NEW YORK (GenomeWeb) – Results of a new meta-analysis suggest that across published research studies, subjects who receive genetic disease risk testing appear largely unmotivated to change their behaviors or health habits based on the results.
In the study, which appeared last night in the British Medical Journal, investigators from the University of Cambridge's behavior and health research unit culled the published literature to identify studies which used a controlled design to evaluate how testing for genetic variants that affect an individual's risk of developing common complex diseases might influence their behavior or lifestyle.
Overall, the researchers examined about 10,000 abstracts, narrowing these down to 18 studies that matched their criteria for inclusion. When they compiled the data from these, they found that across the board, individuals informed of their genetic disease risk showed little or no change in their health-related behavior, such as smoking cessation and physical activity, compared to control groups.
The emergence of genomic tests that report variants linked to increased risk of common diseases has spurred much excitement and much debate, with proponents arguing that knowledge of being at a greater risk of a particular disease can spur individuals to make more informed decisions about changing their behavior or seeking specific health interventions.
However, the authors of the new study argue, DNA-based disease risk testing can only translate into health benefits if patients actually act on their results, and if that action in turn modifies their outcomes.
Theresa Marteau, leader of the new study, said in a statement that while the personalized medicine field may have high expectations for the impact of genetic risk testing on patient decision-making, based on the Cambridge team's meta-analysis, this impact has not appeared to pan out.
"We have found no evidence that this is the case," she said.
Though other review studies have addressed the question of emotional and behavioral outcomes of DNA-based disease risk testing, none have quantified the results of randomized controlled studies across the variety of different behavioral effects that the Cambridge team's new analysis did.
To identify appropriate studies to analyze, Marteau and colleagues searched a variety of databases of published studies up to February 2015. The group excluded studies that were not randomized or controlled using an alternative to true randomization, such as alternation of subjects or division by date of birth.
They also excluded studies that communicated risk for diseases like Huntington's, which have no known risk interventions, focusing only on studies addressing diseases whose risk can be mediated, like heart disease, cancer, and Alzheimer's.
This resulted in the isolation of 18 studies in which researchers had compared the effect of genomic disease risk testing with no testing, or with some non-genetic risk estimate, on behaviors including smoking, alcohol consumption, diet, and physical activity, and where possible, associated outcomes like depression, anxiety, and motivation.
If a study reported different outcomes related to the same behavioral change, the group analyzed only the most stringent, for example biochemically validated smoking cessation versus patient-reported smoking cessation. When a study had more than one follow-up time point, the researchers included only the longest available.
The Cambridge team then analyzed the studies in aggregate with an eye to each of several specific behavioral changes. Across most of these, they saw that there was little to no effect on behavior associated with communication of genetic risk results.
For smoking cessation, for example, analysis of six studies that included this endpoint found that among 2,663 total study subjects, there was no significant effect of DNA-based risk communication on smoking cessation.
Interestingly, this didn't change when the group looked at five of these studies in which they could compare participants whose genetic results showed them to be at a higher risk, versus those whose results were negative. High-risk or low-risk genetic information did not appear to affect patient's smoking behavior.
The group also saw little to no effect of genetic testing across three studies assessing self-reported alcohol use, and in six studies assessing physical activity as an endpoint, the authors reported.
Single studies that looked at sun protection behaviors in relation to genetic cancer risk and medication use in relation to genetic Alzheimer's risk also showed little effect amongst those receiving DNA risk testing relative to the relevant control.
According to Marteau and colleagues, the results, while not exhaustive, are a strong indication that there is little support for the idea that that genetic risk testing focused on commonly occurring genetic variations will significantly impact behavior and lifestyle changes.
At the same time, the evidence equally supports a lack of negative impact from genetic risk testing on behavior and motivation. In other words, the team's analysis also found no evidence that testing might demotivate people or discourage them from changing their behavior or lifestyle.
This tracks with results from other studies, such as PGen — a prospective analysis of the psychological and behavioral impact of personal genomics data based on surveys of customers who received results from either 23andMe or Pathway Genomics — which found that changes in mood and anxiety in early adopters of direct-to-consumer genetic risk testing were transient, showing an initial shift at the time results were received, but then settling over six months back to normal levels.
In contrast to the findings of the Cambridge team's meta-analysis, there are areas where it appears that genetic risk information does have significant measurable effects on patient behavior, namely amongst those who receive testing for hereditary cancer risk where research has seen increased rates of screening and prophylactic surgery.
Authors of the new study suggested that this discordance may indicate that individuals who receive genetic tests are more likely to be motivated toward clinical interventions like surgery or doctor's visits than they are to self-directed behavioral modifications like smoking cessation and exercise.
Similar patterns have emerged in other studies. For example, data from PGen found approximately 11 percent of study subjects that received direct-to-consumer genetic testing had sought out some sort of medical follow up six months later based on their genomic results.
However, the Cambridge study authors wrote, increased health-care interventions amongst those who get genetic cancer risk testing is not universal. The one large well-conducted study in their analysis that looked at colorectal cancer screening found that there was no increase in uptake of screening associated with genetic testing.
The team wrote that its analysis is not the final answer to the question of the impact of genetic risk analysis. Confining their analysis strictly to randomized controlled studies, the group's conclusions contrast strikingly with results from smaller non-randomized observational studies, like PGen, in which 42 percent of more than 1,000 surveyed individuals who had received DTC genetic testing reported changes in terms of diet and exercise, supplement, and over-the-counter drug usage.
Because of the meta-analysis structure of the study it is possible that the data misses particular niches where there may actually be statistically significant changes in behavior among certain subsets of individuals with certain types of genetic results.
As these questions are studied further, the Cambridge authors recommended that future trials be conducted using methodologically robust designs powered to detect possible small effects on behavior.