Second in a two-part series: Studies looking at the cost-effectiveness of so-called “theranostics” may enable drug makers to guide product development down more efficiently, and payors to use the data to help them decide what — and whether — to cover the new drugs or diagnostics.
Trouble is, very few of these studies exist, according to a recent study.
“There are few clinical applications of pharmacogenomics that are currently very widely used,” said Kathryn Phillips, a professor of health economics and health services research and director of the Program in Pharmacogenomics and Population Screening at the University of California at San Francisco.
Phillips is the lead author of a meta-analysis investigating the state of cost-effectiveness research into diagnostic-therapeutic companion products. “Until we demonstrate that [cost-effectiveness studies] could have value [to drug makers and others], people are going to be reluctant to pursue [theranostics],” she said.
Step 1: Horse. Step 2: Cart.
There exists a kind of chicken-and-egg scenario in which cost-effectiveness studies depend on the existence of theranostic applications, while greater implementation of theranostics could help spur additional cost-effectiveness research.
It may also be that Phillips and co-author Stephanie Van Bebber, a project director and analyst at UCSF, set study-inclusion standards too high [see Part I of this series: PGx Reporter, 12/2/2004]. But even with few examples of cost-effectiveness studies that the researchers could draw from, they identify technical factors they believe are key to influencing cost effectiveness.
In all fairness, Phillips sees the interaction between these studies and the development of the field of pharmacogenomics as a “feedback loop” wherein basic researchers and drug and diagnostic makers inform economic analyses, while the analyses try to light the path ahead.
“This is very much like the old analogy of crossing the river on a bridge that you’re still building,” said Christopher Webster, the director of regulatory strategy and intelligence at Millennium Pharmaceuticals.
“The technology and the thinking and the construction of new paradigms are happening in real time as drugs are being developed. There are relatively few drugs just now that have been approved for these subpopulations — I mean you’re really talking cancer.”
Even in cancer, where pharmacogenomics has taken a stronger hold, there is precious little starting material for economic analysis of theranostics.
“We don’t have that many very effective targeted therapies yet,” said Elena Elkin of the Memorial Sloan-Kettering Cancer Center Department of Epidemiology and Biostatistics, whose research on the cost-effectiveness of Her-2 testing and Herceptin treatment was reviewed by Phillips and Van Bebber.
“At least in cancer, there are a handful” of situations in which the targeted agent is approved, tests are approved for clinical use, sufficient accuracy data exists, and the long-term outcomes of the therapy in different populations is known, she said.
This supposed vacuum of information seems to be filling, though. The oldest study fitting the researchers’ criteria was published in 1998, and according to Phillips, their number is growing. At the same time, there has been a “huge change in the quality of analyses,” said Phillips.
Moreover, a fuller implementation by industry of theranostics might necessitate better — and more — economic analyses, said Phillips. “With the switch to combo products — where [in order] to market a drug, you need the test — I think that is going to change the whole paradigm,” because a diagnostic accompanying a drug will attract some of the scrutiny reserved for drugs alone — scrutiny that may require cost-effectiveness studies, she said.
Yet contributing to the paucity of cost-effectiveness research could be something as simple as the design of Phillips’ and Van Bebber’s meta-analysis, which appears in the December issue of the journal Pharmacogenomics. As a field, the study of cost-effectiveness, if not in its infancy, is in its “adolescence,” said Joe DiMasi, director of economic analysis at the Tufts University Center for the Study of Drug Development.
The researchers went through an “incredible list of exclusions,” and only a small number of studies probably meet their exacting requirements of a full cost-effectiveness analysis, he said. “It’s not that people aren’t looking in this area.”
In the end, seven of the 11 studies examined by Phillips and Van Bebber found a pharmacogenomic strategy to be cost-effective — that is, less expensive per “quality adjusted life year” gained relative to a non-theranostic option.
Two studies found a pharmacogenomic strategy unfavorable from a cost-effectiveness standpoint. The remaining two were reported as equivocal, meaning that changes in basic study assumptions could alter results from favorable to unfavorable cost effectiveness.
One of the equivocal papers was Elkin’s investigation of Her-2 testing and Herceptin therapy. Elkin said that, currently, many cost-effectiveness studies in the field would be considered equivocal because the great variety of situations with which these studies must contend oftentimes leave the door open for assumptions.
“It’s not an ‘all-or-nothing’ proposition,” she said. For example, some situations show cost-effectiveness or cost-ineffectiveness depending on whether a patient has the target that is sought. “The factors that matter are going to be different in a situation where having a little bit of the target means you’re going to benefit a little,” she told Pharmacogenomics Reporter.
Phillips and Van Bebber contended in previous efforts that a combination of factors increase the cost-effectiveness of pharmacogenomics, and they find corroboration in some of the studies examined for their meta-analysis, they wrote.
A very important — and likely prominent — issue driving cost-effectiveness is the efficacy of diagnostics, said Phillips. “We really haven’t sorted out how accurate these tests are and how we’re going to measure that,” she said.
The false-positive rate was a “big factor” in the finding by Elkin and colleagues that a FISH-based test for Her-2 used alone or in combination with an immunohistochemical assay was more cost-effective than the IHC test alone, said Elkin.
Other drivers of cost-effectiveness identified by the researchers include the prevalence of a genetic mutation and the underlying disease; the severity and cost of the outcome a test is designed to predict; gene penetrance; and the test’s cost and turn-around time. The authors suggest these drivers as topics of further research to guide the use of pharmacogenomics in the clinic and in health policy.
“Industry people and regulatory people” have been showing the most interest in the meta-study, said Phillips. Drug and diagnostic company representatives have been interested mainly because they “know that ultimately their products are going to be scrutinized, and they want to know how people are doing this,” she said.
Even though they don’t do economic analyses of drugs, “people like” the Centers for Medicare & Medicaid Services, are “thinking ahead toward, ‘Are we going to be reimbursing for all of this? What’s going to be reimbursed and what isn’t?’”