance on transmission standards
Transmission approaches need implementation guidance, per my previous blog posts. Stating that REST and SOAP are acceptable does not foster interoperability among various implementations.

4. Need for additional guidance on security standards
Encryption needs clarification. Is encryption of EHR data at rest required in certified EHRs? Although the intent is that data in transit and data on mobile devices needs to be encrypted, should vendors provide encryption in their products because they cannot know how they will be installed and used.

5. Need for additional guidance on quality measurement
The NQF Healthcare Information Technology Expert Panel work on quality measures is based entirely on CDA/CCD, yet patient summaries are summary may be sent in CCR or CCD. Thus, the NPRM's quality measures cannot be extracted from CCR using the current retooling. How do we reconcile incompatible definitions of patient summaries and quality measures?

I look forward to reviewing all the public comments, which will be made available by ONC via the web.


David said...

Hi John,
Your five top themes are all well chosen, and we applaud you for continuing to inform the industry in terms that are understandable. However, we suggest that there is another “big picture” comment that has at least as high a priority.

Provide Continuity of Direction and Adequate Lead Time. The HIT Standards Committee did an admirable job setting a “glide path” from now through Stage 3. That provided a sense of direction, and enabled an orderly progression toward increasing degrees of interoperability. Developers could aim for Stage 2 or even Stage 3, to advance rapidly toward ARRA’s goals. Unfortunately, the glide path concept was mostly lost in the IFR, as it only mentions Stage 2 in a fuzzy way such as “we look forward to receiving recommendations from the HIT Standards Committee” for a single patient summary record standard, whereas those recommendations (i.e., CCD) had already been made! And several new requirements not previously signaled by the HIT PC, HIT SC, CCHIT, or anyone else, appeared in the IFR. The net result has been:
• uncertainty as to direction,
• risk of previously endorsed directions now seemingly questioned (e.g., XDS/HITSP TP13, which have been HHS-recognized standards for transport since 2008),
• causing “churn” and rework,
• discouraging the industry from aiming strategically, and
• forcing people into a reactive mode.
Once the IFR is finalized, especially if it contains any more changes, there will be practically no time to recalibrate EHR development plans or for providers to implement those EHRs in time for 2011 MU.

Fortunately, there is a remedy to this problem, which is to remove those new items that were not in previous HIT PC/SC recommendations but introduced into the IFR. Also, to minimize such churn in the future, we urge ONC to set a strategic direction and glide path (through Stage 3 at least). As an example, related to your fifth theme, the direction for quality measures standardization via QRDA is an area where a glide path was recommended by the Standards Committee, and we hope will be reinstated (e.g., claims, PQRI OR QRDA in Stage 1 converging on QRDA by Stage 3).

While it’s difficult to retool the current IFR with a three-stage glide path, we suggest that ONC and its Federal Advisory Committees make every effort to incorporate the strategic glide path as they define subsequent stages, and that they establish continuity of direction, so that each Stage does not dismiss work that was previously done to follow standards, and new requirements with short lead times do not suddenly appear without adequate time to discuss and prepare.

David Tao and Hans Buitendijk, Siemens Healthcare

Anonymous said...

This is an excellent high level summary and I would also echo the comments by David.

Given that the IFR also deals with certification criteria, I would also highlight the very pressing need for further guidance on how certification will be accomplished and how these criteria will be used in practice. Vendors and providers alike eagerly await additional ONC guidance on these issues given the looming FY/CY 2011 start dates and the limited time available for software revisions. There is also a need for further clarification on the concept of modular certification and how it will work in practice.

Mark Segal