AHRQ Releases Draft Guide for Registry Interoperability: Does Public Health Have a Role?

Noam H. Arzt, Ph.D.On January 11, 2019, the Agency for Healthcare Research and Quality (AHRQ) released a draft Addendum to the Third Edition of Registries for Evaluating Patient Outcomes: A User's Guide called Tool and Technologies for Registry Interoperability. AHRQ has long written about registries - largely from a research standpoint - and I have been following this from afar for some time. This new guide is focused on helping those who both create and use registries understand the issue surrounding leveraging external data to improve registry completeness, accuracy, and usefulness.

This report covers lots of ground and does a good job of summarizing important subtopics. Each chapter is overflowing with footnotes and sources. In Chapter 1, AHRQ reviews the context for registries today within the notion of a learning health system, and then quickly jumps into a useful discussion of interoperability barriers and problems. Many research-oriented registries are narrowly focused on a specific issue or problem; AHRQ envisions a more interoperable set of registries that can create a more integrated "national research infrastructure."

The report continues with a great discussion of data sources for registries (Chapter 2) and data standards relevant to registries (Chapter 3) before a much more focused chapter on obtaining data from EHRs (Chapter 4). This chapter is one of the gems of this report, with a good summary of the challenges and opportunities in relying on EHR data. I really like the three sample use cases and architectures they summarize within this chapter, as well as a discussion of technical issues and operational challenges. The report concludes with a short chapter on getting data from sources other than EHRs (Chapter 5) including a small section on application programming interfaces (APIs) and common data models.

Throughout my exposure to this AHRQ material, I have never fully understood whether they consider public health registries (my bread and butter) to be relevant to this discussion. In past works, they have only mentioned them in passing, spending most of their time focused on research-related registries. I read this new draft report, as usual, keenly sensitive to this issue and curious about how they would treat public health registries. I was pleasantly surprised by the more balanced inclusion of these systems in various descriptions of registry roles and issues.

For example, in Chapter 2 on Data Sources, there is a section of Health Information Exchanges. Within that section is a nice discussion of the collaboration between HIEs and public health agencies regarding access to public health registries. The chapter provides some good examples of how HIEs can be leveraged to provide transport for public health reporting as well as be useful in identifying patients with specific conditions.

Chapter 4 on Obtaining Data from EHRs opens with a solid recognition of the role of the Centers for Medicare and Medicaid Services (CMS) Promoting Interoperability (PI) programs which have incentivized providers and hospitals to participate in reporting to public health registries. I was thrilled to find a sample use case titled "EHR-Linked Public Health Registries" summarized along with other use cases in a handy table. Some of the details were not quite right (for instance, an assumption that transactions related to Meaningful Use were necessarily fully-automated, but other transactions were not; and an indication that public health registries commonly have federated or decentralized architectures when centralized architectures are the overwhelming reality).

But the ensuing discussion about EHR-integrated registries and support for clinical care does not mention, for instance, the role of immunization registries in providing immunization histories and clinical decision support directly to the point of care through registry query/response from EHRs. The report provides a separate section on EHR-Linked Public Health Registries but this section like the previous one only addresses data flowing from EHRs to registries, not the flow of data from registries to EHRs.

One more point is worth mentioning. Though not yet finalized, this document makes no mention of the Office of the National Coordinator for Health Information Technology (ONC) Trusted Exchange Framework and Common Agreement (TEFCA) or the draft U.S Core Data for Interoperability (USCDI). This becomes most relevant in the section in Chapter 5 on Common Data Models which seems incomplete without reference to this emerging 21st Century Cures Act initiative or the Common Clinical Data Set (CCDS) which preceded it.

So, my primary recommendations to AHRQ include:

  1. Clarify up front the scope of this Guide, especially how "registry" is defined for the purpose of understanding the context for the document even though this is part of a larger series of guides where this definition may or may not be clearer.

  2. Engage with public health agencies and member organizations to further clarify and update sections of the Guide that describe public health reporting to make sure the most current and accurate information is available to readers.

  3. Update the document to discuss, even briefly, current ONC initiatives like TEFCA and USCDI that will invariably have an impact on registries and interoperability in the coming years.

Comments on this draft report can be submitted to AHRQ by February 8, 2019.