By Turna Ray
While personalized medicine supporters are hopeful that comparative effectiveness research will encourage the adoption of genomically guided medicine for small subpopulations of patients, payors in the post-healthcare reform environment are more likely to use CER to demand a high evidence threshold supporting the clinical utility of pharmacogenomic treatments in order to reimburse for such products.
At a conference on CER and personalized medicine at the National Institutes of Health's campus in October, representatives from the payor community discussed how efforts under the healthcare reform bill to improve outcomes and reduce costs might impact payment for personalized medicine products. Based on comments from a representative from a large private insurer and another from the Centers for Medicare & Medicaid Services, it seems that for the most part payors are still of the opinion that personalized medicine lacks the evidentiary support to justify broad coverage. However, payors do see CER as a way to gather data on whether to reimburse for certain genomically guided medicines.
Speaking at the meeting, which was hosted by the ECRI Institute, Lee Newcomer, UnitedHealthcare's senior vice president of oncology, defined personalized medicine as the practice of "selecting evidence-based therapies based on the unique characteristics of the patient."
"It's evidence based ... It does not [mean] we have a carte blanche to try the theory outside of the clinical research and clinical trial area," Newcomer said. Then, linking CER to personalized medicine, Newcomer defined "comparative effectiveness research" as the process of "developing the evidence to support personalized medicine decisions."
Another representative from the payor community, Barry Straube, chief medical officer and director of the Centers for Medicare & Medicaid Services' Office of Clinical Standards and Quality, highlighted decisions earlier this year by the Medicare Evidence Development & Coverage Advisory Committee to require high evidentiary standards for genomic and genetic tests, and to not lower requirements compared to other types of medical tests. "There are a lot of stakeholders who want us to have a lower standard and give a higher precedence and higher priorities to genetic testing than other laboratory testing that we do," Straube said.
Echoing Newcomer's sentiments about the too-rapid adoption of personalized medicine products without adequate clinical utility evidence, Straube expressed concern about research hospitals like Brigham and Women's Hospital, where researchers and doctors are not just studying the impact of genomic strategies in clinical trials, but implementing programs to broadly roll out prognostic and predictive genomic testing in the hospital setting, and to broadly educate doctors and patients about genomically guided personalized medicine.
Referencing an earlier presentation given by Elizabeth Nabel, president of Brigham and Women's Hospital, Straube said he was concerned that people outside the Brigham system would learn of these efforts and start asking their doctors for similar genetic testing, creating an untenable situation for the payor community.
"The reality, although all this is very important and absolutely essential to clinical research, is that when the rubber hits the road, and patients, just as they are doing at the Brigham, start coming in to medical offices and requesting access to various genetic tests and treatments to potentially influence those genetic test results, the enormity of the cost to society is frightening," Straube said.
Given these new realities and under the new healthcare reform law, CMS is trying to get a handle on this "tsunami of decision making," and assess the financial implications for the Medicare program. Under the leadership of CMS Administrator Donald Berwick, the government payor responsible for providing benefits to 114 million people in the US has a "triangle aim" to provide better care for individuals, improve care for the population, and lower costs by offering incentives for improving outcomes and providing higher quality care. The last of these aims, focused on lowering healthcare costs through improved outcomes, could include a variety of genetic tests, Straube indicated.
"Comparative effectiveness and personalized medicine would be a boon to helping us achieve all three of these aims," Straube said. At the same time, for CER and personalized medicine to work toward the common aim of providing better and more efficient healthcare, private payors and the government have a lot to figure out in terms of the types of evidence to require, who will conduct and pay for these studies, and how CER-based coverage decisions will impact the populations they serve.
Straube outlined several challenges for CMS in integrating CER into its existing legal mandate. Under the healthcare reform law, while CMS will have to consider data from CER when making coverage decisions, such comparisons cannot be the sole data source for such decisions.
Furthermore, CER data cannot be used to make coverage decisions for products deemed "reasonable and necessary" by the HHS Secretary. Under the Social Security Act §1862(a)(1)(A), CMS can pay for medical interventions or services that have evidence showing them to be "reasonable and necessary for the diagnosis or treatment of illness or injury or to improve the functioning of a malformed body member." Straube noted that CMS is consulting its general counsel about how to interpret its mandate to cover that which is "reasonable and necessary," and still use CER to inform coverage decisions.
Earlier this year, as part of the healthcare reform law, an independent Patient-Centered Outcomes Research Institute was created, tasked with conducting studies that inform the public and healthcare providers about the comparative risks and benefits of marketed drugs, devices, and medical products (PGx Reporter 03/24/10). The bill also instructs the institute to conduct CER into the utility of medical products in "various subpopulations" differentiated by race, ethnicity, sex, age, co-morbidities, as well as genetic and molecular subtypes.
