(Updated)– We may live in the age of information but we really have no idea how to use all that information, especially when it comes to health. One new study, published in the BMJ, is a good illustration of this principle.
“There are high expectations that advances in genetics will usher in a new era of personalized medicine, and that because communicating genetic risks will motivate risk-reducing behavior changes, such communication has a role in risk reduction strategies aimed at improving population health,” write the study authors. But the study demonstrates how difficult it might be to translate these great expectations into reality, as it suggests that by itself communicating the results of genetic tests “has little or no impact on risk-reducing health behavior.”
Gareth Hollands and colleagues updated an earlier 2010 Cochrane review on the impact of communicating genetic risk information to people. They identified 18 studies in which risk-reducing behavior was measured in people randomized to receive genetic assessment of disease risk. The review included six studies on smoking cessation, seven studies on diet, and six studies on physical activity. The authors reported finding no evidence to support the use of genetic testing to encourage risk-reducing behavior.
There is an important limitation to the review. The authors reported that the “studies were predominantly at high or unclear risk of bias, and evidence was typically of low quality.” They also observe that non-randomized studies have found that genetic testing in people with family histories of breast, ovarian and colorectal cancer has resulted in increased screening and prophylactic surgery in people found to be at high risk by genetic testing. These findings “suggest that DNA based risk assessments are more likely to motivate clinical means of reducing risk (such as undergoing surgery or attending screening) than behavioral means (such as altering smoking, diet, or physical activity behaviours).”
But for now there is a glaring absence of evidence supporting the use of genetic tests to improve health behavior. The authors concluded that “the available evidence does not provide support for the expectations raised by researchers and proponents of personalized medicine as well as direct-to-consumer testing companies that the receipt of results from DNA based tests for gene variants that confer increased risk of common complex diseases motivates behavior change.”
Saurabh Jha (University of Pennsylvania) sent the following comment in response to this study:
Debunking of Information Calvinism
The theory, a seductive one, is that knowing that your genetic risk of a disease is high will induce healthy behavior. What separates us from healthy behavior is more than a generic pamphlet about health, but salient information about ourselves. Ignorance is sloth. For example, smokers who know they have an above-average risk of lung cancer may be more inclined to quit. I call this “Information Calvinism.” According to a recent meta-analysis (MA) this may be wishful thinking.
After rigorously combing the literature for studies examining the impact of genetic information (knowing your risk) on behavior, researchers isolated 18 randomized (or nearly randomized) studies which compared the behavior of people who knew about their DNA-based risk with those who didn’t. The control group often had the same intervention as the experimental group, except knowing about the DNA-based risk. The outcome was change in behavior. Behaviors included smoking cessation, reduced alcohol intake, sun protection, better diet, more physical activity, or attendance at screening clinics.
There was no difference in outcomes between those who knew and those who were ignorant of their genetic risk, of either clinical or statistical significance. The Forest plot tells an interesting story. Of all the behaviors, the two which nearly had an effect size of statistical significance, meaning the 95 % confidence interval almost escaped the zero effect line, were sun protection and diet. The lack of effect on alcohol intake and physical activity does not surprise me. But I am surprised that knowing about genetic risk doesn’t increase attendance at screening programs.
For some entities the link between behavior and disease is tenuous. For example, if you know you’re at increased risk of Alzheimer’s disease what are you supposed to do? Solve the Times Crosswords? Take selfies?
Overall, I was surprised by the findings, even though I’m not a proponent of direct-to-consumer genetic testing. So I looked for weakness in methods. But shooting an MA is like shooting the messenger. MA is only as good as the studies it analyzes. One thing difficult for a study assessing the impact of information on behavior to control, is motivation. The motivated may be more likely to seek genetic information and to alter their behavior.
Randomization should distribute motivation equally between the experimental and control arms. There are two caveats – both bias against genetic information. The more motivated in the control arm may seek genetic information later, which could affect their behavior. The less motivated in the experimental arm may be even less inclined to alter their behavior on knowing that their risk is below average. The follow-up period in the studies, in the order of months and maximally 18 months, was too brief to mitigate the possibility of the former.
Despite real flaws in methods of studies pooled, it’s fair to say that widespread knowledge of DNA-risk won’t be a game changer for the healthcare system. Proponents of personalized medicine must choose another Messiah, or another God.
Being a libertarian I’m compelled to emphasize one point. Just because most people like Eggs Benedict doesn’t mean I must like it. I can’t stand Eggs Benedict. Just because most people don’t alter their behavior when they know about their genetic risk doesn’t mean that you won’t. But if you want to know about your ancestry, please pay for the test, and all the downstream tests, counselling and procedures it induces.