Originally published March 26.
By Turna Ray
Abbott Molecular plans to contact the British Journal of Cancer to contest the conclusions of a recently published cost-effectiveness analysis involving its Vysis ALK Break Apart FISH Probe Kit.
According to Stafford O'Kelly, president of Abbott Molecular, the authors of the study published in BJC last month based their economic analysis on two erroneous assumptions: the list price charged by labs for the ALK test and the prevalence of ALK rearrangements in the advanced non-small cell lung cancer population. As such, the modeling performed by the researchers to determine the circumstances under which the pharmacogenetic test is cost effective is flawed and should not be considered by healthcare providers and payors, Abbott maintained.
"If clinicians were to act on this article, patients will suffer," O'Kelly told PGx Reporter. "The whole premise of the paper is based fundamentally on very incorrect assumptions."
In the BJC paper, University of Colorado researchers Adam Atherly and Ross Camidge modeled the health economics of administering Pfizer's non-small cell lung cancer drug Xalkori to patients whose tumors are ALK mutation-positive. They found that broadly testing all advanced NSCLC patients in order to identify the small subset of ALK-positive individuals who should be treated with Xalkori did not meet a cost-effectiveness bar of less than $100,000 per quality-adjusted life year gained.
The US Food and Drug Administration last August simultaneously approved Pfizer's Xalkori and Abbott's Vysis ALK Break Apart FISH Probe Kit. The drug costs more than $115,000 per year. The $1,400 price tag for FISH-based ALK testing cited in the BJC analysis was established by "expert opinion" gathered by the researchers.
"Prices for the different tests vary depending on the payor and system. In the US, for example, different insurers reimburse charges at different rates. To limit this complexity, we have therefore taken charges, not reimbursements, as our base values," the study authors detail in the BJC article. "We estimated costs for pathological testing, including both technical and professional fees, utilizing Medicare list prices and the associated University of Colorado charges."
Assuming testing costs within a range of $600 to $1,400 per patient, Atherly and Camidge found that PGx testing for Xalkori is not cost-effective because ALK rearrangements occur in less than 5 percent of advanced NSCLC patients. However, the researchers demonstrated that by applying enrichment strategies to narrow the NSCLC population receiving testing — for example, if physicians only tested those NSCLC patients who have adenocarcinoma histology, are non-smokers, and are known to have EGFR and KRAS wild-type tumors — payors could potentially more than double the "mean health gain" to around 0.29 QALYs gained per person from 0.013 QALYs gained per person if all advanced NSCLC patients were genetically tested (PGx Reporter 3/21/2012).
In O'Kelly's view, the researchers should not have based their economic modeling on the list price labs charge for the test, but should instead have pegged the analysis to payor reimbursement rates. As the manufacturer, Abbott said it charges labs less than $170 per patient for the ALK FISH test kit. The laboratory then factors in costs associated with performing the test when billing for it.
The lab costs of between $600 and $1,400 cited in the study are "highly exaggerated," O'Kelly asserted. Furthermore, the list price doesn't accurately reflect what payors are reimbursing for the test, which in his view is the most important number when it comes to calculating what a medical intervention costs the healthcare system.
An independent reimbursement expert PGx Reporter consulted agreed that the reimbursed amount, not the list price, reflects the "real economic value" of a test.
Abbott said that Medicare reimbursement for manually running its ALK test ranges from $329 per test to $571 per test, with average reimbursement at $440 per test.
This reimbursement amount covers what the manufacturer charges, the lab's cost of labor, and any additional markups by the lab. The lab list price is usually higher than what payors reimburse, and it is common for labs to take a loss on the difference between the list price and the reimbursed amount, according to experts PGx Reporter consulted.
The median age of onset for lung cancer is around 70 years old. As such, most people considered for treatment with Xalkori and companion ALK testing will be covered under Medicare. Private payors reimbursing for the test will likely follow Medicare's lead in setting rates.
"We have our own service lab and when we look at what private payors will pay for the test, it's actually not more than what the Medicare reimbursement is," O'Kelly said. "Occasionally, you will have an outlier, where an institution, which in this case is the University of Colorado lab, will bill out at $1,400. First of all, they can't bill that to Medicare, and if they did bill that to a private payor, the private payor would be unlikely to pay that."
PGx Reporter contacted several labs to confirm the list prices cited by Atherly and Camidge, as well as the reimbursement amounts discussed by Abbott. A General Electric Healthcare spokesperson said that the company's molecular diagnostic subsidiary Clarient lists Abbott's test at approximately $1,300 and the firm typically receives $530 in reimbursement from Medicare and third-party payors.
The GE spokesperson explained that the list price offers a starting point for the payor, and noted that it is standard practice in the industry for labs to "accept whatever the insurance company deems reasonable and customary."
Another error in the BJC analysis, according to O'Kelly, is the biomarker prevalence rate the study authors used. "There are multiple studies out there that show that between 3 percent and 5 percent of NSCLC patients have the ALK translocation," O'Kelly said.