There is some uncertainty as to how prominently genomically guided personalized medicine strategies will be investigated by PCORI's CER efforts. Advocates of the discipline were encouraged when PCORI announced in September that several members of its newly formed board of governors came from the genomics field. For example, Eugene Washington, chair of PCORI's board of governors and dean of the David Geffen School of Medicine at the University of California, Los Angeles, has focused his research on medical technologies and national health policy for prenatal genetic testing, cervical cancer screening and prevention, health care quality, and racial and ethnic disparities in health outcomes (PGx Reporter 09/29/10).
Another board member, Freda Lewis-Hall, chief medical officer of Pfizer's medical division, was nominated by the Personalized Medicine Coalition for her "commitment to and knowledge of both comparative effectiveness research and personalized medicine."
According to statements from Newcomer and Straube, however, it seems CER will, at least initially, primarily serve payors as a mechanism to look at the relative utility and efficacy of treatments for the general population. CER has traditionally been used to make decisions affecting a broad population. As such, adding personalized medicine to the equation will require a shift in thinking, and a lot of discussions between stakeholders as to how to conduct this type of research for products that work in a small group of patients.
For manufacturers of these products, conducting CER for treatments in genomically defined subpopulations has its own set of challenges. If payors continue to require data from prospectively designed randomized controlled studies in order to justify payment for personalized medicine products, CER studies will require large cohorts and will be costly.
"Right now it's all exploratory and people are trying to make sense of it all. It will be a new set of relationships that will take place in the marketplace around evidence and comparative effectiveness," Genomic Health and PMC co-founder Patrick Terry told PGx Reporter this week. Terry recently joined Scientia Advisors, where he will lead a new practice focused on helping drug, diagnostic, and device companies in North America and Europe develop and implement reimbursement and pricing strategies.
[ pagebreak ]
In aligning the interests of CER with personalized medicine, "it's all pretty squishy right now because it's all high level and theoretical discussions at this point," Terry said. "Payors are saying, 'We want evidence,' and manufacturers are saying, 'Okay, we can produce evidence, but we need to understand what is the cost of producing the evidence, what's the probability of getting paid to produce that evidence, and will we actually see a return on that investment in personalized medicine product development?'"
No Leaps of Faith
Using the metastatic colorectal cancer drug Avastin as an example, UnitedHealthcare's Newcomer discussed some of the challenges payors face when the physician community adopts treatments and interventions before clinical evidence shows that they are medically useful.
According to Newcomer, while Avastin is "extremely useful" in prolonging life in metastatic colorectal cancer, doctors started using it for adjuvant treatment before data came out on how well the drug worked in that setting. "What we saw in our data was that clinicians in the community began treating people with [the] adjuvant drug even before the trials were launched, with the inference, the theory, that if it was good in metastatic disease, it had to be good in adjuvant disease," he said.
"As we all know, those studies came out and showed that there was no benefit [in the adjuvant setting]. But it's that type of leap of faith that we see constantly," Newcomer continued. "And it's that type of tension that comes out in personalized medicine."
In Newcomer's view, payors should be supportive of CER in so far as it allows insurers to pay for drugs that improve patient outcomes and make better use of their resources. "The nice thing about comparative effectiveness is that it will help develop the evidence to inform a good decision. It shouldn't restrict a good decision. It should, in fact, make that decision all the better," he said. While some therapies become obsolete with comparative effectiveness, Newcomer felt that "they will go away for the right reasons."
Furthermore, as far as Newcomer is concerned, there is room for better decision making for personalized medicine products currently on the market. Even though personalized medicine is supposed to make the healthcare system more efficient, Newcomer gave an example involving Herceptin — a drug given to breast cancer patients who overexpress the HER2 protein — to suggest that there is unnecessary healthcare spending even in this arena.
He cited evidence published in the New England Journal of Medicine in which patients treated with 52 weeks of Herceptin showed a comparable reduction in the risk of breast cancer relapse to patients treated with 12 weeks of the drug. Furthermore, patients had fewer cardiac complications in the trial with 12 weeks of treatment than in the trial with 52 weeks of treatment.
Despite this data, "in the US, we continue to treat with a full year of [Herceptin] therapy," Newcomer observed. "It's the standard of care. We're not even considering dropping to 12" weeks of treatment.
What is Acceptable?
In such situations, there is little incentive for a drug company to conduct a trial that would encourage a shorter course of treatment for a drug, but CER would ideally attempt to close such gaps between clinical evidence and healthcare practice. "Comparative effectiveness would take a hard look at these things, and maybe it's possible for us to go to 12 weeks in the US," Newcomer posited. "Or, if this evidence is so compelling it [may] justify a randomized control trial.
[ pagebreak ]
Stakeholders are currently discussing whether a government entity, such as the National Institutes of Health, should step in and conduct CER studies that sponsors lack the financial incentive to do, but which would save the healthcare system money. "NIH is going to get directly involved in translational research and conducting some of these studies," Terry said. "There's no identified single interest that will benefit from this type of evidence evaluation. So, perhaps the federal government should produce some of this data a priori."
This week, an NIH senior working group supported the creation of a new center that will focus on advancing translational medicine and therapeutics. The proposed center will support TMAT research and serve as a resource for the business sector.
