Last week preliminary data from a Scripps Health study, looking at effects of DTC genomewide testing, were published in the New England Journal of Medicine (link to pdf article). The study represents the first published data of this kind. Up until this point, the dialogue surrounding the potential benefits and harms of DTC testing has been mostly anecdotal guess-work. So, needless to say, these results are important. Media outlets were quick to report that ‘consumers can handle the truth’ and that testing has no impact on health behaviour. But, beyond snappy headlines, these results warrant a closer look. I thought I’d give a quick run-down of my reaction here, in the hopes of getting a good discussion going.
In full disclosure, I recently joined The Medcan Clinic in Toronto as a genetic counsellor. We offer personal genome testing using the Navigenics platform in the context of a comprehensive genetics assessment (see ‘Putting GC into DTC’ guest post from last year for details on this model.) As you can imagine, I am particularly interested in these findings.
Some important things regarding study design:
- Study subjects participated in health assessments (assessing dietary fat intake, exercise behaviour, anxiety symptoms and uptake of screening tests) using an electronic survey tool. No physical exam or blood work was taken into account in assessing baseline or follow-up parameters here.
- These results represent data from the baseline assessment and a 3-month follow-up. This is a 20-year longitudinal study, so essentially this data is the tip of the iceberg.
- The study protocol used the Navigenics Health Compass testing, but Navigenics did not provide any financial support for the study, nor were they part of the study design, analysis of data or manuscript preparation.
The most interesting points (as I see it):
- At 3-month follow-up, there were no significant health behavior changes made by study participants measured by amount of fat intake and exercise behavior, except for in the 26.5% of participants who reported sharing their results with their doctor. These participants did have lower fat intake and increased exercise activity.
- Those who shared their results with a Navigenics genetic counselor only (10%) did not show any significant behavior change.
- At 3-month follow-up, there was no significantly increased anxiety or test related distress. Whether or not an individual had genetic counseling did not affect this parameter.
What I take from this:
- Sharing results with a physician is more likely to impact health behavior. This lends evidence against the direct-to-consumer model.
- As genetic counselors, maybe we need to be more focused on the potential for us to add value to consumers of genome wide testing, and less focussed on the potential psychological harms of the testing.
- We all know these tests do not take into account family history. There is a role of GCs to help consumers understand their risks in the context of their family history, to assist consumers in sharing this information with their doctors, and to work with general practitioners to help integrate this information into their practice.
What I can’t wait to see:
- More long term data! This is predisposition testing combined with 3-month follow-up info. I am curious to see whether more subjects decide to share this information with physicians down the road (presumably many did not have doctor’s visits scheduled in the months immediately following receipt of their results). I’m also curious about whether consumers or physicians will use this information to guide future investigations, when these subjects experience an issue requiring attention.
There are obviously a huge number of issues beyond those I’ve highlighted here. Please share your thoughts and reactions below.
The DNA Exchange 
Great post, Allie.
One quick comment:
Not really. It seems plausible (in fact probable) that the types of people who are likely to share their results with a doctor would also be more likely to change their behaviour based on genetic results; these data can’t be used to show that sharing with a doctor, in and of itself, was actually helpful. Correlation != causation, etc.
However, this highlights an important follow-up experiment that someone needs to do: assign DTC customers at random to a group that receive detailed genetic counselling and a group left to their own devices, and then measure long-term outcomes. I don’t suppose this is being done for the Medcan study? If not, is anyone aware of researchers adopting this study design?
Daniel, thanks for the comment and clarification. I should have said that it helps to strengthen the argument for physician involvement over no physician involvement, but as you pointed out, this is more personal opinion rather than fact!
I agree that a randomized control DTC group vs. genetic counselling group would be interesting. Everyone who undergoes personal genome testing at Medcan meets with a genetic counsellor before and after the testing, so we don’t have a group here with no genetic counselling at all. We have some data regarding outcomes for individuals who undergo testing, and the intention is to publish it at some point. I’d also be interested to hear of other researchers looking at this type of study.
Good to see a professionals perspective and reading of the results. It’s not surprising that it required help of a counsellor or physician in order to have significant behavioural change. This is seen in most aspects of health and is particularly well documented for weight management – attempts at weight loss without professional help tend to fail.
I don’t agree though that it argues against the DTC model – this gives the choice to the individual whether to and which information to share with the professional. Having said that, the DTC model so far has not been successful in Europe, with most buyers preferring to go through some sort of practitioner (doctor, nutritionist, etc).
I absolutely agree with the potential for the counsellor to add value – much more so than a medical doctor and I think you are correct that the counsellor profession has the opportunity to approach this area with a very positive outlook – as you say, this study says that personal genetics can be useful and lead to behavioural change when the counsellor is involved. But this high percentage of success is probably due to the DTC user wanting and seeking out the help of the counsellor.
Thanks Keith– I appreciate the feedback!
The always-on-top-of-it Beth Balkite alerted me to a NY Times site that offers a learning/discussion module for a classroom discussion of direct-to-consumer genetic testing. The module does an excellent job of covering the issues that DTC testing raises. Check it out at:
http://learning.blogs.nytimes.com/2011/01/19/the-pros-and-cons-of-genetic-data-debating-personal-dna-testing/
See also the Jan 2011 ish of Genetics in Medicine for a nice commentary by Wylie Burke and James Evans on another article in the same issue by David Walt et al in which medical students use (anonymous) SNP testing through a DTC company and utilize the results as a focus for a class discussion on the ins and outs of SNPs and their role in medical care.