NEW YORK (GenomeWeb) – Phamacogenetic testing firms have been hit particularly hard when it comes to reimbursement.
One such firm, Genelex, saw its lab testing revenue slashed by as much as 70 percent after Medicare contractor Noridian issued a local coverage determination (LCD) several months ago choosing to cover only a handful of gene-drug indications. Genelex markets a pharmacogenetic test panel that gauges a range of CYP450 and other genetic markers known in the medical literature to be associated with drug response. Many of these associations are noted in guidelines and are described in drug labels that have been updated by the US Food and Drug Administration.
Genelex is heavily dependent on Medicare reimbursement, since the elderly tend to be patients who take multiple drugs at once, which in turn increases their chance of adverse events from drug-gene and drug-drug interactions. Medicare beneficiaries stand to potentially benefit the most from PGx-guided treatment decisions, according to Genelex, because elderly patients using its tests take 11 drugs concurrently on average. Medicare covers specific PGx indications, but generally does not pay for multi-marker panels like the ones Genelex and many other companies are selling. One exception is Assurex's GeneSight Psychotropic PGx panel for guiding depression treatments, which received Medicare coverage last year.
Hoping to improve the reimbursement prospects for its PGx panel, which has a list price of around $900, researchers from Genelex and the University of Utah recently published a cost-effectiveness and clinical utility study in the Journal of Medical Economics. In that trial, called IMPACT — involving 200 prospectively enrolled Medicare-eligible patients and 800 historically matched subjects from a claims database — researchers showed that patients who received PGx testing had fewer hospitalizations and emergency room visits than patients in the comparison cohort. Using mean cost estimates, researchers led by the University of Utah's Diana Brixner estimated potential savings of $218 per patient who received PGx testing.
Despite these findings, IMPACT did not impress Medicare contractors, Genelex told GenomeWeb last week. The company was hoping that IMPACT would convince Medicare to pay for testing in the context of a randomized-controlled trial.
"We sent results from this study to Medicare after it came out but they don't seem to understand the clinical utility of CYP450 testing," Genelex CEO Kristine Ashcraft said. "Payors are married to a randomized-controlled trial for every disease state and for every drug-gene combination, maybe because this is what they're used to in drug development. But they're missing the fact that this is next-generation pharmacokinetics."
In certain healthcare sectors, PGx testing is not only happening but taking off. At least two large academic healthcare providers — Vanderbilt University and the University of Florida — are expanding their pilot programs after successfully implementing multi-marker PGx testing in a preemptive fashion.
But these programs within leading academic centers haven't convinced payors to cover PGx testing more broadly throughout the nation. Even when gene-drug associations find their way into FDA labels and guidelines, payors are looking for specific assurance of medical necessity. And this language of medical necessity has been challening for molecular diagnostics firms to translate.
"This is a field where the science has outpaced its clinical applications," said Girish Putcha, who is the director of laboratory science at Medicare contractor Palmetto's MolDx — a program that identifies tests, evaluates evidence to determine analytical validity, clinical validity, and clinical utility, and determines coverage and reimbursement. "Just because we can doesn't mean we should," said Putcha, who spoke to GenomeWeb as an industry observer and not on behalf of CMS or Palmetto.
"Labs and companies have in general not been willing to step up to the plate and do properly designed clinical studies to prove clinical utility for their test or test service for specific clinical applications," he said.
Just because we can doesn't mean we should.
A few months ago, Medicare contractor Noridian issued an LCD for CYP2C19, CYP2D6, CYP2C9, and VKORC1 genetic testing. Noridian agreed to cover CYP2C19 testing for acute coronary syndrome patients undergoing stent procedures who are considering treatment with Plavix (clopidogrel); as well as CYP2D6 testing before giving depression drugs amitriptyline and nortriptyline, and the hyperkinetic movement disorder drug tetrabenazine. In the same LCD, Noridian said it would cover CYP2C9 and VKORC1 testing for the anticoagulant warfarin when it is done as part of a prospective, randomized-controlled trial that meets CMS' coverage with evidence development criteria.
