Moving Targets: Personalized Medicine and Targeted Therapies

July/August 2010, Vol 1, No 3 -


Atlanta, GA—Even as the development of targeted therapies has advanced rapidly in the past several years, the evidence base, evaluation, and reimbursement structures surrounding the use of these products in the real world have lagged behind. A panel presenting research, governmental, and private payer perspectives discussed these issues, and some potential solutions, at the first plenary session of the 15th Annual Meeting of the International Society for Pharma coeconomics and Outcomes Research (ISPOR).

Fully realizing pharmacogenomics and personalized medicine will allow physicians to “identify diseases at an earlier, more curable stage,” suggested Scott D. Ramsey, MD, PhD, of the Fred Hutchinson Cancer Research Center and the University of Washington, Seattle. “If we can do that, we can afford our patients much greater survival and at the same time lower the cost of care for individuals.”

This is especially important, he pointed out, when cancer products are being introduced that commonly cost $10,000 per month. “The proportion of patients that respond to these cancer therapies has remained fairly constant,” which means, of course, “we’re spending ever more money for a flat response in terms of improvement.”

Targeted therapies may improve those response rates, but their use raises other clinical and cost concerns, as Dr Ramsey indicated.

“We expect a huge number of these biomarkers to be introduced over the next 5 to 10 years, and there are several key questions,” Dr Ramsey outlined. “Who will drive their use? Will patients be asking for these tests or will clinicians? When clinicians review the evidence, when will they know that there is sufficient evidence to use in practice?” Payers will also need information to determine payment policies and encourage the use of these products.

Process Costly, Often Flawed

The current situation regarding paying for genomic tests may give a helpful perspective on where costs are headed. According to Dr Ramsey, data from a medium-sized health plan in Washington State (the Regence Group) indicated that in a 12-month period “there were almost 100,000 claims for genomic tests and $85 million billed, just for the tests alone.” That figure does not include all the counseling and other costs.

“If you do the math, we’re looking at about $1000/test,” he concluded. “So there’s a lot of testing going on, but not a lot of certainty about what we’re getting for that money.”

When evaluating genomic tests, Dr Ramsey emphasized the need to key in on 4 elements—does the test measure what is intended (analytic validity); does it measure that item consistently and accurately (clinical validity); and is the test likely to improve patient outcomes (clinical utility). The fourth quality—value, or evaluating whether the test influences care and represents good value compared with not using the test—is increasingly gaining in importance.

The current model of cancer genomic test development is lacking, Dr Ramsey suggested, with gene candidates selected through the idiosyncrasies of individual researchers who are “not necessarily motivated by clinical priorities, [but] are motivated by the area of expertise.”

Later on, the difference between the population used to validate the drug and that in the real world throws off test results. Finally, limited funding is also a factor for evaluating new tests once they are released. “Ninety-seven percent of genomics publications are in that very early phase” of the discovery process, said Dr Ramsey, compared with almost no publications later on, when the population health impact might be assessed.

Evaluating genomic tests ideally would involve understanding the clinical context for that test earlier on, followed by a cost-effectiveness analysis and professional assessment and incorporation into clinical practice guidelines if deemed worthwhile. Currently, a rigorous test evaluation does not happen or happens haphazardly, and tests are moving directly from development into clinical practice, Dr Ramsey outlined.

This process echoes what happened with prostate-specific antigen (PSA) testing, and that is “a lesson we don’t want to repeat,” emphasized Dr Ramsey. Since being approved by the US Food and Drug Administration (FDA) in 1994, 30 million men are screened annually with PSA testing at a cost of $3 billion. Only in 2009 were 2 studies published evaluating the efficacy of this testing; one showed “no impact of testing on prostate cancer deaths,” Dr Ramsey reported, while the other indicated that for every man helped by PSA testing, 48 received unnecessary testing.

