By Turna Ray
"Post-CLIA bioinformatics services" — a term introduced in a California bill seeking to legally set apart consumer genomics firms from clinical laboratories — may be too broad a term that could have unintended consequences for the broader bioinformatics field, according to some experts.
The bill, SB 482, introduced by California state Senator Alex Padilla (D-San Fernando Valley) earlier this year and sponsored by California-based direct-to-consumer personal genomics firm 23andMe, would amend the California Business and Professions Code to exempt so-called “post-CLIA bioinformatics services” from certain requirements that clinical laboratory services providers have to meet under the Clinical Laboratory Improvement Amendment [see PGx Reporter 06-24-2009].
Given that SB 482 would essentially codify new laws and new terms by which to regulate the consumer genomics industry, at least one leading bioinformatics researcher has found the term “post-CLIA bioinformatics services” confusing and the scope of the bill potentially problematic for players in the bioinformatics field — particularly those who are analyzing patient data for research purposes.
“Some of the legal language is tough for me to parse … I don’t understand what they’re trying to say in this [bill],” Atul Butte, associate director of the translational informatics program at the Stanford Center for Clinical and Translational Education and Research, told Pharmacogenomics Reporter last week. “Are they trying to tie bioinformatics services to clinical labs? Are they saying that if I offered a service to interpret [genomic] data that I wouldn’t be able to do it unless I’m partnering with this kind of a licensed clinical lab? What exactly are they making illegal?”
Sheila Walcoff, a partner at the law firm McDermott Will & Emery, agreed that the language in SB 482 could have implications for the bioinformatics field. Consumer genomics “is a new industry, and should California statutorily define and regulate it in a new or different manner, it could, at least at the state level, have an unanticipated impact on the regulatory requirements for the broader bioinformatics industry,” Walcoff told Pharmacogenomics Reporter this week.
The bill, which has been sitting with the California Senate Judiciary Committee since April 30, defines “post-CLIA bioinformatics services" as the “postproduction interpretation, by means of an algorithm, of biological data.” Clarifying this further, the legislation also defines “algorithm” and “biological data.” However, SB 482 does not specify what “bioinformatics” means in this regard.
A 23andMe spokesperson told Pharmacogenomics Reporter this week that the term “post-CLIA bioinformatics services” is “supposed to refer to algorithmic interpretation done by computer after the actual generation of the raw data in a CLIA lab. The idea is that CLIA regulations don't cover bioinformatics and we wanted to emphasize that they're distinct.”
While such services would be exempt under SB 482 from meeting CLIA requirements, the bill would require them to have a contractual relationship with a “licensed clinical laboratory” for receiving and handling biological specimens; disclose to customers the CLIA status of the laboratories being utilized; and employ an individual with a Master's degree or a PhD to approve algorithms used to interpret genomic data.
Furthermore, under SB 482, “post-CLIA bioinformatics services” would have to provide the California Department of Public Health with a "transparent description of the validity of biological data sets and how to perform the algorithm on biological data sets." Although the bill requires these companies to form an external physician advisory board to provide guidance on the interpretation or presentation of analyses to customers, SB 482 does not require approval or review of the algorithm by a government regulatory body.
In Butte’s view, the bill and the terms within raise questions about the impact of SB 482 on bioinformatics researchers like himself who are using investigational algorithms to interpret genetic data generated from laboratory-developed tests that aren’t necessarily CLIA-certified.
“This law doesn’t just cover companies, it covers what’s done in academic institutions, too,” Butte said. “Nothing in this bill blocks that.”
For example, Butte and a team of academic researchers are currently sequencing a particular patient's entire genome with the intention of publishing a paper on their findings. “Would that be illegal by the regulations set forth in this bill?” Butte posited. “I’m not sure if [our research] was done in a CLIA-certified lab, probably not. It was done on a research-grade machine. And I’m providing a bioinformatics service, perhaps as a post-CLIA, I’d be concerned as an academician whether that would actually now be illegal.”
Yet, consumer genomics firms are making a business out of selling genetic risk data, which by their own admission is in the research phase. Consumer genomics firms such as 23andMe, Navigenics, and Decode Genetics issue disclaimers to customer that the genetic risk data sold by their services is for research use or for informational purposes only [see PGx Reporter 06-04-2008].
[ pagebreak ]
Critics of the nascent consumer genomics industry have particularly taken issue with the fact that some of the algorithms and genetic risk associations used by the firms are not sufficiently validated. Since SB 482 doesn’t insist that these algorithms be approved by a government regulatory body, but requires clearance from an external physician advisory board, it raises questions about how other entities using such algorithms would be impacted.
“These algorithms are still a work in progress in most cases, and how [,if] regulated, would the academics be trying to continue to refine the science here?” Butte questioned.
