The S&I Framework is the place to be to understand where the industry (including Meaningful Use) is headed. It has some of the strengths (open participation) and challenges (difficulty integrating the workgroups) of HITSP, but has more direct buy-in from ONC, and also has the admirable goal of pilot testing standards to prove their viability before incorporating them into regulations. However, not everything done in S&I Framework is connected to Meaningful Use stage 2; some is more futuristic.
So here are my three “Aha!” moments.
- Transitions of Care. The dots still need to be connected across the many pieces in this initiative. While there’s convergence upon a single standard (CDA Consolidation Guide, which is close to passing its HL7 ballot), that in itself is both too much and yet not enough for ToC. Too much? Many parts of the guide aren’t necessary for ToC. Not enough? The sections that are necessary don’t have the core data element constraints of the Clinical Information Model (CIM) specified clearly enough to know what must be implemented and piloted. I was also surprised to learn that the current reference implementation work in ToC is based on HITSP C32/C83/C80, not yet based on the CIM and CDA Consolidation work because they’re waiting finalization. So they’ll have to flip over quickly to address the emerging standard in order to pilot it. Piloting tools based on C32 would only be nice to have for developers who haven’t already done it, and it wouldn’t make sense to have to pilot them prior to Stage 2 final rulemaking.
- Query Health. I had thought a “query” was something like an SQL query or an IHE document requestry or patient identity query, and assumed it would be synchronous. As Gershwin would say, “it ain’t necessarily so!” While those are indeed queries, QH has a broader definition, and seems to be leaning strongly to asynchronous messages. For example, a query might be composed of two separate (but correlated) push transactions, like a Direct message conveying the “query” (question to be asked, such as “How many diabetic patients with elevated A1C were on X therapy?”). Then the EHR or HIE would run the query and send back the results (again possibly via a push transaction). So it could be like a secure email and a reply to that email. Or it could be like an HL7 Lab Order and a subsequent Result observation message. The QH Technical Workgroup will evaluate the various technical standards to satisfy the use cases. But beyond whatever technical standards are endorsed, I’ll be interested to see how QH defines a mix of predefined query parameters vs totally ad hoc queries, and how the systems are expected to understand what is being requested.
- Data Segmentation (for Privacy?). I was only at about an hour of these discussions, which were moving slowly. But what struck me was the debate about whether this initiative should focus solely on privacy. Clearly, data can be “segmented” for multiple purposes, but given the difficulty up till now of getting a standardized privacy solution deployed even for simpler use cases, I was surprised that broadening the scope was being advocated by some. Other efforts have struggled between “narrow vs broad” use cases. In this particular case, I think the challenge is big enough on privacy that it makes sense to have a narrow focus rather than to “boil the ocean” with other uses of data segmentation. Perhaps a clear definition of phases and scope would help: Phase 1 is solely focused on standards and pilots for privacy data segmentation; later phases can expand the scope. I view this as primarily a problem (granular privacy consent) in search of a solution, rather than a solution (segmentation) in search of many problems.
So all in all, it was a productive meeting. Much work still remains to be done, and there’s way too much going on for most organizations to stay on top of. Still, I appreciate ONC’s leadership in convening and “uniting the tribes” as Dr. Mostashari said at the Care Transitions meeting.