The highest profile direct-to-consumer genetic testing company in the world, 23andMe, has opened shop in Canada. In addition to providing information about your genetic ancestry (turns out I am 100% Irish), 23andMe will, after your genetic information is pulled from the required tube of spit, provide you with a whole bunch of genetic risk information that may, or may not, be relevant to your health.
Should Canadians care? Should Health Canada convene an emergency session to shut down 23andMe, just as the FDA did in the United States in 2013?
I have been following the genetic testing story for years and one clear and consistent theme has emerged: there is a lot of hype and misinformation on both sides of the debate.
Let’s start with what the evidence tells us about promised benefits. At the core of the direct to consumer business model – whether it is 23andMe or one of the many less well-known outfits promising to genetically-personalize your diet, workout or love life – is the idea that receiving your genetic information will empower you to make more informed decisions. You will get your genes tested, find out if you are at increased risk for heart disease or some form of cancer and change your lifestyle. You will go for long runs with your dog. You will eat more fresh fruit and vegetables. You will lose weight, drink in moderation and quit smoking.
Now, I am not going to say that these direct to consumer genetic services are totally ineffective for all. Some may find this information “empowering”. (Heck, some people feel empowered by their horoscope.) But there is absolutely no reason to believe that providing genetic risk information will, for most, lead to healthier decisions or motivate these kinds of changes.
First, the vast majority of the information provided by these companies – or by genetic testing generally – simply isn’t that predictive. Indeed, I got my genes tested by 23andMe in 2011. It was a fun and intriguing process – a bit like eavesdropping on a couple of strangers talking about you. (They are right! I do have blue eyes, my hair is kinda curly and I am slightly taller than average!) But the information I received from 23andMe was (and will forever remain) far less predictive of future health problems than, say, my weight, blood pressure, family history or postal code.
While there are a few well-known genetic mutations that are highly predictive of a future disease, these mutations are relatively rare. Most genetic risk information is largely useless when set against other, well-established risk factors such as smoking, obesity and inactivity. For example, a recent study out of Cambridge published in PLoS Medicine found that the weight of an individual is far more predictive of type 2 diabetes than genetic risk information. As a result, the authors of the study conclude by noting the “importance of universal rather than targeted [i.e., genetically-informed] approaches to lifestyle intervention.”
Second, we should also not forget how incredibly difficult it is to change behaviour. Providing a bit of risk information – whether it comes from your bathroom weigh scale or a multi-million-dollar gene sequencer – is unlikely to have a large impact on population health. Studies have consistently found that genetic risk does not lead to dramatic changes. To cite just a few examples, it does not seem to help smokers quit or the obese lose weight. And a 2013 study on point came to the conclusion that genetic risk counselling does not significantly alter motivation or adherence to weight loss strategies for individuals at risk for diabetes.
OK, so the benefits – at least in terms of lifestyle changes – seem marginal or non-existent, but what about the harms? Just as there is little evidence that people benefit from receiving genetic risk information, there is also little evidence of significant harm. Studies have found, for example, that people do not freak out when they receive their results and they don’t behave fatalistically to bad genetic news (e.g., “bring on the fries, I’m gonna die anyway!”). In addition, despite lots of hand waving (and fear) to the contrary, there is very little solid evidence of widespread genetic discrimination.
There are, of course, other problems that may be associated with direct to consumer genetic testing, including increased (and perhaps unnecessary) utilization of health care resources and an inappropriate framing of chronic disease as a personal, rather than a social, concern. But our first response to the 23andMe development should be to take a deep breath and dial back the exaggerated claims of benefit and harm. Let’s develop an evidence-based direct to consumer policy.
When I got my results from 23andMe I found out that I am at (ever so slightly) increased risk for a whole bunch of nasty things, including coronary heart disease and various cancers. My personalized advice? Drink in moderation, exercise, watch my weight, don’t smoke and eat a healthy diet that includes lots of fruits and vegetables. Solid. Advice we should all follow. No test required.
The comments section is closed.
In your last paragraph you mention your personalized advice. Many people do that and still end up with cancer or some other health problem related to having done all of the right things. There are no guarantees. I am not disputing that is great advice, I am just wondering what then when you have done all that and you still end up with disease.
What is your take on GMO’s. Apparently the governments won’t let anyone do any studies on GMO’s and their affect on health. And apparently they have approved the sale of them without government studies. Is this so. Frankly I am tired of hearing about this.
Linda Swityk
I agree with a lot of what you said Linda, but there’s one thing I do disagree with…
You said, “I am not disputing that is great advice, I am just wondering what then when you have done all that and you still end up with disease.” I think you were referring to the comment in the article of “eating a healthy diet that includes lots of fruits and vegetables. Solid. Advice we should all follow. No test required.”
