|Home | About | Journals | Submit | Contact Us | Français|
As technology improves, barriers to payment are beginning to fall for costly diagnostic tests that may identify good candidates for even pricier therapies. Part 1 of 2.
When warfarin hit the market, it was used as rat poison. But decades ago, a chance discovery revealed the poison also worked as a potent blood thinner; today, physicians prescribe 30 million doses a year.
The drug’s considerable success has come at a high cost. It is the second-leading therapeutic culprit to send a patient to the emergency room, accounting for more than 40,000 emergency visits each year. Now, though, geneticists at the U.S. Food and Drug Administration say a simple genetic test can cut that number dramatically.
Scientists have long known that people with particular variants of the CYP2C9 gene metabolize warfarin differently, retaining potent quantities longer than other patients do. Later on, researchers added variants of the VKORC1 gene as a signpost for possible adverse reactions. And at the strong urging of Lawrence Lesko, PhD, director of the FDA’s clinical pharmacology office, the FDA agreed in August to place a genetic test recommendation on the drug’s label to help improve its safety profile, potentially shaving more than $1 billion off the tab for ER visits.
Lesko’s victory has been heralded in some quarters as a big step toward a new era of personalized medicine, when genetic tests will be common tools for matching patients with the right drugs. In other quarters, it’s been criticized for advancing a technology that’s not ready for prime time. But wherever the experts stand, the movement of a high-profile drug like warfarin into the ranks of select therapeutics where prudence dictates use of a genetic test has given the diagnostics field an overnight publicity jolt.
That sudden news blitz comes at a time when most health plans and a host of physicians are having their first, often awkward encounters with diagnostic tests. Estimates of the time it will take for tests like these to become the standard of care gyrate widely between a few years to two decades. But that day will arrive, experts say — and already, more than a dozen tests have burst onto the scene, demanding attention and reimbursement.
“I think the field is catching up,” observes Alfred G. Merriweather, chief financial officer of Monogram Biosciences, which just weeks ago commercialized a diagnostic test (Trofile) to identify HIV patients most likely to respond to Pfizer’s maraviroc (Selzentry). “There’s a lot of interest in genomic markers; a lot have been right on target. There’s also a lot of talk about personalized medicine. But it has not translated into many viable products. As tests get more sophisticated, we’ll move beyond talk, toward more substance.”
Monogram’s coreceptor tropism assay, he says, “is a terrific example of truly personalized medicine, with our test being necessary to determine whether a particular patient is appropriate for Selzentry. For many years, Herceptin has been the only real example of a drug linked to a diagnostic. Now we have another.”
“Payers would, in theory, have a significant interest in companion diagnostics,” says Rick Carlson, JD, clinical professor of health policy at the University of Washington. “If they can get the drug to those who can benefit from it, or exclude it from those who won’t, then it’s a cost-effective tool for payers. But they’re just getting around to understanding pharmacogenetics.
“Clinical medicine is more disorganized and inefficient than most people would like,” Carlson continues. “Even when there is evidence, we’re dealing with enormous illiteracy, particularly in the medical community. We have only about a thousand clinically trained medical geneticists in the country. Most clinicians are not even familiar enough to know that there are diagnostics.”
There are exceptions, but some consultants say that most health plans have difficulty keeping track of genetic tests on the market.
“There’s a lack of institutional framework or guidelines similar to what is used in reviewing a pharmaceutical, like the AMCP dossier, for payers to assess this emerging technology,” says Burt Zweigenhaft, CEO of BioPharma Partners, which consults with molecular developers and health plans. “In addition, [payers] lack internal expertise to assess this technology. The technology assessment committee of a health plan would be hard pressed, except in, say, San Francisco or New York, to put together a panel of independent experts to review and then develop coverage policy determinations.”
Several other obstacles also need to be overcome.
First, there are only a handful of pioneering tests on the market, says Carlson. Abbott and Roche have combined diagnostics and pharmaceuticals, he adds, but most of the companies working on developing new diagnostics are small, and, he says, “Big pharma historically has had problems working with a company in a garage.”
