Name: Michelle Mello
Title: Director of the Program in Law and Public Health, Harvard School of Public Health;
Chair of Institutional Review Board;
Professor of law and public health, Department of Public Health and Management, Harvard School of Public Health
Education: JD, Yale Law School;
PhD in health policy, University of North Carolina at Chapel Hill;
M.Phil, Oxford University;
BA, Stanford University
The noninvasive prenatal testing market has exploded since Sequenom first launched its MaterniT21 test in October 2011. Currently, there are four companies in the US with sequencing-based tests that screen for fetal aneuploidies — Sequenom, Ariosa Diagnostics, Natera, and Illumina-owned Verinata Health — as well as companies in Europe and Asia.
Patients are rapidly adopting these tests, with both Sequenom and Ariosa reporting annualized test run rates of over 100,000. Neither Verinata nor Natera have disclosed their test run rate.
Yet, despite the benefits of these tests over traditional screening methods — they are noninvasive and require only a blood draw, can be done as early as nine weeks gestation, and have low false positive and false negative rates — the rapid adoption has sparked concern among some experts.
Michelle Mello, director of the Program of Law and Public Health at Harvard School of Public Health, recently laid out such concerns in a perspective piece published in the New England Journal of Medicine.
The foremost concern, she and her colleagues wrote, is that physicians and patients do not fully understand what the test can and cannot do, particularly when it comes to an important statistic known as positive predictive value.
Clinical Sequencing News recently spoke with Mello about the importance of these tests' positive predictive values and why physicians and patients should not be satisfied with simply knowing the tests' sensitivities and specificities.
First, can you give me a little background about how and why you became interested in this issue?
I'm interested generally in issues around the commercialization of diagnostics. I got interested in this specific topic, frankly, just from personal experience. I had a lot of low-risk friends commenting that they had been offered the test, which I understood had been marked by professional organizations for high-risk women and I got interested in how that practice had started.
What is the problem with the way in which these tests are being offered?
It's really more a matter that there is a disjunct between where the professional guidelines are in terms of the types of women for whom they recommend this test, and where the marketing is.
There are three leading professional organizations that have all come down in favor of offering this for high-risk women, although that term high-risk is not well defined. In contrast, there's not any specificity in the marketing of these tests for low risk women.
That's relevant because although the tests represent an exciting advance over standard testing, most of the validation has been in populations of not just high-risk women but in biological samples that have been enriched to be super high risk, in the sense that they deliberately contain a large number of abnormal karyotypes. As we explain in the paper, that really affects how you calculate an important piece of statistical information that women should care about — positive predictive value.
The information we have about most of these tests comes from studies that involve samples [with a high frequency of] genetic abnormalities. And that information doesn't necessarily indicate how well these tests are going to work for lower risk women.
Why is the positive predictive value so important, and how does it compare to specificity and sensitivity?
There are some really important measures in epidemiology for which the population prevalence of the condition does not matter. Sensitivity and specificity of the test are two of them, and [those values] are what the companies really emphasize in their marketing materials, and those properties of a test are really important. However, from the patient's perspective, what they want to know is, if I test positive, what is the probability that my baby really has the genetic disorder?
The question you're going to ask yourself when you get a positive test back, is what should I do? And, the answer to that might be really different if you have high confidence that the test is right, versus if you have low confidence that the test is right.
That's what is going to matter to a lot of women. And that's called the positive predictive value. What you calculate for that depends very much on how prevalent the health condition is in the group of people from which you're calculating it.
There's nothing wrong with what the companies have said about the sensitivity and specificity of their tests. We're just making a point that those are not the only things that patients should care about. And, one of the things that patients should care about a lot has not really been investigated in the right group of people.
How do you propose that these tests be used?
My sense is that clinical practice ought to be evidence based. If the evidence is there to tell us that the benefits outweigh the risks and negatives of this test for certain populations, there ought to be a recommendation for it, provided it's also cost-justified. It seems like we're there for certain groups of women, but not for others. I would say that if you want to market this test to low-risk women, you ought to commit to doing the kinds of studies that demonstrate the utility of the test for that population. Companies have started to do that but not all have, and as we make the point in the article, [testing among low-risk women] has already begun to diffuse out into practice among a lot of obstetricians.
How could potential regulation by the US Food and Drug Administration impact the use of these tests, either if the FDA decides to regulate LDTs or if there is 510(k) cleared version of the test?
I don't know if I've reached a conclusion about that. There are some real pros and cons to being more heavily regulated. Probably if you talk to the companies, they would say that being relatively free from the strictures of FDA approval has enabled them to innovate very rapidly. Indeed, the growth in this space has really been tremendous in a short period of time. Even just in the six months since we started writing the article there's been a lot of innovation in terms of what these companies can do and get out on the market.
What you're losing when you impose additional regulation is that flexibility. It's going to take companies longer to develop the evidence base needed to get FDA approval than it would if they didn't have to get FDA approval.
But, what you're gaining is certainty around the test. So, it's a social tradeoff that I think we just maybe haven't grappled with as fully as we need to as more and more of these genetic diagnostics come online.
How do you foresee noninvasive prenatal tests being implemented into clinical practice going forward?
I suspect that things will ultimately settle down. This is a really competitive marketplace, so the companies have a lot of incentive even apart from FDA regulation to try and do the studies necessary to demonstrate that their test is best. And they seem to be doing that. It's just that the practice has gotten out ahead even of that quite rapid evidence-based development.
I suspect if we look back in five to seven years, things will have equilibriated a little bit. But in the meantime, I think it's important for obstetricians, and also for women, to understand that there are cons to these tests as well as pros.
Do you think that there needs to be more physician education around these tests?
I suspect that there is a little bit of confusion on the part of physicians. These epidemiologic concepts are not easy to grasp and most physicians have not had a great deal of formal training in epidemiology. So unless somebody really instructs them on it, they may not appreciate this nuance. And, this hasn't been a thrust of the companies' marketing materials, nor is it really even brought out in the guidelines that are out there on this issue.
It would be helpful to have some clear statement from an authoritative body that explains to the physician what you should be thinking about, and explaining the tradeoffs to patients. And then patients also of course have to participate. The good news is, this is one decision where patients are already pretty engaged decision makers. For a long time they have had to make difficult decisions about screening. The availability of these new tests just puts a sharper point on those decisions and makes it more complex. But people at this stage of their lives are pretty active about making the right decisions about their babies. I think there is a good prospect for there to be some active decision making if we can get folks to understand what the factors are that are on each side of the ledger.