How to Move Forward

According to Dr Ramsey, there are several ways that personalized medicine can be advanced. First is to consider genomic testing as a partnership between test developers and research groups on the one hand, and the health system on the other. “I think we need the health system involved,” said Dr Ramsey.

Lawrence J. Lesko, PhD, FCP, director of the Office of Clinical Pharmacology at the FDA, emphasized this point, saying, “It’s important to design these studies with the right questions in mind. It’s amazing to me how often studies are conducted that simply don’t ask the right questions.”

Dr Lesko discussed some other obstacles to the development and widespread adoption of personalized medicine, the first of which is too much information. “Many of the predicted findings of the technology involve hundreds of genes that technology developers have to boil down to something that they can make useful for clinicians,” he pointed out. Second, the FDA and other regulatory agencies that influence the direction of research have not been clear about what re search needs to be done for approvals.

Another tool that may advance genomic testing is the Evaluation of Genomic Applications in Practice and Prevention (EGAPP) process. Ralph J. Coates, PhD, associate director for science for the Office of Public Health Genomics at the Centers for Disease Control and Prevention, discussed how this effort attempts to establish a systematic, evidence-based process for evaluating genetic tests, and move them from research to practice. The criteria look at factors including health outcomes and utility in decision-making, and in many of the examples given by Dr Coates—depression, breast cancer, warfarin dosing—there was insufficient evidence, uncertainty, or a lack of trial-based data available to allow a recommendation about the use of genetic tests. The Centers for Medicare & Medicaid Services has employed the EGAPP process when making a coverage decision regarding genetic testing for response to warfarin.

Payers Ahead of the Pack

Despite the imperfections of the current evaluative system, payers are eager to expand the use of targeted therapies and realize their benefits, said Robert S. Epstein, MD, MS, chief medical officer and president of Medco Research Institute, Medco Health Solutions, Inc. According to a Medco survey of payers from 2008, pharmacogenomics was tied with consumerism as the second most important topic on their radar screens. “Payers find this topic fascinating—they love it,” remarked Dr Epstein, elaborating that pharmacogenomics allows targeting of therapy, reduced adverse outcomes, and improved efficacy—“all things payers would like to see.” Payers are now paying Medco to identify patients on certain drugs to ensure that the right patient is receiving that drug at the proper dose, Dr Epstein noted.

Payers are now offering their membership to do prospective studies, “and I wouldn’t have thought that would happen.” Dr Epstein discussed a collaborative study of this type looking at warfarin and hospitalizations. It was not randomized, but was large, and was “powered to look at the real question on a payer’s mind—not a surrogate end point, but ‘can genotyping reduce a hard end point like hospitalizations?’”

With personalized medicine, Dr Epstein continued, payers are looking for the same thing as with drugs—value for money. They want real-world comparative data with real outcomes of interest, and evaluation of consequential, matters. For the private sector, however, determining return on investment is difficult, and this may be the key factor holding back a wider adoption of pharmacogenomics.

An audience member wondered what would happen if the genomics and personalized medicine achieved the hoped-for result—what if all the best candidates for a drug were successfully identified? What happens to pricing if the market for drugs is cut by two thirds? Dr Ramsey acknowledged the disconnect between drug development and this testing, saying “it’s putting companies in a bind.” Using the example of cetuximab, and how its marketing strategy had to change once KRAS information became part of the labeling, Dr Ramsey offered a guess “that companies are going to have to get on top of it.” Nevertheless, narrowing the population “might not be a bad thing if the efficacy is there,” he suggested.

“We are definitely moving from blockbuster to nichebuster,” said Dr Epstein. He pointed out that Gleevec, which has a small population, is a multibillion-dollar drug. Also, in 1998, there was only one billion-dollar cancer drug; there are now 20 “and they’re all targeted.”

Payers might pay a higher price on a smaller subset of people for a more predictable response, he concluded. “As long as you have better efficacy, or better safety, payers are not upset about that.”