For this field to really take off, Carlson says, it needs an effective business model for joint development of new drugs and the genetic tests necessary to identify patients who could benefit from them. And that model will need encouragement from insurers and regulators, both of which have largely been missing from the equation.
But that’s changing.
“All the major payers are looking at this area. They’re beginning to invest the time to understand it,” notes Carlson. “And they’re going to be confronted with propositions to cover something.”
Monogram Biosciences, for example, has been knocking on insurers’ doors to highlight the role of its Trofile assay when prescribing maraviroc, and last month Medicare published coding guidance for coverage of Trofile. Over the past three years, Genomic Health has been hard at work making a case for Oncotype DX, a genomic test that predicts the likelihood of recurrence in patients with early-stage breast cancer and identifies those who are most likely to respond to chemotherapy.
Oncologists already know that chemotherapy is recommended for the vast majority of early-stage breast cancer patients, even though the data tell them that only 25 percent will benefit. Using the test, they’ll find the 50 percent of patients who will not benefit from chemotherapy, the 25 percent who fall into a gray zone, and the 25 percent who will get a “huge benefit,” says Genomic Health President Kim Popovits.
“Payers really understand the need for the tool,” she adds. But getting there was neither easy nor cheap. The company has invested $140 million in advancing the test, a process Popovits compares to drug development. Payers wanted multiple studies in peer-reviewed journals before covering a test.
Genomic Health’s strategy of aggressively pursuing research publication in peer-reviewed journals and persistently making its case to payers has yielded dividends.
In its quarterly report released in early August, Genomic Health tracked a string of successes in gaining insurer support for its Oncotype DX diagnostic. Blue Cross Blue Shield of Michigan, HealthNet, and HealthPartners all issued positive reimbursement policies. In July, the tech assessment group of the Blue Cross Blue Shield Association gave the test its stamp of approval. And, in a notably high-profile case, UnitedHealthcare agreed at the beginning of the year to cover it as well — with an important caveat. According to one report, when physicians and patients decide to ignore test results and go ahead with chemotherapy even when it is not recommended by the test, UnitedHealthcare would go back and make a billing adjustment. Details were not discussed.
“We haven’t had a trial without a 25 to 30 percent change in treatment. Plans are basically seeing chemo right-sized.”
President, Genomic Health
For Genomic Health, which has staff devoted to seemingly everything, from one-on-one discussions with insurers to a help desk for individuals trying to gain coverage, pioneering this business requires flexibility. Officials are perfectly willing to negotiate the details of performance-based contracts with payers.
“We understand that payers fear they’ll see a $3,500, $3,600 increase in costs” for the test without any payback, says Popovits. “We encourage payers to track treatment decisions so they can see the impact the test is having. As performance-based contracting goes, we are fans of putting our money where our mouth is. If we can’t change the balance or the way patients are treated with chemotherapy, then we don’t deserve to have a $3,600 test in the marketplace.”
But she also is confident that the test will deliver results.
“We haven’t had a trial without a 25 to 30 percent change in treatment. Plans are basically seeing chemo right-sized.”
That $3,500 price tag for Genomic Health’s test can save $20,000 in unneeded chemotherapy, says BioPharma’s Zweigenhaft. That’s to say nothing about the side effects patients endure without a clear clinical benefit and the additional episode-of-care costs.
“Genomic Health is the poster child for this business,” he says, ticking off the reasons: clear data supporting the use of its test, obvious benefits for patients, an economic offset for health plans, and — when he started looking into the field two years ago —acceptance and favorable coverage policy were “painfully slow.
“Two years ago,” he continues, “you had no policy or coverage determinations, and thus all the tests were automatically declined.”
That was a serious impediment to physician adoption of molecular technology.
“Doctors perceived that there’s a huge barrier in test approval,” he adds. “Even more educated and motivated doctors are very resistant to fighting costly battles and dealing with the administrative delays and grievance procedures needed to utilize this technology.”
Genomic Health says its recent successes should make it much easier for patients to get the Oncotype DX test. So far, 6,000 oncologists have ordered it. But with only about a third of breast cancer patients getting the test, Genomic Health also knows it has a way to go before the diagnostic test becomes a standard of care.
