23andMe uses a CLIA-certified lab to perform its test and so is within its rights, like all CLIA labs, to offer a test developed and run in its own lab (a lab-developed test, or LDT). FDA only regulates tests sold in kit form (IVDs), which can be performed in any clinical lab. The only thing "kit-like" in the 23andMe test is the sample collection swab, which, while unique for an LDT, seems like a stretch as a candidate to push the test into the kit category. 23andMe also markets the test extensively (exclusively?) to consumers, which is also unique for an LDT developer, but not established grounds for an LDT to require approval. While FDA has stated in the past that it assumes it has the authority to regulate LDTs, it continues to exercise what it calls "enforcement discretion" in allowing LDTs to be sold without approval, at least for now.
FDA has made numerous statements about the possibility of regulating LDTs, dating back to the 2006 draft guidance regarding "In Vitro Diagnostic Multivariate Index Assays," but until now it seems to all have been talk. Meetings have been convened, additional draft guidances have been issued, statements have been made, but no discernible action has been taken.
The most recent statements by FDA regarding the agency's declining to exercise enforcement discretion over LDTs include this summer's statement by Margaret Hamburg during a speech at ASCO that FDA "is working to make sure that the accuracy and clinical validity of high-risks tests are established before they come to market."
And now we have FDA's October, 2013 publication Paving the Way for Personalized Medicine (PWPM) in which the agency states that "FDA has been developing a risk-based framework for regulatory oversight of LDTs." Sounds encouraging, but also continues the tradition of only providing generalities.
For the sake of test developers, both IVD and LDT, and indeed for all stakeholders in personalized medicine, it is likely past time for FDA to make its intentions clear. Will FDA put high-risk LDTs through PMA? Will LDTs be required to become kits or will there be a new PMA category? Will there be a pre-market clearance process for lower risk LDTs (and for that matter, what is the future of 510(k) generally)?
These issues and many more need to be sorted, and soon, because the ambiguity is a significant drag on the development of novel tests. The agency acknowledges this drag by stating in the PWPM that "FDA believes that clarifying the regulatory oversight framework for IVDs would facilitate the development of in vitro diagnostic tests…" Indeed, how could it not. And the challenges are only going to grow, as the 23andMe case demonstrates.
IVD and LDT developers are already frustrated by reimbursement ambiguity and variability (a topic for another post), and regulatory ambiguity only adds to the disincentive to invest in novelty. Any business plan for a new IVD or LDT suffers tremendously from the necessary acknowledgement that the regulatory landscape is far from settled. As a Silicon Valley investment banker was recently telling me in relation to this topic, it used to be that you could afford 3 strikes, but these days 1 is more than enough.
And let us be fair to IVD developers that the current landscape challenges them more than it does LDT developers. The cost to develop an LDT is almost always significantly lower than the cost to develop the IVD against which it competes, so it should be no surprise that an average of only 3 novel targets per year have made their way into new IVDs over the last decade.
The risks/reward imbalance for IVD developers extends beyond the development side of the commercialization process, as post-market conditions are also more favorable for LDT developers since they do not have to conform to the guidelines that govern the marketing of IVDs, and so are more free to promote their tests directly to doctors and patients.
When FDA does act there is serious cause for concern about what additional effects new regulatory oversight could have on LDT innovation. FDA review periods for IVDs, both PMA and 510(k), have steadily lengthened over the last decade, while the total number of annual clearances has steadily declined. At the same time there has been no concurrent increase in the average complexity of IVDs submitted that might account for the review process getting longer, as complex, algorithm-interpreted tests tend to be offered far more often as LDTs than IVDs (Pathwork Diagnostics being a notable exception, and its recent failure will likely only encourage the trend).
Imposing the same challenged review processes and infrastructure that are used by FDA to clear and approve IVDs may bring LDTs onto the same playing field as IVDs, and may improve public safety (FDA stated in the PPM that it knows of harm having come to patients from poorly validated LDTs), but it may also contribute to stifling innovation in LDTs.
Indeed some of the factors that make the pre- and post-market conditions more attractive for LDTs might make sense for all tests. For example, tests that are validated using data from retrospective analyses of tissue libraries (common for LDTs) rather than validated by far more costly and lengthy prospective studies (often required of IVDs) may yield the same quality product, at least for lower risk tests.
One thing that LDT and IVD developers can agree on is that a lack of transparency is frustrating and stultifying. Medicare's inability to get its act together on test reimbursement is not just a drag on the market but is literally closing projects and companies down, and the lack of transparency in the rate-setting process only compounds these negative effects.
FDA has an opportunity to be a much more positive player in advancing personalized medicine, by clearly stating the steps it intends to take in regulating all tests, and by ensuring that those steps do the most they can to encourage innovation while not compromising public safety. Hopefully having to deal with tests such as 23andMe's, which blur the line between IVD and LDT, motivates FDA to define very clearly what the standards are for regulatory oversight.