NEW YORK (GenomeWeb) – A new study this week has added mixed results to the evidence base for whether using pharmacogenomic information and associated decision support tools can help stave off adverse events and drug interactions for patients taking multiple medications.
In the study, which analyzed data from a clinical trial funded by Genelex and appeared this week in the Journal of Managed Care & Specialty Pharmacy, investigators randomized a cohort of about 350 polypharmacy patients eligible for medication therapy management (a service provided for patients who are expected to be at a high risk of adverse drug reactions or interactions).
Under Medicare, certain plans are required to offer MTM programs for certain beneficiaries, including those with multiple chronic conditions, taking multiple covered medications, and/or likely to incur annual medication costs that exceed a certain level.
In the study published this week, researchers from the University of Utah and management firm Magellan Health looked at differences between three trial intervention arms: standard MTM, MTM that also incorporated decision support via YouScript (which Genelex spun out as a separate company in 2017) and decision support plus PGx test results for CYP450 and VKORC1 enzymes.
Jonathan Magness, currently senior director for Medicare Part D clinical pharmacy operations at Magellan, said in an interview that the initial trial was conducted by Genelex and a small pharmacy benefit manager in Utah called Veridicus Holdings, which was later acquired by Magellan.
"They said they would donate … tests and we could do what we wanted with them and see what the outcome is," he said.
The team recorded the number of adverse reactions or drug interactions in each arm of the study and compared them. They also specifically measured the severity of these "drug therapy problems" (DTPs), comparing how many serious (grade 3 and above) versus lower grade instances were noted for the three groups, as well as how often a DTP — serious or minor — was accepted for use in managing prescriptions going forward.
"At the conclusion of the study we actually sent all of our recommendations to a couple of clinical pharmacists who also work in this in the same area to evaluate each of our recommendations and give it a numerical score based off of how they felt the outcome of that recommendation could have played into the life and the healthcare of that patient," Magness explained.
An initial stumbling block in the study was the fact that there was a high dropout of patients from the PGx arm, with half failing to even return a buccal swab test for genetic analysis. This meant that researchers had to do a post hoc analysis rather than an intention-to-treat analysis, reassigning individuals who failed to provide the PGx samples to the non-PGx arm.
Also disappointing for PGx advocates was the that there was no real difference in the rate of DTPs identified between the three arms.
However, investigators did conclude that the proportion of serious DTPs identified, and the rate at which these were accepted for future avoidance, were significantly higher in patients undergoing PGx testing.
In other words, blinded pharmacists identified more serious problems that could crop up for patients when they had PGx results available than they did in patients without those results. They also identified more PGx-based DTPS as serious— 13 of 42 or about 30 percent — versus only 5 percent of the non-PGx-related DTPs.
According to the authors, this suggests that "genetic profiles of patients help capture potential problems that would not be identified through standard MTM programs," and that the "increased seriousness of DTPs identified through the utilization of PGx test results demonstrates a patient-safety opportunity that is currently being overlooked in many MTM programs."
Considering that serious DTPs were more likely to be accepted than non-serious findings, the fact that PGx led to more of these findings would imply that patients should see improved outcomes, as prescribers put into use the identified proscriptions.
However, the team hasn't yet done a direct analysis of outcomes or health economics, though Magness said that this was initially planned as a second phase of the study.
The study is not the first that Genelex has sponsored to try to establish the value of its testing. The firm completed a study called IMPACT in 2015 which compared 200 prospectively enrolled Medicare-eligible patients and 800 historically matched subjects, but at the time, this failed to sway Medicare toward embracing its testing holistically, rather than only for specific indications.
In another trial published in early 2017, investigators randomized approximately 100 patients discharged from a hospital home health agency and studied their health utilization at different timepoints.
In that study, researchers found that 60 days after discharge from the hospital, patients in the PGx testing group had 52 percent fewer re-hospitalizations and 42 percent fewer emergency room visits compared to controls. Of the 124 PGx-guided treatment recommendations pharmacists relayed to physicians, doctors followed 77 percent.
But that also does not seem to have persuaded payors. On its website, the company currently calls both private insurance and Medicare coverage "limited."
PGx companies have struggled across the board to gain wide clinical adoption. One issue, according to speakers at the Association for Molecular Pathology annual meeting earlier this year, is that there is a catch-22 in trying to make a case for using assays as part and parcel of medication management rather than paying piecemeal for specific drug-gene pairs.
