News blog

Policy paper: Myriad turns cancer genetic data into trade secrets

If the information Myriad Genetics has collected about breast cancer mutations remains proprietary, costs of gene tests could increase while quality declines, argues Robert Cook-Deegan, a policy researcher at Duke University in Durham, North Carolina. In a paper published today in the European Journal of Human Genetics, Cook-Deegan and co-authors John Conley, James Evans, and Daniel Vorhaus urge health-care payers and policy makers to encourage the company to share clinical data and proprietary algorithms.

In the U.S., Myriad’s patents make it the only commercial provider for tests analyzing the genes BRCA1 and BRCA2. And it performs a lot of tests: nearly a million patients have had BRCA testing, according to information cited by Cook-Deegan. About 5% of breast cancers and about 15% of ovarian cancers are related to mutations in these genes, says Susan Domchek, who studies the genetics of breast cancer at University of Pennsylvania’s School of Medicine.

17-year-old Myriad Genetics turned its first profit this month, but its patents expire in 2015, and that’s assuming the U.S. Supreme Court upholds them. Meanwhile, the company plans to expand its offerings into Europe, where patent laws permit companies other than Myriad to perform the tests.

By now, though, the BRCA mutation data Myriad has collected is becoming more valuable than the original patents. When Myriad finds a new mutation, Cook-Deegan explains, it offers free testing to family members. These and other analyses allow it to interpret results of more variants. Gene testing by other companies finds variants of unknown significance in about 20 to 30% of cases, while Myriad’s rate is closer to 3% because it is able to assign significance to a greater variety of mutations. Since around 2005, mutation data and analysis algorithms have been available only to Myriad researchers.

“They haven’t done anything illegal, they found a hole that no one else thought of and built a business model around it,” says Cook-Deegan. But there are definite drawbacks for patients, he says. Those who do not use Myriad tests are more likely to be told that their results are uninterpretable, while those who use Myriad’s tests receive analysis that has not been vetted by the greater scientific community. (Myriad did not respond to Nature’s requests for comment. In a statement, Martina Cornel, chair of the European Society of Human Genetics’ Professional and Public Policy committee says, “we are very concerned that such important data is being withheld from those who most need it.”

Cook-Deegan and his co-authors argue for policies that would encourage data-sharing, such as health-care payers demanding the evidence behind clinical determinations. Alternately, regulators could require disclosure, or science funders could support the re-creation of such data. In fact, Richard Nussbaum of the University of California, San Francisco, is soliciting patients to submit their data and using it to build another, public database.

Another recent commentary from Domchek, Sean McElligot, and others at the University of Pennsylvania focuses on policy measures that would apply to testing in the United States, such as amendments to the U.S. Health Insurance Portability and Accountability Act. CLIA certification, required of clinical testing laboratories, could require data sharing. Or the FDA could be granted explicit authority over genetic testing and require disclosure for regulatory approval.

David Ledbetter, CSO of Geisinger Health System in Danville, Pennsylvania, is coordinating other projects that allow testing laboratories to share information about other genetic variants. The benefits not only allow researchers to study more variants but to develop systems and cross-checks that flag questionable results, he says.

Still, he does not believe that these issues are on the radar of health care payers; instead, he would like to see data sharing measures required as part of certification for clinical laboratories. “When your techniques or your data are private secrets, it’s difficult to participate in any quality control program.”


  1. Report this comment

    Thomas Ryder said:

    So researchers demand to have for free that which others have already paid for, often paid for by the very person from whom the researchers intend to accrue benefit by studying.

    And that isn’t even the most offensive aspect of the demand. It would be one thing if the phenotype/genotype associations could be done anonymously but the modern reality is that genotype infomation complex enough to have much value in association studies becomes a highly precise fingerprint for identifying a particular individual. No more unliked studies I’m afraid. I guess privacy is highly overrated – well, other people’s privacy anyway.

    But if the “researchers” want the data that badly, they do have recourse – they simply need to find out what it is worth to the other stakeholders, obtain informed consent and compensate them at some premium over what the stakeholders feel the information is worth.

Comments are closed.