CHICAGO (GenomeWeb) – Data integration and interoperability have been longstanding challenges in the world of electronic health records. As the 21st Century Cures Act requires, officials from both the Obama and Trump administrations have been less than subtle in calling out vendors and health systems alike for "information blocking" on the clinical side, so integrating genomic data hardly seems like a priority across the industry yet.
Still, even one of the pioneers of the EHR field has begun merging pharmacogenomics into clinical decision support and has plans to expand its genomics capabilities.
Meditech may have been founded in 1969 by one of the developers of the Massachusetts General Hospital Utility Multi-Programming System (MUMPS) — an oft-derided language that still underpins some of the technology in modern EHRs — but the vendor also has a senior leader who champions precision medicine.
Hoda Sayed-Friel, executive vice president of Westwood, Massachusetts-based Meditech, has had an interest in pharmacogenomics for at least five years, well in advance of the launch of the Precision Medicine Initiative.
That interest has shown up in Meditech's EHR in the last year when, at the urging of Meditech, formulary services provider First Databank added drug-gene interaction checking to its engine that powers clinical decision support for many EHR vendors.
Meditech's EHR can now scan prescription orders for potential drug-gene interactions and alert clinicians of any issues during the computerized physician order-entry process.
For example, if a doctor is looking at a gene that controls drug metabolism and the test shows an issue with CYP2C19, the Meditech lab results dictionary will be able to understand that the patient has slow metabolism. So, if the physician orders antiplatelet drug clopidogrel (Plavix), for example, the system can warn of a potential metabolism issue and suggest a different anticoagulant, Sayed-Friel explained.
Genetic testing comes into play for certain conditions even before the drug order goes in.
"Part of what we're doing along with clinical decision support is that the cardiac stent order set will have a recommendation in that order set that you should do CYP2C19 genetic marker testing before that patient goes in for that procedure," Sayed-Friel said. "If you don't order it, you will get a warning."
When the gene test result comes back, the system will alert the clinician whether it is safe to proceed with Plavix or pick another drug.
Even if the patient needs an emergency stent, the system can query whether the person has had genetic testing performed, Sayed-Friel noted. The EHR has surveillance tools on a dashboard that displays all patients on Plavix or who have had the CYP2C19 test. "We do that already. That's part of our system," she said.
Given all the interoperability issues that exist with EHRs, the pharmacogenomic testing is only as good as the available data integration.
"The execution of the rules is already in the Meditech EHR," Sayed-Friel said. "The feeding of the data to drive that rule will depend on whether that data resides in that lab database or if we have to get it from a genomics lab database."
This is easy to do when the care provider owns or has ready access to the lab database. It is more difficult when genetic information is stored elsewhere, Sayed-Friel said.
Sayed-Friel is a member of a National Academy of Medicine workgroup for genomics and EHR integration known as DIGITizE, which stands for Displaying and Integrating Genetic Information Through the EHR. The group is made up of geneticists, representatives of EHR vendors, and standards organizations.
"All of those different participants try to figure out from that workgroup how to get clinical decision support with genomics into EHRs," she said.
"What we need from the EHR side is if a clinician is going to provide care to a patient where a particular gene marker is critical to providing the right care, we want to be able to ask for that information if we don't already have portions of it in our system, and ask for it in an objective way so we can run it through our algorithms," Sayed-Friel said.
That means it should come in not as a report, but with the condition of the gene marker itself so the vendors can run it through a clinical decision support engine, she explained.
Meditech can query genomic databases several ways, depending on whether the hospital has a genetic-testing lab on site (as community hospitals tend to send out their tests).
"We could set up a point-to-point interface with some [Health Level Seven International] specifications to get that data," Sayed-Friel said. The first iteration of this was a standard HL7 communication with a protocol that includes the gene marker tested, the result, and the interpretation. Soon, the group will incorporate the Fast Healthcare Interoperability Resources (FHIR) Genomics specification, an HL7 standard that remains under development.
As for Meditech, its ambitions in genomics remain modest. Sayed-Friel said that the company chose to start with pharmacogenomics because it is not controversial. Soon, the vendor will look at genomics in oncology.
"The area that we're not touching is the genomics of disease probability," Sayed-Friel said. "That's too controversial. It's not an EHR thing. It's a physician-patient counseling thing."