In tomorrow’s clinic, health care providers diagnosing a patient or prescribing a drug are likely to tap the tools of a growing field: pharmacogenetics, or using genetic variation among patients to better understand how different patients will respond to a given therapy. Using pharmacogenetics, for example, health care providers could learn that two patients who share similar symptoms actually have different sub-types of a disease and hence require different drugs. Providers could also determine whether two patients metabolize a drug at different rates, and thus need a higher or lower dose of that drug.
As the science of genetics marches forward, researchers generally agree that it’s becoming scientifically possible to identify genetic variation among patients and to link that variation to differences in drug response. Now, this emerging field opens the door to a host of health care questions: How can these new diagnostics and new drugs get to the market? What does that mean for the health care industry? And what regulatory policy and oversight are needed for this emerging field?
These issues informed “Pharmacogenomics: Is This the Drug for You?,” a Genetics Perspectives on Policy (GenePOP) Seminar held April 4, 2006, at the National Press Club in Washington, D.C.
The Genetics and Public Policy Center, supported by The Pew Charitable Trusts at Johns Hopkins University, hosted the seminar. The panelists were: Mara Aspinall, president of Genzyme Genetics (a division of Genzyme Corporation), a genetic testing company with labs across the United States performing nearly one million genetic tests annually; Gail Javitt, director of law and policy at the Center, whose work explores the regulatory landscape for genetic testing laboratories and the tests themselves; and Kathryn Phillips, a professor of health economics and health services research at the University of California, San Francisco, whose research focuses on the use, cost effectiveness, and regulation of pharmacogenetics. Center director Kathy Hudson moderated the session.
Launching the discussion, Aspinall described how far researchers have come in recognizing and fighting disease. Fifty years ago, she noted, health care providers diagnosed cancer vaguely, as “something wrong in the blood.” Today, researchers recognize 39 types of lymphoma and 33 types of leukemia. This sub-typing characterizes every major disease field. Moreover, similar success has swept the health care field, bringing more specific drugs, more tailored therapies, better educated patients and physicians, and more precise and rapid diagnostic technology than ever before.
Amid these advances, however, health care still has serious gaps. Adverse drug reactions remain a major threat to patients—and may even represent the sixth major cause of death today, according to some estimates, Aspinall noted. Drug compliance rates remain low, as many prescriptions go unfilled or are improperly used. In addition, although disease sub-types increasingly are recognized, effective treatments for those sub-types frequently are unavailable. Finally, hospitals and health care providers do not consistently use the best new tools, illustrating the need to educate both them and patients.
Against this backdrop, traditional medicine is evolving. Historically, health care providers have used a trial-and-error approach to therapy, using observation to prescribe a therapy, followed by observing a patient’s response, followed by changes or tweaks to the therapy when needed. But the days of trial-and-error are gradually giving way to a more progressive approach, combining clinical evaluation and quantitative testing to improve the standard of care. With new tools becoming available, health care providers can move away from trial and error.
That’s important, Aspinall emphasized, because today, common drugs for depression, lower cholesterol, heart disease, and cancer fail to work because the wrong drug is going to the wrong patient at the wrong time. These errant prescriptions may cost the U.S. health care system between $7 billion and $17 billion annually. Moreover, ineffective prescriptions are likely to lead to poor compliance. The situation may be worst with drugs for cancer, which carries a general survival rate of just over 50 percent. These drugs are not working for all people. Now, we don’t have the answer, she said. Both from a health and cost standpoint, better drug efficacy could make a big difference.
In another example of prescriptions gone awry, Aspinall pointed to the “wheel of misfortune” that characterizes increased antibiotic use. Health care providers frequently prescribe antibiotics ineffectively, without doing tests to determine whether a patient has a type of bacterial infection resistant to certain antibiotics, for instance. The wheel begins with ineffective, increased antibiotic use, which then increases resistant strains, leading to more health care use and limited treatment alternatives. Drugs that were once considered a last resort are now being routinely used to fight bacterial infections.