John Watkins, RPh, pharmacy manager in charge of formulary development at Premera Blue Cross, has been a particularly keen observer of the progress of Oncotype DX and the field of genetic tests.
“We’ve been seriously looking at it ever since we started our biotech initiative in 2003,” says Watkins. As new oral cancer drugs came along, the data didn’t make a convincing case for providing them to broad populations. But with population subsets responding to these new targeted therapies, it also became clear that genetic tests were essential for identifying the right patients.
For Watkins, who prides himself on maintaining a collaborative approach with physicians in the plan’s network, it’s also important not to get ahead of the science. Trastuzumab (Herceptin) would certainly fit the profile of an expensive oncology drug tailored to a patient subset. Widely heralded as one of the first in a new generation of personalized therapies, trastuzumab is intended only for breast cancer patients whose tumors express the HER2/neu protein. But Premera has no requirement for physicians to conduct the HER2 diagnostic test before it covers trastuzumab.
“I learned quickly from talking with experts that there are issues around diagnostic tests for HER2,” Watkins says. “There are multiple tests and no agreement on the validity of those tests. It was sort of a muddle.”1 Over time, the testing process has been improved such that there is now general acceptance of the results, but Watkins also is hearing that physicians in his network are still routinely prescribing trastuzumab without ordering the test — an issue that was reported in July in the Journal of the National Cancer Institute.
A test for imatinib (Gleevec), a powerful cancer drug prescribed for a subset of leukemia patients, is required to meet the plan’s mandate on prior authorization. In that case, says Watkins, there is clear evidence that the test works as billed.
For now, Premera is reviewing each genetic test to determine its appropriate role in predicting treatment outcomes.
“There’s a lot more coming,” says Watkins, “but I think we are still in the infancy stage.”
For health plans, it’s all a work in progress. Watkins’s pharmacy group has been investing more time in reviewing tests now in the hope it will pay off with a more streamlined approach to approvals.
“Our process for drugs is better established,” he offers. “The process for these tests is still under development. Hopefully, after 10 or 12 of these, some patterns will emerge. The goal is to establish our pipeline surveillance process to identify a test no later than six months before it’s available, and we can then discuss it internally.”
As drug developers link therapies to new diagnostics that better identify appropriate patient populations, you’ll see more and more drug/test pairings take hold, says Popovits. But it’s no surprise that some big pharmaceutical companies aren’t very enthusiastic about the diagnostics revolution.
“You can imagine having an Avastin or Erbitux available — a large market — treating many to find few who benefit, and then along comes a company that says it can find the few,” she adds. “That can be challenging for a drug already commercialized, but we believe developers are beginning to value this approach in earlier stages of development.”
“Do you think that if you were a drug executive that you would want to be tied to the hip to this technology?” Zweigenhaft responds. “If the doctor doesn’t know who should get the drug, I sell 100 doses. If they do know who benefits, I sell 50. And if health plans don’t mandate that test, then why should I? Inappropriate and off-label drug utilization is a big part of the economics in this business, and if we don’t start personalizing the targeting of clinical response, the health system will push back — or even worse.”
“Inappropriate and off-label drug utilization is a big part of the economics in this business, and if we don’t start personalizing the targeting of clinical response, the health system will push back or even worse.”
There’s another catch.
To mandate these tests, observers note, health plans have to learn how to evaluate their cost-effectiveness. The silo mentality still pervades managed care, though, and sticker shock alone can be a problem.
“The Oncotype DX test is very expensive,” at least in comparison to most generic lab tests that payers have covered up to now, says Carlson. “Health plans treat diagnostics as commodities, because they have been relatively inexpensive. They buy lab tests in bulk.”
In another respect, though, price commands payers’ attention.
“If the price of the test is considered excessive, then they will look at something more closely,” says Carlson. Insurers want to avoid any sudden spike in costs associated with new therapies. They also want to learn about anything that can help rein in expenses on their 10 most costly diseases. Tests that can control expenditures on members with diabetes, cardiovascular disease, and other major cost drivers are much more likely to gain consideration than something aimed at a small patient population.
“If it fits into one of the top 10,” says Carlson, “it will get a closer look than some rare genetic disorder.”
