NEW YORK – Researchers at the US Department of Veterans Affairs are hoping to make it easier to incorporate results from its pharmacogenomics (PGx) program into patient records, while also studying reasons providers don't always follow the resulting recommendations.
Rona Relova, a research health scientist at VA Palo Alto Health Care who studies precision medicine, is heading up this effort through a fellowship awarded by the Veterans Health Administration's Office of Healthcare Innovation and Learning that enables her to investigate how to improve the return of results for the VA's National Pharmacogenomics Program (NPP).
The NPP builds off of the pharmacogenomics testing for veterans, or PHASER, program, a PGx testing pilot that initially launched at the Durham VA Medical Center in 2019 and has since expanded to additional VA sites with the Durham facility continuing to serve as the coordinating center for the program. The NPP is expanding the PGx program across the VA.
As part of the program, providers at certain VA medical facilities can order PGx testing for their patients within the electronic health record (EHR) system, have blood samples collected from patients, and send the blood draws to third-party laboratories for testing.
Relova was excited by the prospect of getting precision medicine into actual patient care.
However, she was dismayed that PGx results typically were returned from labs in a PDF file, so the data weren't easy for physicians to review or enter into a patient's medical record.
"It cannot integrate with our healthcare records," Relova said. That can pose challenges for physicians who might not realize that PGx results are available for their patients or may not have time to hunt down a PDF and dig through results for each patient before making a prescribing decision. This challenge, she added, isn't unique to the VA.
Relova is working to refine and scale the PGx program, including incorporating a clinical decision support system to guide prescribing decisions based on testing results. She's conducting this work as a senior fellow within a program at the VHA Innovation Ecosystem, which gives leaders at the VA time and resources to work on projects to improve clinical outcomes, healthcare costs, and veteran experience.
Currently, the process to get PGx results from a PDF and into a patient's chart is manual, as the data has to be entered by a technician. That means there's a lag between when results are returned by the lab and when they're available in the EHR, so may not be ready when physicians are making prescribing decisions. Relova's goal is to automate this process.
Earlier this month, her team installed the application they've been developing to make results available to providers more quickly at three initial VA sites in Charleston, San Francisco, and Palo Alto. The application is built using the VA's Arches data platform to upload PGx data from labs directly into the EHR, which are made available to providers as unsigned clinical notes. Once providers sign and review the data, they become available for patients to view within the EHR patient portal.
Relova and her colleagues will be tracking whether this application affects how PGx is used by providers.
Still, even without automation, the VA had observed early adoption of the PGx test.
The NPP, which has so far launched at about 140 sites of the VA's roughly 170 medical centers, analyzes about 20 genes with implications for nearly 90 commonly prescribed medications within cardiology, infectious disease, oncology, pain management, psychiatry, and other specialties. Physicians have ordered the PGx panel for patients more than 82,800 times.
Physicians receive a report in the patient's medical record detailing their PGx results, including the genetic variants detected and what that means for the patient's ability to metabolize various drugs.
Once data are entered into a patient's medical record, if a provider orders a drug that a patient is unlikely to respond to based on PGx, an alert will pop up within the EHR with recommendations on dosing considerations or alternative drugs, with citations of where the recommendation is from. For example, an alert might cite information from the US Food and Drug Administration or guidance from the Clinical Pharmacogenetics Implementation Consortium (CPIC), an internationally recognized guidelines body.
"It takes out the one-size-fits-all approach [to medicine]," Relova said.
There are also designated pharmacists as part of the NPP that providers and patients can connect with for support understanding PGx test results.
Between July 2021 and February 2024, more than 22,000 PGx alerts related to nearly 6,500 patients fired within the clinical decision support system, according to a poster presented at the Precision Med Tri-Con conference in March.
Providers prescribed drugs that were concordant with PGx recommendations about one-quarter of the time or canceled medication orders in a way that was concordant with the recommendations about 35 percent of the time. Interestingly, 38 percent of the time, or in 8,174 instances, providers ordered in a way that was discordant with the recommendations.
That's an area of ongoing study, Relova said. Her team is continuing to collect data on concordant and discordant decisions and patient outcomes based on those decisions and is planning to prepare a manuscript on their findings. They are aiming to publish that paper early next year.
In some cases, prescribers were prompted to select a reason for overriding a PGx recommendation. The most common response, selected in 41 percent of cases, was that it was a refill of a medication that a patient was already taking, followed by physicians stating the indication was justified (34 percent) or that they intended to monitor the patient for issues (9 percent).
These trends also may change as providers get used to the PGx program and prescribe new drugs, and as the program expands into new sites.
"As we get more numbers, the trends are changing," Relova said.