As these efforts are still in the planning stages, there are still a number of unknowns regarding the types of evidence manufacturers will need to provide, and how these studies will be paid for.
Although prospectively designed, randomized controlled trials are the gold standard for evidence, there are ongoing efforts within the personalized medicine community to discuss with payors alternative study designs and evidence requirements. Terry noted that sponsors investing in molecularly guided medicine are considering a variety of methods for evidence generation, such as conducing prospective clinical trials on archival repositories, literature bridging, and bundling previously presented evidence in new ways.
"Right now CER is going to be based on already conducted clinical trials. So, we are going to be looking backwards at the seminal clinical trials, comparing head-to-head trials, comparing the literature and the conclusions about the effectiveness of a particular therapy or intervention. We can do that with molecular technologies now as well," Terry noted.
Terry cited Genomic Health as an example of a molecular diagnostics company that has tried and tested some of these alternative evidence-gathering methods, and been successful at garnering reimbursement for a multi-gene expression breast cancer recurrence test.
The company's business model was based on "creating evidence by looking backwards in clinical trials, molecularly deconstructing those trials, and having an explanation of why people benefitted and why others didn't benefit by using these tools," Terry said. "That was an effective way for Genomic Health to drive coverage payment and utilization in the marketplace."
However, whether other molecular diagnostics firms can emulate Genomic Health's example and achieve reimbursement success, especially in the CER realm, remains to be seen. It will likely take several years before payors and drug/diagnostic sponsors settle on the necessary evidence standards for CER in personalized medicine.
"There are a whole host of strategies that can be applied, but what is going to be acceptable?" Terry wondered. "Who's going to go out on a limb and demonstrate some of these alternative methods? And will the payors help support these alternative evidence reviews to set a prototype or a predicate example of how this can be done?"
According to Straube, before healthcare reform, CMS traditionally did not make head-to-head comparisons when making coverage decisions. Instead, CMS has traditionally made its decisions by looking at "what would happen if the patient got the usual care without the benefit of the new treatment," he explained. As such, CMS is currently seeking advice from its general counsel on how to interpret its mandate to cover treatments that are "reasonable and necessary," while still using CER.
[ pagebreak ]
Also, when crafting the healthcare reform bill, many legislators were afraid that a CER-conducting body sponsored by the US government, such as PCORI, would end up looking too much like the UK's National Institute for Health and Clinical Excellence, which uses cost-effectiveness estimates to advise the National Health Service about which treatments it should reimburse. As such, legislators restricted PCORI from the use of cost-effectiveness analysis in CER to make coverage recommendations.
"Through congressional lobbying and other interest groups, we've made sure we've excluded CER to be utilized in cost-effectiveness analysis," Terry said. But "there are workarounds that CMS can prototype … to look at cost."
While CMS is also restricted from using cost considerations to make coverage decisions under its own legal statute, there is an exception under the law that would allow the government entity to use cost-effectiveness data when it comes to paying for preventative services. According to Straube, CMS is in the process of figuring out when it can factor cost data into its coverage decisions for certain personalized medicine products that would fall into the preventative services category. "Whether personalized medicine is, or parts of it are construed, as preventative services remains to be seen," he noted.
Another unanswered question for CMS is what to do for populations of Medicare patients that respond best to a drug that it has decided not to cover, partly based on CER data.
"What about preserving program flexibility for outliers? We know from CER that there is a broad population that can benefit more than another broad population from two different treatments," said Straube, noting that there will need to be a process for determining how patients that respond best to the treatment excluded by CER will receive that drug under Medicare coverage. Of course, it is exactly in these types of situations, when clinical evidence is needed on subpopulations of patients or outliers, where CER and personalized medicine strategies would likely intersect.
Emerging technologies pose yet another problem for CER-driven coverage, one that looms particularly large for personalized medicine as pharmas embrace drug/diagnostic codevelopment strategies. For example, CER data might show that a drug personalized with the help of a genetic test yields better outcomes in a subset of patients than another treatment that doesn't require a genetic test. However, Medicare beneficiaries in disparate geographic locales may not have access to genetic testing and therefore would not be able to get the treatment covered by CMS.
"If you do comparative effectiveness research, and it tells you that one treatment is far better than others, you're left with that one treatment if you mandate coverage for the one treatment only," Straube said. "If something goes wrong, and if that treatment is not available … if we put all our eggs in one basket with one treatment because comparative effectiveness research says that's the best, what about back-up if something happens?"
The availability of the companion diagnostic on a national and global scale when the drug is launched is a logistical problem that pharmaceutical firms are particularly sensitive about. Broad distribution channels and laboratory networks are a must-have when drug and diagnostic firms negotiate the development of a combination product. And now, this availability problem is also on the minds of payors in the CER era.
"This is a very difficult area … We will improve the quality and value of care, I believe, with personalized medicine and comparative effectiveness research," Straube said. "But in terms of translating [this to] coverage decision-making, we have a lot of issues to deal with."
Have topics you'd like to see covered in Pharmacogenomics Reporter? Contact the editor at tray [at] genomeweb [.] com.