Regardless of broader economic costs, the real loss behind ineffective medicine is personal: Time is of the essence. Among lung cancer patients, the one-year survival rate averages 40 percent, and five-year survival averages roughly 10 percent. Those patients, Aspinall noted, lack time for a trial-and-error medical system. They need educated health care decisions from their very first visit to a health care provider. The same is true for patients who have cancer, heart failure, renal disease, and other serious conditions.
As pharmacogenetic tests evolve to place patients in sub-categories and treat them accordingly, Aspinall argued, health care providers should use them. This “new, traditional medicine,” she said, will include observation, test, action, and a predictable response. Repetition may not be necessary, as patients experience improved drug use and targeted treatment.
In fact, personalized medicine exists today, she added—just not in the vastly revolutionary style hyped by the media and others during the 1980s and 90s. Health care providers increasingly choose therapies to maximize a patient’s likely response, taking into account a patient’s variety of cancer, say. Health care providers also monitor patients to assess a drug’s effectiveness and the optimal duration of therapy. Increasingly, health care is about individual biology. The future, Aspinall said, is about predictive testing. This marries testing and treatment as personalized medicine that brings new tools to the task of ultimately improving patient outcomes.
To implement pharmacogenetics effectively into health care practice and policy, however, its economic value must be considered, asserted economist Kathryn Phillips. Why is economics important? First, Phillips suggested, economics equals value, and pharmacogenetics will only succeed if it is valued by patients, providers, insurers and other payers, industry, government, and society. Second, she added, economics also equals incentives; people need incentives to adopt pharmacogenomics.
The documented value of pharmacogenetics remains ambiguous. A recent survey of academic literature, she says, turned up just 11 studies of pharmacogenetic interventions, exploring a limited range of conditions with mixed results on whether the interventions were cost-effective.
One example of successfully-adopted pharmacogenetics is the selection of patients to receive an antibody-based drug, Herceptin, to treat breast cancer. Herceptin is costly, at roughly $4,000 for a month’s prescription. The drug is deemed appropriate for patients whose tumors over-express the “HER2” protein, for whom therapy increases median survival by a few months. However, several economic analyses, including a study by researchers at Harvard, found minimal value for the drug’s cost. Outside the United States, in countries that review the economics of new drugs before approval (such as the United Kingdom), regulators were slow to approve Herceptin because of concerns about its cost and value. Concerns are expected to increase because new research shows that Herceptin may be useful in primary cancers, as well as metastatic cases, indicating a much wider group of patients may receive this drug.
On the flip side, a slowly-adopted pharmacogenetics example is a test that analyzes mutations in cytochrome P-450 (CYP 450) enzymes. The FDA has approved the so-called AmpliChip, developed by Roche Diagnostics to identify mutations in CYP 2D6 and 2C19. In an economic analysis, Phillips and her colleagues determined the new test could have a large positive impact on health care. That’s because these enzymes metabolize common drugs, such that the test is relevant to 189 million prescriptions and $12.8 billion annually in the U.S., particularly regarding drugs for mental health and heart disease patients.
However, insufficient data currently exist to assess the impact of using this test. In particular, few data exist on clinical outcomes of the test, or what really works in the clinic. Does the test improve outcomes for patients?
To illustrate the challenges in determining the value of pharmacogenetics, Phillips noted that the popular pain reliever codeine is metabolized by 2D6. Because some women rapidly metabolize codeine, the drug is unsafe for post-partum pain in women who are breastfeeding, who pass the drug along to newborns. The AmpliChip test could identify which women should avoid codeine. But that raises a host of questions, according to Phillips. Who is going to pay for the test? Should everyone take it? Will only people who are insured get the test? How much will the test cost? How reliable is it?
More broadly, to realize the value of pharmacogenetics, the health care industry will need to develop new paradigms to integrate the development of diagnostics and drugs, two historically separate industries and regulatory fields. Early consideration of diagnostics will become critical, Phillips asserted. Moreover, diagnostic drug combinations are complicated and relevant to multiple drugs and diseases.
