NEW YORK – Electronic health records vendor Epic Systems recently saw the first EHR-to-EHR exchange of structured genomic data on its Care Everywhere interoperability platform, heralding a step forward in pursuit of precision medicine, according to Catherine Procknow, a senior software developer at Epic who largely specializes in genomics.
"With that structured, computable data, you're actually able to make that result report actionable, especially downstream so you can build decision support into Epic so that you can provide prescribing support down the line," Procknow said during a meeting at Epic headquarters in Verona, Wisconsin. "Now, a patient's genomic data can follow them wherever they go to receive care, just like every other piece of data."
A company spokesperson said that "at least 19" Epic customers now incorporate structured data in variant results reporting from in-house labs. Nine commercial genetic labs are sending structured genetic data to institutional Epic users, the majority to more than one health system. The company did not name any of these users.
This activity is the result of the May 2022 launch of a dedicated, cloud-based sequence server to store large genomics files as part of a regular update of the core EHR. The sequence server is the product of at least four years of planning and development.
Epic updates its EHR quarterly, and Procknow said that institutions tend to upgrade twice a year, though there is often a lag of up to a year between a new release and customers switching to that version.
"As soon as groups began adopting the version we released [the sequence server] in, we saw this exchange [of structured genomic data] happening," Procknow said. "You don't need to have gone through a genomics module install to actually exchange this data" because the Care Everywhere interoperability platform supports it.
The sequence server is not a single database, but rather a platform of which each customer has its own instance. However, it is all hosted on Epic's Nebula Cloud, which sits in the Microsoft Azure cloud environment. Integrations can be with laboratory information systems either in commercial testing facilities or at in-house labs.
Currently, the sequence server is storing only VCF files. The Epic EHR keeps track of the relationship between test orders in patient records and corresponding files in the sequence server. Since VCF files usually do not contain patient identifiers, data in the dedicated server is also suitable for research use, Procknow said.
Accompanying reporting tools allow users to search both the EHR and the sequence server for VCF files of patients with a common diagnosis or phenotype, regardless of which lab performed the sequencing.
She said that the company wants to support emerging multiomic technologies as well to help healthcare providers leverage such data for clinical care in the future.
With current capabilities in place, Epic now wants in particular to be able to support genome interpretation and reinterpretation and is building tertiary analysis tools and workflows. "I think reinterpretation workflows is a … case where we can provide a lot of value," Procknow said.
Epic is starting with interpretation workflows, taking VCF files into the EHR, then outputting results reports directly into each patient's record.
Procknow said that the company eventually wants to store up-to-date versions of knowledgebases such as ClinVar on the sequence server, alongside the VCFs and lab reports that are there now. "We're working on workflows that connect the variation that can be contained in VCFs and other sequencing artifacts and those knowledgebases," she said.
Those workflows include automating tertiary analysis and ways to flag results for reinterpretation. "[We will be] building tools to help with underdiagnosed and misdiagnosed rare diseases, diagnostic-odyssey support, and also generating risk score algorithms" and even performing genome-wide association studies, Procknow said.
Reinterpretation would be helpful when patients present with new symptoms or as children develop.
Procknow said that Epic seeks to apply genomics to support diagnoses of rare diseases, particularly in neonatal intensive care units. "The idea is to help shorten the diagnostic odyssey with this type of integration to support rapid whole-genome sequencing in the NICU and other emerging sequencing use cases," she said.
There are a number of groups studying rapid WGS for newborns, and some, including the Rady Children's Institute for Genomic Medicine and Genomics England, are pushing for the practice to become universal. Rady Children's Hospital-San Diego is an Epic user.
In the US, Epic tends to have some of the largest integrated healthcare systems, including academic ones, as its customers. The vendor also has agreements with genetic testing companies including Invitae, Myriad Genetics, and Foundation Medicine to integrate test results into EHRs.
Procknow said that all the data flow happens behind the scenes thanks to "native" integration into the core Epic EHR as well as ancillary systems like the company's Aura specialty diagnostics suite.
Epic continues to tout standards-based data exchange, and the firm became a corporate member of the Global Alliance for Genomics and Health a year ago. The company is also heavily invested in Health Level Seven International, though it had to be cajoled by dozens of users to adopt elements of Minimal Common Oncology Data Elements, specifications from the American Society of Clinical Oncology, Mitre, and the Alliance for Clinical Trials in Oncology Foundation.
She said that the "next frontier" is to support genomics integration with users of other EHRs, though Epic has been the subject of criticism for years that it is not fully committed to interoperability with other vendors. The company has been trying to shake that perception.
Procknow said that Epic is "especially interested" in standards that can introduce genomics into clinical decision support, including by connecting genotypes and variant lists to existing medical knowledge. Expressing genotypes in EHRs has been an ongoing struggle in the standards world.
"As new testing methodology becomes a part of practice, as new types of variants need to be represented, the standard bodies and us work to incorporate that into our data models, into our software, so that it can be used just like any other piece of patient data," Procknow said.