The U.S. Food and Drug Administration recently held a public meeting of their direct-to-consumer genetics advisory panel (agenda here). Luckily it was held in Gaithersburg, Maryland, ahout 15 minutes from where I live. For more complete coverage of the events see the twitter hashtag #FDADTC and Dan Vorhaus' blog posts.
The point of the meeting was to gather information on what the FDA's stance should be on direct-to-consumer genetics tests, such as 23andMe's. There was a lot of discussion about whether such tests, if they give information of medical relevance, should be given only through a physician instead of being directly available to the consumer.
Unfortunately these kinds of tests are likely to be viewed differently by different people. For instance, I have a pretty good knowledge of genetics and human disease so I have no problem seeing my results directly without any filter by a physician or a genetic counselor. I'm guessing that most people who are 23andMe customers are like me. But there are a lot of people who don't know anything about genetics, or worse, are misinformed. There have been studies of how people react to the results of genetics test, and two of those studies were presented at this panel.
One of those was by researchers at the Scripps Research Institute, recently published in the New England Journal of Medicine. It says "the subjects in our study are clearly not representative of the broader U.S. population, and we therefore cannot draw conclusions about the effect of genomewide testing on the population at large." This is because they drew from a population local health and technology companies.
The second such study is the Multiplex Initiative which is being run by the NIH. There are no publications about that study yet so I have go by my memory. Colleen McBride, who runs the study, said that there were people who opted not to take the test once they read more about it on the project website. The study population seemed to more representative of the general population than the Scripps study in that it was drawn from patients at the Henry Ford Health System in Detroit, Michigan, USA. But the fact the people could opt out might have affected the results.
Both studies, in general, found that genetic test results did not any obvious adverse effects on the study populations. My opinion is that there are likely to be some issues when people who don't understand genetics get these results. Think what might happen if somebody got a BRCA result that indicated that they had a very high risk of getting breast cancer, at possibly a young age. People have responded to this type of result by getting preventative mastectomies, for instance. A person who had an incomplete understanding of the meaning of such a result, and their options, might react unpredictably.
But is that reason enough to restrict access to these tests? I confess to going back and forth on this all the time. Ideally anyone who wants access to such a test should be able to choose one and take one and interpret the results to their heart's content. People who didn't want to take one wouldn't have to, although there might be some situations where they would need one to help with diagnosis. And people who think they want one but don't really know what it is or how to interpret it should also be able to get one and get some help in understanding it. To summarize:
Wants test and understands it:
- No problem!
Doesn't want test:
- Fine, unless he/she needs a test to diagnose something.
- In that case it's still fine since a doctor would presumably order it since the person didn't want it.
Wants test but won't understand results:
- If they know they won't understand the results then they'll go find somebody to help them understand it. Maybe a doctor or maybe a genetic counselor.
- If they don't know they won't understand it, then maybe this is a problem.
So outside of the last one, everything seems to be fine. If people get convinced by advertising or a cool web page to get a genetic test, but aren't capable of understanding the results, they might have trouble if they get a result that is dramatic, like a greatly increased risk of breast cancer or Alzheimer's disease or Parkinson's. So how do we imagine this kind of risk will be addressed?
First of all, we to have some assurance that the results themselves are an accurate reflection of the underlying research. Of course, this is true for all the cases listed above, it's just that this last case is the one most vulnerable to these kinds of mistakes. My take is that should be solved by somebody, and it might not be the FDA, reviewing the risk models that companies are using. Maybe the American College of Medical Genetics or the American Society of Human Genetics could step up here? In addition, the genetics testing companies should step up and make their risk calculations more open, not just to an approval or regulatory body, but to everyone.
Secondly, there is a risk that somebody will misunderstand a dramatic result. Some might say that this is where doctors come in. Doctors could review the results and talk about them with patients and avoid the drama. This would work only if the doctor was knowledgable about the tests and their results. I've asked several of my doctors about 23andMe and they'd never heard of it. I think the genetics testing companies themselves should step up and do more to help. 23andMe currently has an extra level of access control on results such as this. For instance, when you look at your results for breast cancer for the first time, it asks you for approval before showing them to you. I suggest another step. If the results show high risk, I recommend that the testing company not reveal the results until the customer calls a genetic counselor, retained by the company, or by the patient, to discuss them.
Now it's worth pointing out that people who are skilled in bioinformatics could probably take their own genetic data and analyze it themselves. They wouldn't need 23andMe to tell them what risk alleles they had. But these people most likely already fall into the first category, those that want the test and understand the results.
If you've made it this far, I have one last point. A recent Genomes Unzipped blog post on this topic points out that treating genetics data differently than any other kind of medical data is genetics exceptionalism and that the tests should regulated based on their utility and the potential risk associated with the results. I agree with that completely, but of course, a whole genome test like 23andMe's contains a handful of results that can be considered dramatic and the rest of them are not likely to ruffle anyone's feathers. So do you regulate the entire test of just specific results? I think my suggestion above about particularly dramatic results being treated differently might be a good path to follow here.
But there would be one other way to fix this issue. Why not make all medical tests direct to consumer? For instance, to pick a relatively innocuous example, cholesterol testing. Why do I have to go to a doctor for that test? Because only a doctor knows when I need one? Because I can't interpret the results? I think in this example plenty of people know when to get one and how to interpret the results. Again, this is not true for everybody, but why limit access based on only a part of the population?
Many tests are of such as specialized nature that the average person, even with general medical knowledge, might not be qualified to call for and assess them. Many of them are also expensive so that insurance companies would want some assurance that the test was relevant before paying for it. But again, an educated person might know when it was justified and be able to make the case directly to the insurance company.
When I started this post I wasn't sure where it would end up, and it ended up in a somewhat unexpected place. To be clear, my thinking on this issue is constantly evolving. Any comments or insights from physicians, consumers, regulators, or geneticists welcome.