Atherly and Camidge projected an initial ALK positive frequency of around 1.6 percent, noting that although previously published studies have estimated that around 4 percent of patients harbor ALK rearrangements in an unselected NSCLC population, "most of the resection cases analyzed have been heavily biased towards adenocarcinoma histologies, which may predominate among the early-stage lesions present in such series."
O'Kelly suggested that if Atherly and Camidge had considered the reimbursement amount and a higher prevalence rate for ALK rearrangements in NSCLC patients, then their analysis would have improved the test's chances of being cost-effective. "There's no reason why the researchers couldn't apply the $440 [average Medicare] reimbursement, instead of the $1,400 [lab] charges, within their model," he posited. "That would give a very different answer."
Camidge declined to address Abbott's concerns with PGx Reporter. He said he will reserve a response until Abbott publishes its editorial in BJC, and avoid any "side discussions."
"The answers to all [the] questions are in our paper already, regarding us using charges as reimbursement varies and we then model multiple different prices," Camidge said. "The basis for the frequencies in the model is all outlined there, too."
The Benefit of Enrichment
Beyond the technical issues with the study methodology, Abbott ultimately challenges the notion that enrichment strategies to increase cost-effectiveness of PGx testing would be beneficial for patients.
"None of the sorts of programs" involving enrichment strategies to hone in on patients most likely to test positive for a particular marker of interest "have been validated at all," O'Kelly noted. "The problem that you run into there, absent a validated study … [is] you run the risk that patients are going to suffer. By this paper, if an enrichment plan [was used by a doctor], it would miss a significant number of patients.
"None of us want to be that one patient because a payor is trying to reduce spending, especially if they're trying to reduce spending based on an incorrect assumption based on a health economics and outcomes research study that's incorrect," he continued.
In the BJC paper, Atherly and Camidge readily acknowledge that enrichment strategies will miss a portion of patients that would have responded to the drug. According to their analysis, by applying an enrichment strategy the cost per QALY gained would decrease to around $4,756 compared to the more than $100,000 per QALY gained when the entire advanced NSCLC patient population is tested. However, such a paradigm would also miss around 56 percent of patients with ALK rearrangements.
It is currently unclear whether insurers are considering implementing strategies to specifically narrow the population receiving ALK testing for Xalkori. Several payors have said they require healthcare providers to garner their authorization before performing ALK testing to ensure that patients are being tested with the FDA-approved test developed by Abbott.
However, insurers have also employed prior authorization strategies to rein in inappropriate utilization of genetic tests, for example in the case of Myriad Genetics' BRACAnalysis test to assess risk of hereditary breast and ovarian cancer.
Difficult economic conditions and the healthcare reform law have placed pressure on payors to lower unnecessary spending by ensuring that they are paying for interventions that truly improve patient health. As such, standards-setting bodies and payors have launched cost-effectiveness and comparative effectiveness studies for some molecular tests.
In the area of BRCA testing, for example, Aetna is surveying de-identified data on 13,000 of its members who have received BRCA testing in the community care setting in an effort to track how physicians are making decisions to administer such testing. The insurer has a hypothesis that "a large amount of BRCA testing that is being performed is not being performed in an evidence-based manner" (PGx Reporter 9/22/2010).
Meanwhile, the US Preventative Services Task Force, an independent panel of non-government experts, is performing studies to determine under what circumstances genetically testing asymptomatic women for their risk of developing hereditary breast and ovarian cancer has a positive impact on their health (PGx Reporter 3/7/2012).
PGx products, by definition, are indicated for a small patient subset, which translates into less revenue for drug and test makers compared to what they would earn for treatments and diagnostics intended for a molecularly undifferentiated population. However, there are different economic factors at play for developers of therapies and diagnostics under this model.
On the therapeutic side, drug developers can receive additional market exclusivity and tax breaks if they advance PGx drugs that receive orphan status from the FDA for meeting an unmet need in a small disease subpopulation. Furthermore, Atherly and Camidge note in their paper that when a treatment is targeted for a small genomically defined patient population, a higher drug price is "a precedent already partially established by some cost-effectiveness bodies," who reason that despite the higher cost, "the overall impact on society … will be low, whereas the gain for the individuals affected may be great."
It is more challenging for test developers and labs to garner a return on investment for companion tests. As such, it is in Abbott's best interest to cast a wide net within the advanced NSCLC population and discourage healthcare providers from employing enrichment strategies that would reduce the number of ALK tests performed for Xalkori response.
The company has efforts underway to educate physicians about which patients they should consider for ALK testing. "We have an active education program at major oncology symposiums where physicians typically go and interact with one another," Kathryn Becker, director of Abbott Molecular’s oncology business, told PGx Reporter. "We demonstrate the clinical effectiveness or our product and educate them about the different kinds of tests and the fact that [Abbott's] FDA-approved test is the one that's linked to the therapeutic outcome. We're doing that on a global basis."
Additionally, Abbott is in the middle of conducting its own cost-effectiveness analysis for ALK testing in lung cancer. The study will be completed in six weeks.
Have topics you'd like to see covered in Pharmacogenomics Reporter
em>? Contact the editor at tray [at]
genomeweb [.] com.