There are also unresolved problems trying to track test usage.
“The coding system for billing is considered woefully out of date in the molecular diagnostic field,” says Peter Keeling, CEO of Diaceutics, a United Kingdom–based company that advises pharmaceutical and diagnostics makers on pairing up. “Most are a compilation of multiple codes — nothing to indicate a test’s use for oncology, breast cancer, or neurological disorders.”
“We have an antiquated medical claims processing system for molecular test reimbursement that can’t tell the difference between tests,” agrees Zweigenhaft. “Basically what happens when a payer approves one test, it ultimately and unintentionally opens the door to similar unapproved tests using a generic HCPCS code that fly under the radar to get paid.”
“As we speak, the vast majority of insurers don’t link formularies test claims,” notes Premera’s Watkins. “Those systems don’t talk to each other. We could use codes for molecular tests that describe the test to the specificity level of NDC numbers.”
Zweigenhaft goes so far as to predict the emergence of genomic benefit management companies similar to pharmacy benefit managers that will deploy a centralized contact-center model to handle coverage determinations, manage payment, and provide consultation with the treating physician and patient. Already, early start-up companies, such as YgeneX, have sprouted to help payers and genomic developers alike to implement efficient and accurate access to this technology.
The science of diagnostics also has a long way to go, says Gonzalo Laje, MD, a National Institutes of Health researcher and coauthor of a recent American Journal of Psychiatry paper linking the effectiveness of the antidepressant citalopram (Celexa) to a particular genetic variation. Laje believes that as our understanding of the field gets better, the tests will get better as well. And though the goal will be to say with near certainty that a therapy will work as intended, he says, single nucleotide polymorphisms may not account for the whole picture: copy number variation, epigenetics, and even environmental factors could have significant interactions. In the case of warfarin, for instance, a diet or supplement with excessive vitamin K would cancel out the desired anticoagulant effect.
“We’re looking for 90 to 95 percent likelihood” that a test result is reliable, says Laje. “That would be the number to start looking at. A minimum of 90 percent.”
Some experts in the field say this technology will pass a critical milestone when diagnostic tests are developed hand in hand with the new biologics that crowd developers’ pipelines.
“In 10 years, about 20 to 25 percent of new therapeutic products in the pipeline will depend on some degree on a related test,” Keeling predicts. “One of the underlying problems, if not the underlying problem, is that a new diagnostic can take 10, 12, 16 years to feed into the standard of care. And pharmaceutical companies may be looking at providing sufficient resources to drive a therapy to peak sales in a three-year time frame. That gap does not work when a therapy is dependent on a companion test.”
“Views have been changing profoundly over the last two or three years, and it’s still shifting,” adds Carlson, at the University of Washington. “The delivery system in healthcare is so ponderous that everything takes a long time.”
Looking down the road five years, he says, diagnostic tests won’t yet be common for a drug. Ten or 15 years from now, we’re likely to see broad improvements as diagnostic tests linked to therapies eventually become a standard of care.
Test developers say if you persist, you can make headway. Forecasts of a slow-motion acceptance of the tests, they think, are overly cautious.
“I think we’re going to get there in probably five years,” says Genomic Health’s Popovits. “I may be more bullish than most, but after three years’ experience with Oncotype DX, we’ve gained the acceptance of the professional community, and payers have seen an impact on treatment decisions. The fact is that the business model is successful in luring others to move into the space, and a rising tide lifts all ships.”
It is still common for new technology to take 15 years to become fully accepted in healthcare, notes Zweigenhaft, “but that was before you had $100,000 drugs and the Internet.” That’s why he thinks this technology can take off in five years.
“When health plans understand the clinical utility and overwhelming economic power in this knowledge,” he adds, “they’re going to insist that drugs and tests be combined.”
1The picture may be changing. At the American Society of Clinical Oncology meeting in June, Monogram Biosciences presented clinical data from its VeraTag assay in separate clinical cohorts of trastuzumab-treated patients with metastatic breast cancer. The test identified different subpopulations of patients with different clinical outcomes, regardless of whether patients were selected for therapy by immunohistochemistry or fluorescence in situ hybridization.