The personal genetics company 23andMe has begun offering its customers a new report informing them of their likelihood of developing type 2 diabetes.
The report was developed and validated using data 23andMe has gathered from over 2.5 million of its users who have opted into research. The information was used to create a polygenic score from over 1000 genetic variants associated with type 2 diabetes.
The type 2 diabetes score has just been added to the company’s Health & Ancestry service, which costs $199 and also provides information on genetic risks for some cancers, familial hypercholesterolemia, celiac disease, and some forms of dementia and Parkinson disease, among other conditions.
A saliva collection kit is mailed to the customer’s home. Users spit into the tube provided, add a barcode to their sample, and mail it to the company’s laboratory in a pre-paid package. Results are available in 3 to 5 weeks by email, which includes a link to a report.
The type 2 diabetes report informs customers about how factors such as age, weight, and lifestyle affect their risk of the disease and actions they can take to mitigate the risk, along with educational resources.
Links are also provided to 23andMe collaborator Lark Health, an artificial intelligence platform that offers diabetes prevention counseling.
But whether or not consumers are equipped to know what to do with such information, without any medical professional input, has been the subject of debate, as reported by Medscape Medical News.
“It is our hope that receiving personalized information about type 2 diabetes predisposition and prevention can motivate and facilitate healthy choices for 23andMe customers and contribute to ameliorating this ongoing public health concern,” the company writes in background information about the science behind the report.
Unlike the company’s genetic report products that required US Food and Drug (FDA) Administration authorization, the new report on type 2 diabetes was developed under 2016 FDA guidelines for low-risk general wellness devices. Other 23andMe products related to “genetic weight” and saturated fat were similarly developed, a 23andMe spokesperson told Medscape Medical News.
Three Main Considerations
Asked to comment, Mark McCarthy, MB, BChir, MD, Robert Turner Professor of Diabetic Medicine at the University of Oxford, UK, pointed out that the approximate 65% predictive accuracy of the genetic score is similar to that found in other research settings.
“We’re all finding more or less the same degree of power in these genetic tests to predict people with or without diabetes, so that’s good,” he told Medscape Medical News.
As for clinical utility, McCarthy said there are three main considerations.
First, “Does it tell you anything you don’t already know from clinical risk factors? That depends on the age of the person. At age 20, it probably does, especially if there is no family history. But by age 50, you probably will have quite a lot of evidence, so it probably won’t add a lot.”
Another important issue is the test’s accuracy in various ethnic groups, he said, noting that “most of the data are in Europeans, so the risk prediction is much better in [that population] and it’s dangerous to try and apply it to other ethnic groups because it could give misleading results. It’s a problem for the field. We’re working on it, but it’s not yet solved.”
Indeed, the 23andMe spokeswoman acknowledged this as an “industry-wide problem,” but also pointed out that their database includes one of the largest cohorts of African Americans — over 78,000, comprising 4% of the total population — for whom individual patient-level data are available.
23andMe is also participating in several projects along with Ancestry.com to collect more data from people with African ancestry, she said.
The third issue, McCarthy noted, is whether people will actually change their behavior based on the genetic information gleaned from the report.
“Will people with a high risk score be motivated to change their behavior in ways that will be advantageous for their health? Most studies that have tried to look at this have found not much benefit. But there appears to be no negative effect either…By and large, it makes precious little difference.”
A Potentially Useful Contribution
McCarthy said the true value of such genetic risk scores will be realized when a single test can give patients a comprehensive report about their overall risks for many different common conditions in addition to diabetes, such as cardiovascular disease, breast cancer, and prostate cancer, for which subsequent individualized mitigation measures can be taken by those found to be at high risk.
“I think it’s a really great experiment…We’re moving towards a time when many adults will have this kind of genetic data sitting in their medical records to be used for tailoring drugs, predicting side effects, and predicting many diseases in ways that are clinically useful,” he observed.
And, “We really need to have datasets where we can explore how to present this information, how to use it, and how to present it in ways to motivate people to do good things to reduce their risk of disease,” he added.
“If done right, this is a useful contribution to the debate. Have they oversold it? Sure. But is it still useful? Yes, I think it is.”
McCarthy has served as a director, officer, partner, employee, advisor, consultant, or trustee for Merck, Pfizer, Eli Lilly, and Novo Nordisk. He has received research grants from Pfizer, Merck, Takeda, Servier, Sanofi Aventis, Roche, AstraZeneca, Novo Nordisk, Eli Lilly, Janssen, and Boehringer Ingelheim. He has received income in an amount equal to or greater than $250 from Merck, Pfizer, and Lilly.