A big shift is under way in the world of standards for electronic clinical quality measures (eCQMs). The HL7 FHIR (Fast Healthcare Interoperability Resources) standard is at the heart of a two-phase transition to new standards that seek to harmonize eCQM and clinical decision support (CDS) development.
The shift will affect vendors, developers, implementers, and quality reporting agencies, and while the transition may be costly in the short term, the long-term benefits should greatly outweigh the initial cost, said KP Sethi, senior software architect at Lantana Consulting Group and co-chair of HL7’s Clinical Quality Information workgroup, in a recent webinar presentation.
The new standards should allow developers to automate and simplify many manual tasks. Other benefits of the transition, he said:
• Reduces implementation burden and learning curve;
• Reduces long-term maintenance burden for measure developers and decision support rule authors so they can share and across their domains; and
• Encourages sharing of artifacts across several domains.
Sethi used the following example of a clinical decision support guideline and a quality measure to make his point:
• Clinical Guideline: Patients with ischemic stroke and history of atrial fibrillation or atrial flutter should be prescribed anticoagulants
• Quality Measure: Percentage of patients with history of Afib/Aflutter, with diagnosis of ischemic stroke that were given anticoagulant at discharge
He noted that there are similarities in the data elements being used. There is a diagnosis common to both. The fact that patients should be given an anticoagulant at discharge is common. One is telling the clinician to do what the guideline is suggesting and the quality measure measures whether they did it or not. They are driving toward the same guideline in different ways, he said.
“However, in today’s standards we cannot describe those data elements in ways that can be commonly used by the quality measure and decision support rules,” he said. “That is because the quality measure standard and the CDS standard really have nothing in common. Both developed in silos and have no overlap, even thought they talk about almost the same thing. So the goal with the new standards is to develop a set of shared standards that can be used by both CQM and CDS. Those common standards include a common data model and a common expression language.”
There were other influences on the need for new standards. One of those is that the current standards are based on HL7 version 3, which didn’t really have a whole lot of adoption. “That is because version 3 was found to be too complex and lacked the ease of implementation that was necessary,” Sethi said.
At the same time, HL7 FHIR burst onto the scene and is simpler to implement. “Because of its implementer-friendly design, the industry is rapidly shifting toward FHIR and away from HL7 version 3,” he said. “Besides those influences, we had a recognition of a need for shareable, executable artifacts between CDS and CQM that would reduce developer burdens, and maintenance. This could increase the accuracy of the logic in authoring.”
Phase 1 of this shift, happening between the third quarter of 2016 and the third quarter of 2017, involves a new Common Expression Language called Clinical Quality Language (CQL), which replaces Quality Data Model (QDM) Logic.
Here are some of the benefits of CQL:
• Replaces QDM logic expressions
• Supports logic reuse and sharing
• Data model agnostic
• Provides both author-friendly and machine-friendly syntax
• Automatic transformation to machine-friendly syntax
• No need for manual mapping (unlike QDM/HQMF)
Phase 2, which does not have a proposed timeline yet, would move to a new common data model. That model, called Quality Improvement and Clinical Knowledge (QUICK), replaces QDM Data. It harmonizes data models used by CDS and CQM, and unlike the QDM, contains data elements only (no logic).
These two are based on a Common Backbone of FHIR. There is a FHIR Measure Resource, which represents an eCQM in FHIR and replaces Health Quality Measures Format (HQMF).
The FHIR Measure Resource represents an eCQM in FHIR syntax. It contains:
• Metadata of the measure
• The data elements need by the eCQM
• The logic of the calculation
Compared to current standards:
• Replaces HQMF
• Benefits from FHIR syntax (simpler, easier to implement)
• Supports an operation to execute the measure
The FHIR Measure Report Resource represents quality reports in FHIR and Replaces Quality Reporting Document Architecture (QRDA), a Clinical Document Architecture (CDA)-based standard for reporting patient quality data for one or more quality measures.
The phase 2 timeline would depend on the success of phase 1, but also on maturity level and adoption of FHIR across the industry, he said. “A lot of the standards we talked about use FHIR as the backbone. Depending on how quickly the industry moves toward FHIR, the timeline could be pushed up or pushed out. It also depends on the readiness of FHIR. FHIR is a new standard. It has been around for four or five years. A lot of it is not mature enough to be put into production yet. IT hasn’t gone through a normative release.”
On the quality measure reporting side, Sethi pointed out that phase 2 will require that validation tooling be updated and vendor software exports will need to produce FHIR instead of QRDA. But the benefits include:
• Simpler measure reporting format vs. QRDA.
• Reduced data validation burden on agencies like CMS.
• Reduced burden on standards developers and maintainers.
• Reduced reporting burden on implementers.
• Move to common FHIR-based standards consistent with industry.
Sethi closed by saying he was optimistic that these standards will give the industry stronger tools to improve delivery of healthcare. “We have learned a lot in the six years that eCQMs have been under development. We have made great strides. These tools are a further step in the right direction to improve the delivery of healthcare.”
