Does genomic risk information motivate people to change their behavior?

For families with rare, highly penetrant genetic conditions, genome medicine is already a reality, with genetic tests that can identify the family members at high risk of disease. The rationale for testing is clear: it saves lives in families with conditions such as multiple endocrine neoplasia type 2, hereditary breast ovarian cancer syndrome and familial hypercholesterolemia by directing the use of prophylactic surgery, intensive screening strategies and specific treatment regimens. With these successes in mind, many people reasonably hope that similar benefits can be achieved on a population level by screening for more common genetic variants associated with disease risk.The rapidly expanding number of known risk variants following from the dramatic success of genome-wide association studies [1] has fueled this vision of 'personalized medicine'. The logic is that identification of even modestly increased risks for common diseases enables providers to make personalized recommendations for screening and risk reduction. This assumes that genetic risk information will motivate behavior, because the greatest gains in prevention for common complex diseases will come from lifestyle improvements, such as smoking cessation for heart disease and lung cancer risk reduction, diet and exercise changes for diabetes and cancer risk reduction, and adherence to recommended screening guidelines.But this vision of individualized genome medicine must be approached with caution. Most variants emerging from gene-disease association studies have very small effect sizes, often with odds ratios of 1.5 or less [2,3], confirming that susceptibility to common chronic disease reflects a complex interaction between many different genes and the environment. Despite an increasing number of direct-to-consumer tests offering information about common disease susceptibilities, risk information of this sort may be most useful as an adjunct to current risk assessment, refining rather than replacing o