Does genetic risk information improve healthy behavior? Let’s not throw out the baby with the bath water!

Posted on by Muin J Khoury, Director, Office of Public Health Genomics, Centers for Disease Control and Prevention

a double helix in a green circle, a glass of alcohol in a red circle wth a slash through it, a No Smoking stop sign, excercise equipement, an apple and a Screening door tagIn a recent systematic review with meta-analysis, Hollands et al evaluated the impact of communicating genetic risk information on risk-reducing health behaviors and motivations for behavior change. The authors reviewed 18 studies with 7 behavioral outcomes, including smoking cessation, diet and physical activity. They found no significant effects of communicating DNA based risk estimates on smoking cessation, diet or physical activity. They also found no effects on other behaviors (including alcohol use, medication use, sun protection, and screening programs).  The study generated interest and press coverage.   One widely expressed sentiment is to point out “the limits of personalized medicine” emphasizing that “knowing genetic risk of disease doesn’t motivate people to change their behavior.”

A closer look at the review gives plenty of reasons to pause and reflect on the limitations of the studies themselves. The review might at first seem to prematurely condemn a nascent field that is still in infancy and requires additional rigorous research. However, the authors of the review pointed out the limitations of available evidence, and stated that “studies were predominantly at high or unclear risk of bias, and evidence was typically of low quality.”

In a series of letters to the editor, further explanations of study limitations were offered. They are worth repeating here. Janssens explains that communicating risk estimates based on one or a few single nucleotide polymorphisms with weak disease risk associations is hardly expected to change health behavior for complex diseases that are caused by interactions between many genetic and environmental factors. Most genetic tests included in the review are already outdated. Currently, more genetic variants for different diseases are available which can improve the predictive ability of genetic tests.

Burton et al point out that we already know that there is a paucity of evidence linking risk communication, genomic or otherwise, to the motivation of sustained behavior change. Motivating behavior change is extraordinarily difficult, and there is no reason to suggest that genomic information will have a greater influencing power than non-genetic information such as personal, physiological or biomarker data – even though these more traditional factors are widely considered to have utility. Furthermore, there is little evidence that the accuracy of the risk prediction for these diseases will be improved. Thus, for any individual, the disease risk estimate will be only slightly higher or slightly lower than the population “average” risk, making the case of stratified or individualized disease prevention less than compelling.

Hay et al point out the limitations of studies included in this review such as selection biases, relatively small sample sizes, the methods of risk communication, and the failure to consider gene-environment interactions.

The bottom line is that a rigorous review of methodologically-limited studies can be misleading. The field of genomic risk communication is at an early stage, and this is a time of rapid change in risk information. So let us not throw out the baby with the bathwater.  Rather than dismissing genomics on the basis of limited data, we should focus on generating appropriate evidence on how the combination of genomic and other personal and environmental information can be effectively used to drive behavioral and medical interventions that improve health and prevent disease.

Posted on by Muin J Khoury, Director, Office of Public Health Genomics, Centers for Disease Control and PreventionTags ,
Page last reviewed: April 8, 2024
Page last updated: April 8, 2024