Genelex's PGx panel tests for many other gene-drug associations that this policy would not cover, even though many of the associations are the subject of guidelines issued by the Clinical Pharmacogenetic Implementation Consortium (CPIC) and addressed in FDA-approved drug labeling. So, Genelex applied for a formal LCD reconsideration after publishing the IMPACT study results, based on the advice of Noridian Medical Director Gary Oakes, Ashcraft said. Genelex had designed IMPACT with input from Palemetto Medical Director Elaine Jeter and Dane Dickson, former director of clinical science at MolDx, she added. Heeding their advice to involve research experts, Genelex gave an unrestricted grant to researchers from the University of Utah to gauge the clinical utility of its PGx panel in IMPACT, and to propose a randomized-controlled trial based on the results.
IMPACT showed that healthcare resource utilization was 72 percent in the PGx-tested group, and 49 percent in the untested group. Although overall utilization was greater in the tested group, Ashcraft said this was because patients who wanted to get on different drugs based on PGx testing results had additional outpatient visits. The tested group, however, had much lower hospitalization rates compared to the untested cohort (10 percent versus 16 percent) and fewer emergency room visits (4 percent versus 15 percent).
There was a $168,896 and $15,390 difference in hospital and emergency room visit costs, respectively, between the tested and untested groups. Using mean cost estimates, the researchers calculated that a PGx strategy offset $788 of the $914 list price of Genelex's test, and imparted a net savings of $218 per patient.
However, IMPACT study authors highlighted several limitations in the paper. For example, researchers matched patients in the PGx testing arm in a 4-to-1 ratio with untested patients from a large claims database, but they recognized that "the use of an administrative database for historical controls provided inherent potential bias." Additionally, the statistical methods used to match patients between the two arms did not factor in race and ethnicity, a key factor influencing the prevalence of PGx markers in certain populations.
Given these limitations, a payor evaluating IMPACT must consider what definitively can be concluded about the clinical utility of PGx testing. In Putcha's view, IMPACT supports, but does not prove, the test's clinical utility for the application envisioned in the study. For example, a payor would question, he noted, why the authors decided to compare a prospective cohort with historical controls from a claims database, and what the standard of care for the patients in the control arm was.
The tested group was enrolled at three clinical sites between October 2014 and June 2015, and the information on the controls was taken from the claims database for a period between July 2012 and December 2013. "Why not just enroll the patients from the same clinics over the same period of time?" wondered Putcha.
Moreover, IMPACT showed that using median cost estimates, PGx testing resulted in the healthcare system paying $126 or 14 percent of the retail price of the test. When mean costs were employed, testing saved the healthcare system $218 per patient. Given the distribution of costs, Putcha wondered which is more appropriate, median or mean cost estimates.
Language of medical necessity
Labs would counter that payors have too high a bar, particularly for gene-drug associations that are in FDA-approved drug labels and CPIC guidelines. But FDA and CPIC do not always speak the language of "medical necessity," which makes for a difficult translation job for test developers conducting clinical studies.
Although the FDA has updated numerous drug labels with gene-drug association information, most of this shows up in the clinical pharmacology section describing how therapies are metabolized. Gene-drug information is often not in the indications and usage, dosage and administration, contraindications, or warnings and precautions sections, Putcha noted, which contain more specific guidance as to actions doctors can take.
In very few cases, the agency has explicitly required PGx testing in drug labels, and anything short of that tends to slow physician adoption and results in inconsistent coverage. For example, the label for the HIV drug Ziagen (abacavir) restricts the drug for patients with the HLA-B*5701 allele because they are at risk for life threatening hypersensitivity reactions, and it tells doctors in no uncertain terms to screen for this marker before giving patients the drug. Based on this labeling language, doctors consistently perform HLA-B*5701 testing before prescribing patients Ziagen.
Comparatively, the black box warning in the label of Plavix tells doctors that acute coronary syndrome patients undergoing stent procedures who are poor CYP2C19 metabolizers exhibit higher cardiovascular event rates at standard doses. But the label only informs physicians that tests are available and asks them to "consider" alternative treatment strategies for poor metabolizers. This is certainly one reason why CYP2C19 testing is not widely performed in the context of Plavix, and many cardiologists remain unconvinced that PGx testing is even necessary.
"Unless it is clear what specific actions a prescriber should take for a given drug, given a particular CYP genotype," one cannot claim that "CYP genotyping is medically necessary," Putcha said. In the same way, he observed that CPIC guidelines do not tell payors whether to test, but how to interpret results, if available.
What Genelex does and what [Palmetto's] MolDx does is 180 degrees [in] different directions.