Furthermore, Butte took issue with the level of education SB 482 requires for those handling questions about the algorithm and approving documentation describing the algorithm.
According to SB 482, “an entity providing post-CLIA bioinformatics services shall designate an individual with a PhD or Master's degree in bioinformatics, statistical genetics, biostatistics, or statistics with a biological or medical specialization to be responsible for approving algorithms and documentation thereof and to serve as the point of contact for questions pertaining to the algorithm.”
“This is saying you need a minimum training, but it’s not clear what minimum training should go into this certification. There is no certification set up here,” Butte said. “Also, they don’t specifically list an MD, needing a medical degree, so that doesn’t count I guess.”
Finally, Butte questioned why SB 482 has been introduced in the California legislature and not in Congress. “Why is this a California issue instead of an FDA issue?” Butte asked.
A Patchwork Approach
Currently, clinical laboratory services are overseen by the Centers for Medicare & Medicaid Services under CLIA. CMS has said it is working with consumer genomics firms to see where CLIA may be applicable. The US Food and Drug Administration has only said it is watching the nascent consumer genomics industry with interest. Similarly, the Federal Trade Commission has also said it is investigating the marketing practices of DTC consumer genomics firms.
SB 482 comes after health regulators in New York and California sent cease-and-desist letters to several consumer genomics firms, among them 23andMe, Navigenics, and Decode. New York and California have some of the most stringent certification requirements for clinical labs, and as a result, consumer genomics firms feared at the time that other states might follow the lead of these two states in clamping down on the industry [see PGx Reporter 06-25-2008].
SB 482 is essentially the first legislative effort attempting to legally describe what consumer genomics firms do. As such, if the term “post-CLIA bioinformatics services” gets codified in California law, it may be used by other state and federal regulators going forward.
As seen with the confusion surrounding FDA’s expressed intent to oversee so-called in vitro diagnostic multivariate index assays, a subset of complex, algorithm-based laboratory-developed tests, the introduction of new regulatory terms has the power to turn an industry upside down, sometimes requiring increased investment to ensure that products meet legal definitions. Even after the FDA held a public hearing on the subject and issued two draft guidelines attempting to define IVDMIAs, most diagnostic developers still maintain the agency’s proposed regulations are confusing and a disincentive to innovation [see PGx Reporter 08-01-2007].
Absent federal regulations, which take much longer to develop, 23andMe may have decided to introduce SB 482 on the state level since it’s based in Mountain View, Calif.
[ pagebreak ]
“Though interest at the federal level has been largely limited … legislative action by California could generate more specific action by federal regulators and legislators,” McDermott Will & Emery’s Walcoff said.
When SB 482 was first introduced in April, patient groups objected that the bill did not protect customers’ privacy sufficiently. At the time, there were rumblings that if it looks like this legislation might pass in the California legislature, then federal authorities will get involved in regulating the industry.
Coincidentally, as federal authorities are watching this rapidly evolving industry, industry stakeholders and state regulators may be looking to the feds for guidance.
23andMe informed Pharmacogenomics Reporter this week that “while agreement has been reached on some key issues” with regard to legislative action in this area, “more work is needed, and state action is somewhat dependent on whether there will be any action at the federal level.”
In the past, the heads of several consumer genomics firms have spoken out on the need for federal regulations to guide the nascent industry before technological advances make the issue unwieldy and a patchwork of state laws emerge, further confusing consumers.
As seen with the Genetic Information Nondiscrimination Act of 2008, it took more than a dozen years to pass a federal law providing baseline protections against employment and health insurance discrimination based on one’s genetic data. Meanwhile, a tangle of state laws had been issued, which, some have reported, deterred people uncertain of their rights from participating in research studies involving genetics [see PGx Reporter 05-07-2008].
Not wanting to let slow regulatory processes inhibit advancements in the consumer genomics field, industry leaders are trying to self-regulate as much as they can by coming up with their own standards.
During a December conference on consumer genomics hosted by the Centers for Disease Control and Prevention and the National Institutes of Health, representatives from 23andMe, Decode Genetics, and Navigenics presented voluntary industry standards they had developed with the help of the Personalized Medicine Coalition [see PGx Reporter 04-15-2009]. Still, Muin Khoury, director of the Office of Public Health Genomics at the CDC, has said that more work needs to be done to standardize clinical validation of gene-risk associations in these firms.
In response to criticism that SB 482 doesn’t adequately protect consumers’ privacy, bill sponsor 23andMe has defended that the legislation goes further in providing privacy protections than current law does. Furthermore, the bill is subject to changes as discussions with patient groups and regulators continue, the company has assured.
"The current language is a starting point, not a final product," a 23andMe spokesperson told PGx Reporter in June. "We look forward to continuing to discuss and refine SB 482."