Well, if you’ve ever read anything about Genetics then you’d know that not everyone is capable of breaking down the foods associated with what’s considered to be a healthy diet. Some genetic deficiencies made some people incapable of eating whole wheats, others are lactose intolerant. There’s there’s the allergies to nuts and other healthy foods that are capable of even killing some people when ingested. But there’s other things that aren’t even discussed in mainstream conversations… what about the effects of certain people taking certain “one med fits all medications” and then mixing them with high levels of natural glucose found in a high majority of our healthy foods?
If there’s anything out there that offers society with an all in one picture of what they’re genetic makeup is in need of, or needs to avoid, then it’s to be found in genetic testing. Personal medicine is still years away from where we all need it to be if we hope to live long, healthy lives.
As for what happens to people when they live a healthy lifestyle and still get sick? Well, answers to those questions can be found in the many studies in the field of Epigenetics. We truly are affected by our ancestors, and past social norms, in more ways than anyone ever considered, but scientific research is figuring it out slowly.
There’s a lot of good documentaries on youtube… look up “The Ghosts in Our Genes” if you’re actually interested in how our genes affect our lives then you’ll find that interesting.
There’s also several books out there, that are well written by some Geneticists, and they’re written in layman’s terms that makes it easy for anyone, of any level to understand.
Dr. Sharon Mualem wrote several books and each of them are easy to get and even making learning fun.
There’s lots of information out there that can help anyone, not just researchers or persons touched by rare genetic variances. They can all help us gain further insight into what’s been learned and what Geneticists still need to know. I think most people would be surprised to find there’s a lot more knowledge about genetics than any of us ever thought possible.
Through all I’ve learned, I’m convinced the solutions lay in the words; “personalized healthcare”. A type of care offered by genetic testing that’s specific to each unique individuals genetic makeup.
First, it would be instructive to see a sample of the test received. It certainly isn’t true that healthcare is a ‘social rather than individual’ endevour, not even close. What I read here hardly seems like grounds to make such a practise ‘illegal’ as in the US. That just seems bizarre. But the more important aspect not mentioned here is the security. This private company now holds your genetic information. You may trust them to ‘not keep it’ or something, but thats a big trust.
It’s a Very big trust to allow your health information be “out there”, but it’s already out there due to the research and clinical studies I’ve had to be part of in my attempt to receive a proper diagnosis. Plus, I already face discrimination in Canada by insurance companies due to my extensive list of pre-existing conditions.
There is currently no legislation in Canada that protects Canadians from discrimination, while there is in the US and the EU. I’m a Canadian whose Genome is not protected, but I’ve had to be truthful on what I live with in case I need care here or when travelling.
For me personally, I’ve learned the advantages of letting my data be available for review. It can help Doctors and researchers look at the many aspects of genetic effects of people of all ages, races, and cultures. If my information helps at least one other person besides me, then I’ll consider it well worth the risk.
I would never wish my health on my worst enemy, so why not try to take part in the research required to help future generations? My suffering has to be worth something more than the money made by pharmaceutical companies….. Dont get me wrong. I’m very grateful for the medication options, but I’d much prefer to see someone bypass all the risks associated with continual medication use.
The American Academy of OphthalmologyRecommendations for Genetic Testing of Inherited Eye Diseases – 2014
The following are a few of the Specific Recommendations
Offer genetic testing to patients with clinical findings suggestive of a Mendelian disorder whose causative gene(s) have been identified. If unfamiliar with such testing, refer the patient to a physician or counselor who is. In all cases, ensure that the patient receives counseling from a physician with expertise in inherited disease or a certified genetic counselor.
Avoid direct-to-consumer genetic testing and discourage patients from obtaining such tests themselves. Encourage the involvement of a trained physician, genetic counselor, or both for all genetic tests so that appropriate interpretation and counseling can be provided.
Avoid routine genetic testing for genetically complex disorders like age-related macular degeneration and late-onset primary open-angle glaucoma until specific treatment or surveillance strategies have been shown in 1 or more published prospective clinical trials to be of benefit to individuals with specific disease-associated genotypes. In the meantime, confine the genotyping of such patients to research studies.
Avoid testing asymptomatic minors for untreatable disorders except in extraordinary circumstances.
More recommendations and the full guideline are here:
http://one.aao.org/clinical-statement/recommendations-genetic-testing-of-inherited-eye-d
As a genetic counsellor, I am wary of direct-to-consumer genetic testing, primarily because of the lack of context in which the information is provided. It is important to know, however, that the current version of the 23andMe test being marketed in Canada is significantly scaled back from the original version that you would have taken several years ago. While they still provide trait information, they have removed all of their common complex disease data, and are instead focusing on well-known and generally more rare genetic variants. This poses new issues for the genetics community, but it is an interesting move for this company which is clearly trying to gain approval in the eyes of the medical community and FDA. The updated list of conditions can be found here: https://www.23andme.com/en-ca/health/all/
Tim, Thanks for this analysis. We should all pass it on to our Members of Parliament for their education – a starting point for evidence-based decisions regarding genetic profiling. The only point I would add has been noted by Stuart Hogarth below, and that is the need to recognise that genetic risk can be combined with other risk factors (e.g. life-style, family history, environmental exposure) to give an overall assessment of risk, standard procedure in any genetic counselling facility, but sadly unavailable in a genome screen that cannot take into account the pre-test “a priori” risk .