The genetic testing industry received some bad press this year after Medicare halted payments to New Orleans-based Renaissance Rx and began investigating its billing practices. Some have accused Renaissance Rx of providing kickbacks to doctors taking part in a 250,000-patient study aimed at proving the clinical utility of its PGx test. As part of the study, Renaissance had a registry of physicians, some of whom have said they were paid for enrolling and testing patients for the study.
Although Renaissance denies allegations of wrongdoing, the company had to halt the study after Medicare stopped paying for its tests. Since then, reports have implicated a number of other firms in having inappropriate relationships with doctors and providing kickbacks.
This "bad juju," in Ashcraft's words, compounded an already difficult reimbursement environment for PGx testing, and is squeezing firms like Genelex. The company maintains it has been practicing good science and making every effort to provide the type of evidence needed for reimbursement, but payors just cannot see beyond randomized-controlled trials.
"What Genelex does and what [Palmetto's] MolDx does is 180 degrees [in] different directions," said Paul Seesman, Genelex's director of payor relations, noting that there seems to be unfounded concern among payors that PGx tests are being misused and do not have much clinical utility. After talking recently with Palmetto's Jeter about the latest publication, Seesman came away feeling like there was no path for establishing the clinical utility for PGx panel testing for polypharmacy patients who are at risk of serious adverse events.
Seesman believes the payor community's stance is out of step with other government agencies, like the FDA and National Institutes of Health, who are pushing aggressively to implement the Precision Medicine Initiative. Attitudes are changing on the frontlines of medical prescribing too. Ashcraft noted that accredited pharmacy schools now have mandatory curricula on phamacogenetics. Even the American Society of Health-System Pharmacists has recognized the importance of PGx testing in a position statement and has encouraged education. "It's basic science and the level of evidence payors are requiring doesn't make a lot of sense," Ashcraft said.
In Putcha's view, payors are not "conceptually" wedded to randomized-controlled trials. Specifically, with regard to IMPACT, the design of the study makes it more "hypothesis generating" than "hypothesis proving" from a payor perspective, he said.
This is something the IMPACT study authors recognized in the paper, noting that IMPACT "can inform the design of future studies" where more direct comparisons between tested and untested patients can be made. "The evidence in this study should be further corroborated with randomized observational data in a unified data source to link these outcomes to the impact of these interventions," the study authors concluded.
The power of this is in improving polypharmacy management.
Toward randomized trials
Genelex was hoping that based on the findings from IMPACT, Medicare would at least cover testing for patients within a randomized-controlled trial. Absent Medicare coverage, Genelex is working with accountable care organizations, some commercial payors, and self insurers. Genelex can also bill Medicare for covered PGx indications, such as CYP2C19 testing for Plavix, but doing so would not allow the lab to perform testing in an efficient manner.
When a patient's doctor orders CYP2C19 testing, anti-kickback laws restrict Genelex to reporting only that specific result, even though its panel test might reveal additional markers informative of the patient's ability to respond to other prescribed drugs. Patients have to sign statements recognizing that only CYP2C19 testing is covered, and if they want to know other results, they have to pay out of pocket.
In such scenarios, Genelex runs the whole panel, but does not analyze or report other markers. This wastes the power of panel PGx testing, in Ashcraft's view. "The power of this is in improving polypharmacy management," she said.
The national payment limit for CYP2C19 testing is $291.80. It is most efficient for the lab to run the whole panel, and if all 384 well trays are full, then it costs a few hundred dollars to run the test. "It can cost us way more than we're getting paid if that tray is not full," Ashcraft said.
Throughout the molecular diagnostics industry, labs have complained that Medicare payment rates do not cover the cost of performing testing, and some firms have gone belly up in recent years as a result. According to Ashcraft, a few hundred dollars of reimbursement per test does not begin to cover the firm's sales efforts, development of the clinical decision support tool, operating expenses, and the randomized-controlled trials that the firm is now conducting to meet payors' demands.
For example, Genelex is planning to randomize 300 patients into a trial where a third will have standard medication management, a third will have standard management plus drug interaction risk analysis with Genelex's YouScript software, and another third will receive all this plus pharmacogenetic testing. Another study, in collaboration with Harding University, aims to randomize 100 subjects into two groups — one receiving PGx testing and the other not — and pharmacists will use YouScript to make treatment recommendations for patients who receive genetic testing and those who get the standard of care.
Ashcraft noted that University of Utah researchers Genelex worked with for the IMPACT study are putting together a proposal for a randomized-controlled trial. For all these additional studies, Genelex is currently seeking investment.