Among the technical challenges, she cited lack of data linking pharmacogenetics to outcomes, data comparing the effectiveness of therapeutics, and data on the products themselves. Pharmacogenetics is difficult to evaluate because essentially it measures the value of prevention. Medicare coverage is also complicated, as it historically covers diagnostic but not screening tests.
Ultimately, Phillips concluded, pharmacogenetics is the wave of the future. But its value must be demonstrated, which will require more data, and must occur within today’s policy context of high drug costs and safety concerns, amid the continuing need to bring beneficial drugs to the market.
Ultimately, whether pharmacogenetics succeeds has as much to do with its regulatory environment as with the scientific or economic feasibility, suggested Gail Javitt. She pointed out that genetic testing is rapidly growing, with more than 1,200 genetic tests, at least 900 of them available clinically, a handful of which explicitly are pharmacogenetic. This field offers the promise of improved drug safety and efficacy and more treatment options individualized to patients.
But Javitt noted that several gaps in the current regulatory framework imperil this promise. Those gaps lead to questions including: Are pharmacogenetic tests safe and accurate? Do laboratories know how to reliably perform these tests? Do healthcare providers know how to interpret the information?
Regulatory oversight lies with agencies under the U.S. Department of Health and Human Services. The U.S. Food and Drug Administration (FDA) regulates products, such as drugs, medical devices, and biological products and tissues. The U.S. Centers for Medicare and Medicaid Services (CMS) oversees laboratory quality, including certifying laboratories; ensuring the right people are performing tests; examining whether standards are in place; and determining whether proficiency testing is taking place, meaning a laboratory knows how to accurately conduct tests.
In surveys, many members of the public say they believe the government already regulates the quality of genetic tests and also that the government should play its role as part of its responsibility to protect the public and the public health, Javitt said. However, she added, the reality is not so simple.
The FDA does not subject most genetic tests to pre-market review, primarily because genetic testing has emerged in the laboratory environment. The FDA does review diagnostics, which would include, say, a boxed product with labels, reagents, and directions for use. But Javitt notes that only a dozen or so genetic tests fit this diagnostic definition. The vast majority are laboratory-based and performed in-house, requiring no pre-market review. Moreover, even if a manufacturer does bring a test kit to market, a laboratory can offer the same test without going through FDA.
In recognition of the need to address these issues, FDA has recently issued several guidance documents recognizing the potential value of pharmacogenetics to help the agency achieve its mission of ensuring that drugs are safe and effective. Again, however,
these guidances don’t speak to in-house developed tests, which represent the vast majority of tests on the market.
The other major regulatory body, CMS, has authority over the quality of laboratories based on the1988 Clinical Laboratory Improvement Amendments (CLIA). CLIA applies to all laboratories doing medical tests for patient health, including genetic testing. What’s missing, however, is a so-called “specialty” area, with precise regulatory instructions for labs of a given variety. No such specialty area exists for genetic tests right now, in contrast to many other similarly complicated tests being performed today, Javitt noted. Although laboratories can voluntarily demonstrate proficiency testing with specific programs, none are required to.
Javitt noted that this disparate regulatory environment, full of gaps, repeatedly has been highlighted by federally-appointed groups. The stakes are high. With valid pharmacogenetics tests, a patient is more likely to get the right drug, the right dose, better efficacy, and better safety.
For the public and policy makers to gain confidence in pharmacogenetics tests, Javitt said, high-quality laboratories should demonstrate their proficiency, with some mechanism of oversight by an independent third party. Similarly, a test’s validity needs to be supported by data before going on the market. Healthcare providers need to know how to interpret these tests and give them to the right patients at the right time. Finally, the health care industry needs a mechanism for the use of tests in clinical practice to be evaluated over time, she said.
In discussion, the audience raised several questions. Moderator Kathy Hudson questioned whether pharmacogenetics might catalyze a change in payer behavior about prevention. Historically, payers such as insurance providers have been wary of routinely doing preventative tests on asymptomatic people because it is costly. Phillips acknowledged that payers will need incentives to support prevention-oriented testing. What’s more, she noted, people regularly change insurance providers, and companies are reluctant to pay for tests that will not provide a cost benefit to them.
