Informatics Professionals See Potential for New Paradigm for Our National Health IT Infrastructure
(BETHESDA, MD) – In a pair of comments filed today with the Office of the National Coordinator for Health IT (ONC), the American Medical Informatics Association (AMIA) supports the federal government goals to harmonize network-to-network exchange and establish a new mechanism to develop and disseminate national data policy. The nation’s informatics professionals urged federal officials to develop a comprehensive timeline, which includes establishment of pilots for specific aspects of the TEFCA, and they challenged ONC to establish a “share first, structure later” data policy, so that important data elements are not trapped behind siloed systems.
The ONC Trusted Exchange Framework & Common Agreement (TEFCA) and its US Core Data for Interoperability (USCDI) are two sets of draft policies developed by the ONC under direction from the 21st Century Cures Act of 2016. They are intended to enable: nationwide data exchange among clinicians using different exchange networks; comprehensive access for individuals who wish to use their own health data; bulk record exchange for accountable care organizations; and support for an emerging landscape of open and accessible application programming interfaces (APIs). The TEFCA outlines the legal obligations of stakeholders engaging in such activities, and the USCDI provides the underlying data policy by identifying priority data classes and establishing a process/timeline to make available such data classes for exchange through the TEFCA.
In comments submitted for the TEFCA, AMIA recognized the lack of harmonization among current networks, causing clinicians to rely on point-to-point exchange that is a marginal improvement over fax machines. The group noted that the current paradigm is costly to operational IT, increases the burden to clinical practice, and falls short of the promises of digitized patient care.
“At a national level, the health system simply cannot optimize nationwide investments in health IT based on push-only exchange and contractual relationships that must be created anew for each step in the referral chain. In tandem, the TEFCA and USCDI represent a new paradigm for our national health IT infrastructure,” AMIA said in comments.
But moving to a new paradigm will take research, experience, and a thoughtful approach. AMIA recommended ONC develop a roadmap that details an implementation plan at least three years into the future and ONC should establish specific pilot tests to inform this implementation plan. AMIA also urged ONC to seek feedback from the wider stakeholder community as the roadmap is implemented, and they requested that the lead coordinating body, referred to as the Recognized Coordinating Entity, or RCE, be required to hold open comment periods, listening sessions, and other accountability mechanisms.
AMIA noted that fulfillment of the TEFCA provisions will take several years to be operational across a majority of stakeholders, and that a more detailed timeline and implementation plan will help stakeholders better understand how the TEFCA and USCDI are meant to be operationalized.
“While these goals are laudable, the details of these policies will require further refinement and significant time to be operationalized,” AMIA said. “Specifically, we anticipate that the TEFCA requirements to enable bulk transfer and individual access will require much more work to develop consensus on executable solutions. We also anticipate that more work and experience will be needed with consent management of the kind required by the TEFCA. Ultimately, these early networks will form the foundations for a learning health system. We need to be thoughtful in how we establish this foundation.”
AMIA identified the component pieces of (1) individual access, (2) bulk transfer, and (3) consent management as important aspects of the TEFCA that needed further refinement. “[These] tests [should] be clinically and geographically relevant, which means including representative care settings (e.g. community hospitals, health systems, SNFs, ambulatory, LTPAC, etc.), reflect patient care patterns (rather than two leading academic medical centers on the east and west coast), and capable of informing other use cases like bulk transfer to support data science, analytics, and population-level research. It is important to reiterate that proofs of concept are insufficient. These tests must answer specific questions, or establish specific processes, meant to inform the overall strategy.”
While supportive of the USCDI, AMIA said in companion comments that standardization of data should not be prerequisite for data exchange. Rather, AMIA recommended ONC finalize the USCDI Version 1 with a policy that (1) enables the sharing of all data classes in computable and human readable (e.g. free text) formats, and that (2) gradually structures data classes over time. A “share first, structure later” policy would allow all data classes not yet ready for inclusion in the USCDI to be exchangeable – even if it is only exchangeable as free text. To support a “share first” functionality, AMIA pointed ONC towards current and emerging Health Level Seven (HL7) standards, including the HL7 CCDA Unstructured Document, document-level template, which provides a mechanism to share unstructured data from documents, and/or several emerging FHIR® resource for more granular data, to support this policy proposal. AMIA provided a graphical representation of their policy alternative, available here, that would enable sharing of data classes without the need to have them structured first.
“Ultimately, we want a network that builds toward a learning health system – one that lets every interaction with a patient become an opportunity to learn. That goal will take time, so an incremental approach in which we ‘share first, structure later’ is an important first step,” said AMIA President and CEO Douglas B. Fridsma, MD, PhD, FACP, FACMI. “Building on pilots, learning from experience, and creating an open and transparent governance process will help us develop a network that meets our needs now and can evolve over time.”
AMIA also noted that data collection should be considered an intervention (similar to other tests or procedures) that has both value and costs. “The value of data collection is well-articulated for quality, safety, and other purposes, but the cost of data collection – in terms of time, money, and resources – is not,” AMIA said. “Recent studies have articulated the costs of certain activities, such as generation of clinical quality measures using EHRs, but a more generalizable metric must be developed.”
AMIA recommended that ONC work with partner agencies, including the Agency for Healthcare Research & Quality, the National Library of Medicine, National Institutes for Standards and Technology, and CMS, to develop a generalizable metric that captures both the value and the cost of data collection and identifies ways to leverage this metric in the USCDI, quality measurement, and other data collection activities.
“In the same way that CMS and other payers rely on quality-adjusted life-years (QALY) or Number Needed to Treat (NNT) to examine specific interventions, we must develop a similar measure to capture the cost-effectiveness of collecting electronic data,” AMIA wrote.
AMIA recommended ONC develop a scorecard for USCDI data classes that includes an estimate of a “cost to collect,” leverages important prior work to develop an objective framework, and includes aspects of AMIA’s Health IT Data Standards & Interoperability Policy Principles.
AMIA’s full comments on the USCDI can be found here, and TEFCA comments can be found here.
###
AMIA, the leading professional association for informatics professionals, is the center of action for 5,500 informatics professionals from more than 65 countries. As the voice of the nation’s top biomedical and health informatics professionals, AMIA and its members play a leading role in assessing the effect of health innovations on health policy, and advancing the field of informatics. AMIA actively supports five domains in informatics: translational bioinformatics, clinical research informatics, clinical informatics, consumer health informatics, and public health informatics.