To pay for systemic, preemptive PGx, insurers want to see proof that it impacts patient outcomes in a cost-effective manner. But measuring that impact is hard or impossible when PGx testing isn't implemented in way that is truly, or holistically integrated, which it can't be if insurers don't pay for it.
In the new study, Magness and his coauthors wrote that at the time of the trial, most prescribers would be expected to be unfamiliar with PGx, and thus reluctant to accept the validity of such a recommendation at face value.
This may have played in to the overall low rate (30 percent) at which the researchers saw DTPs actually lead to a therapeutic change for patients.
"Ultimately, assessing clinical utility is limited by the level of familiarity of the prescribers receiving MTM recommendations. PGx familiarity at the time of this study was rather low and, as a result, may have affected the level of acceptance of DTP recommendations," they wrote.
According to Magness, it becomes even more complicated because even if PGx is embraced in MTM, it also must be consistently treated by physicians or other healthcare providers for it to translate into improved patient outcomes.
Anecdotally, he said that he himself participated in an employer-sponsored genetic testing pilot recently, in this case not with Genelex, but with another PGx company.
It turned out that he was on a medication for which his genotype implied little to no efficacy. "I went to my physician with that information and successfully got it switched," Magness said. "But when I went in there with those results, even with my background and my knowledge and being able to explain that to my primary care provider, he was still very confused as to what it was I was even really referring to."
"And then even when he understood what I was saying… when he went into the medical record to try and input that I was a CYP2C19 ultra-rapid metabolizer there's nowhere in the software for him to put that in that's meaningful," and that would provide future alerts, he added.
Genelex spun out its analytics and clinical decision support tool YouScript as an independent business at the time that it published its prior 100-patient study in 2017, hoping that as an independent company it could focus on addressing the systemic barriers to pharmacogenetics adoption in medical care and creating architecture to integrate PGx into electronic medical records and the healthcare workflow of doctors and pharmacists.
And the company has been continuing those efforts, most recently announcing a program with Highmark Health in which the healthcare system will evaluate the YouScript Precision Prescribing Software, integrating it into the EHR system for 3,000 multiple-medication patients from Allegheny Health Network affiliate Premier Medical Associates.
YouScript has also had a deal with Epic Systems through which healthcare systems and providers who use Epic EHRs can license its tools and use them for decision support including PGx integration via Genelex or from other companies.
"When I put on my PBM hat, what we are thinking about on the payer side is how can we advance this," Magness said. "We've had some internal discussions within Magellan to say, look, we see that there is a potential benefit, but how do we identify those members that are the most poised to benefit and how do we get that pharmacology information back into our software instead of relying on the physicians' offices to do this sort of analysis."
Magness also suggested that future studies may benefit from a tighter focus on populations in which a utility and economics case can be made more clearly. "One of the weaknesses of [our] trial was that we took that MTM population and we didn't try to identify a subset of the population that may benefit the most."
"If we were to have preemptively looked at the claims data that was coming in and said OK here's a group of members who have x-amount of drugs that are processed through the CYP450 system … and then did a randomization to treat versus not treat, I think we would have seen some much better outcomes and better statistical value," he said.
Speaking earlier this year at the AMP meeting, Coriell Life Sciences President and CEO Scott McGill discussed how his company is tackling the challenge of cementing PGx clinical utility within its model of "enterprise pharmacogenomics," in ways that are more custom-fit to persuading payors.
In one case study, he highlighted work with a pension fund of about 34,000 retired teachers. On average, Coriell found that members were on 15 medications each, and calculated initially, based only on variant prevalence measures from population data, that about 83 percent of the group would be expected to be taking at least one drug with "clear, credible" evidence for a genetic proscription.
The company has now actually tested about 2,100 individuals from the cohort since January and is seeing a reduction of both medical issues and healthcare costs versus a non-tested control group, he said last month.
Other companies are also trying to collect the right kind of outcome evidence to persuade payors. Genomind has begun to amass evidence demonstrating the ability of its Genecept test to reduce utilization of healthcare services and related costs among patients with mental illness, including an 800-patient study published earlier this year in Depression and Anxiety in which researchers concluded that PGx-tested patients experienced 40 percent fewer emergency room visits and 58 percent fewer hospitalizations compared to the controls.
Optimistically, Magness said that the way PGx testing prices have come down in recent years has made all of this a much more feasible discussion "whereas a couple years ago it was a no-go across the board."