Well there is a large difference between analysing close to 1 million well know SNP locations and providing you your entire exome sequence. The raw data will have very important data contained within it. Even if today we can’t make sense of it, Future developments will allow you to reannotate your exome periodically and find new things on your existing data.
I for one welcome the chance to do this and will be pissed if the UK government blocks this from happening here (whenever it reaches our shores)
Hi Tim
Nice post. I am all for evidence-based policy, but I would also like evidence-based genetic testing.
Your post highlights low odd ratios tied to standard healthy lifestyle advice as evidence of hype, which, as you know, is an argument I am fully in agreement with.
However, there is a more fundamental problem – have they got the science right? Clearly companies like 23andme know how to read the scientific literature but what are their protocols for which genes to include and which to exclude; what is the algorithm which underpins their polygenic risk scores; and how can you give an individual a risk score if you do not have the data to understand their pre-test probability of disease?
The debate about these tests has got far too focused on the question of whether we can find evidence of harm, and is paying far too little attention to the issue of whether the risk predictions are accurate. Absent the latter then we are back to the argument that the most immediate harm posed by these services is that you are wasting your money on something a test which may be quite misleading.
This is interesting. I’ve always wanted to know more about my ancestry, but not disease risk or what have you.
I wonder if there is a way 23andme can provide lineage assessment alone? I do not want my genetic data falling into the hands of disability and life insurance companies.
Deep ancestry (e.g. We are all African if you go back far enough) is available for approx $100 from The Human Genome Project through National Grographic. Members of my family have done this mail away saliva test. Fun! Also interesting. For example the branch of the family that thought they were pure Scandinavian turned out to be most closely related to Mediteranian Semetics. It really changed how we looked at our ancestry which turns out to vary according to how far back we wished to go. Made us feel more connected to the entire human race.
Another issue concerns whether samples should be subject to re-interrogation as technology develops? Experience in the direct-to-consumer personal genomics indicates that unanticipated changes to risk estimates, based on the inclusion of new variants, may be both surprising and concerning (Aldhous 2009; Henderson 2009). As one recent commentary put it “The DNA you are born with stays with you for life, but health predictions based on it can alter dramatically.”(Aldhous 2009). Such changes have the potential to substantially effect individual risk through risk-reclassification. A study by Mihaescu et al., is particularly illustrative, with the authors noting that when risk assessments for type 2 diabetes based on a single genetic factor (TCF7L2) was supplemented with the addition of 18 other polymorphisms, 34% of participants underwent a risk reclassification. Of those classified as below average risk on their TCF7L2 status, 24%were reclassified as above average risk, while 44% of those initially classified as above average risk were reclassified as below average risk. In part this related to the fact that some variants were protective against type 2 diabetes (Mihaescu, van Hoek et al. 2009). In a further analysis in which 1000 simulations of random permutations of possible orderings of polymorphisms were calculated, with polymorphisms added individually, on average 47% of cases underwent at least one risk reclassification. Furthermore, on the basis that protective or defective polymorphisms could be added in varying amounts – to reflect the potential for stepwise changes in technological advances – 30% of simulations switched risk category multiple time, with a range of between 2 and 15 changes (Mihaescu, van Hoek et al. 2009). This not only points to the potential for substantial changes in risk classification over time as technology advances and increases the number of analyses that samples may undergo, but also raises important clinical and policy questions regarding the management, interpretation, and communication of this information. Moreover, it raises the question of responsibility on the part of the healthcare professionals involved in the delivery of genomic testing. Indeed, at least one author has argued that, compared to DTC companies, physicians or health care institutions have an increased responsibility for their patients deriving from their professional duty (Bunnik, Schermer et al. 2011).
Aldhous, P. (2009). “Gene discoveries dramatically alter disease predictions.” New Scientist 203(2719): 12.
Bunnik, E. M., M. H. Schermer, et al. (2012). “The role of disease characteristics in the ethical debate on personal genome testing.” BMC Med Genomics 5: 4.
Henderson, M. (2009). Consumer genomics: your genes don’t change, but your disease risk still might. The Times. London.
Mihaescu, R., M. van Hoek, et al. (2009). “Evaluation of risk prediction updates from commercial genome-wide scans.” Genetics in Medicine 11(8): 588-594.