Public policy analyst Amy Fletcher, of the University of Canterbury in New Zealand, asked about the regulation of DNA collections that might build up at in-house labs doing testing. In fact, Aspinall responded, DNA for genetic testing often is banked only for a limited time and is subject to research regulations.
Jill Wechsler, Washington editor of Pharmaceutical Executive magazine, questioned whether legislation would be required create a specialty area for genetic testing within CLIA. The answer, according to Javitt, is no—and in fact, CMS reported in 2000 that such a specialty area would be created, though it hasn’t happened yet. The Genetics and Public Policy Center called upon CMS in 2005 to issue this proposed regulation because patient advocates consider it important to the quality of testing.
Ed Abrahams, who directs the Personalized Medicine Coalition, asked Phillips whether she could envision a working model for personalized medicine in the marketplace, given her reservations about the economic value. Phillips commented that pharmacogenetics does, indeed, have much potential, but economists do need to determine the economic model. That’s difficult to do with scarce data on diagnostics (who’s using them, for what purpose, who’s paying for them, and for how much). She also added that despite the challenges of laboratory development of in-house genetic tests, this approach does ensure that tests will reach the market, a pipeline process that could be far slower if regulated completely by the FDA.
Joe McInerney, who heads the National Coalition for Health Professional Education in Genetics, wondered what primary care providers need to know about pharmacogenetics. Aspinall highlighted a need for fundamental knowledge of genetics, beginning at medical school. Few medical schools currently maintain a full course curriculum that incorporates genetics. Healthcare providers – doctors, genetic counselors, social workers, nurses – need to understand pharmacogenetics and tap the tasks and drugs most relevant to a patient’s situation. The biggest challenge, she added, is creating a way to keep family and general practitioners up to date on changing medical knowledge. Javitt agreed, adding that genetic tests require labels, much like those found on prescription drugs, indicating which patients should take the test, when, and related information.
In fact, Phillips said, health care providers simply lack the time to truly be educated broadly on genetics. Instead, they’d most benefit from practical guidelines. The Center currently is reviewing guidelines available for genetic tests and the process that professional societies use in developing those guidelines, with the ultimate goal of improving practice guidelines.
Tim Leshan, a senior policy analyst at the National Human Genome Research Institute, questioned what kind of independent third party might regulate genetic tests. To that, Javitt suggested the tasks that could be needed, including: proficiency testing of labs; adverse event reporting; patterns in test use; and long-term evaluation of test use.
Vijaya Melnick of the University of the District of Columbia and Georgetown Medical Center remarked on the relevance of the cardiovascular drug BiDil, created under the rubric of pharmacogenetics with a label for African Americans. Aspinall noted that such specific prescription demographics are sure to grow with the field of pharmacogenetics. She drew a parallel with Tay-Sachs Disease testing, primarily conducted for people of Ashkenazi Jewish descent, and sickle cell testing, often given to people of African American descent. From a testing point of view, she noted, racial and descent indications can help health care providers better pinpoint a diagnosis. BiDil crossed the line into a therapeutic, as a drug more effective to people who have a certain racial makeup.
Freelance writer Maya Pines, representing the Howard Hughes Medical Institute Bulletin, asked about any problems in the clinical pipeline, moving tests from the lab to clinic. In general, the panelists cited not regulatory delays, but challenges in the science of developing genetic tests, establishing reimbursement rates from insurance companies, and similar tasks. Aspinall reiterated the importance of scientific innovation and the need to get genetic tests into physicians’ hands to improve patient care.
Jaydee Hanson of the International Center for Technology Assessment questioned how gene patents affect the availability of genetic tests. Aspinall agreed that it’s an issue, citing complaints among health care providers that commercial companies can hold an exclusive patent and profit from an expensive test. Many tests, however, are not patented, she noted. The challenge is to strike a balance, maintaining useful patents while providing health care access to important innovations. Many geneticists, she added, have allowed research to be done on patented genes or tests, for the benefit of society.