|Prof. David Spiegelhalter said effective risk communication requires a multipronged approach that uses words, numbers, images and varied metaphors.
But effective communication is no simple matter, noted Spiegelhalter, who spoke as part of NIH’s Medicine: Mind the Gap series.
For one, words used to describe risk can be quite unclear. “Might” and “possibly” mean different things to different people. Adopting shared definitions of terms is one solution to this problem. In the Intergovernmental Panel on Climate Change’s likelihood scale, for example, “very likely” means a 90-100 percent probability.
Numbers can be just as confusing, though. Spiegelhalter noted that when asked to identify the greatest risk of getting a disease—1 in 10, 1 in 100, or 1 in 1,000—only a quarter of U.S. study participants picked the right answer. “It’s very confusing because the biggest number is associated with the smallest risk,” he noted.
Spiegelhalter also cautions against “denominator neglect,” or the failure to consider a number in context. “Who does that?” he asked. “The newspapers do it every single day.” A negative event makes a dramatic story, he said, but the number of times that nothing negative happens—the denominator—often goes unnoticed.
Absolute risk and relative risk pose problems, too. A study can find that patients treated with a medication had a 36 percent reduction of heart attack risk compared to those taking a placebo. So, in terms of relative risk the effect sounds impressive. However, in terms of absolute risk, or the likely effect on the population, the number may be much less impressive, such as 1 out of 100 people.
The public should be particularly suspicious of mismatched framing, Spiegelhalter added. A misleading drug ad might report benefits in relative terms to play up the positives, but then describe harms in absolute terms to minimize the negatives. One way to address this problem is to report both benefits and harms uniformly on a drug fact label, such as one that an FDA advisory committee has recommended.
Above all, Spiegelhalter argued, effective risk communication requires a multipronged approach that uses words, numbers, images and varied metaphors.
For example, pictographs help convey a comparison. An image might use rows of stick figures to represent treated study participants on one side and those assigned a placebo on the other. Each possible benefit and side effect could be depicted in a separate line. Other approaches are useful too, including tree-maps, which divide an image into different-sized rectangles to convey the relative significance of compared items.
Spiegelhalter also recommends using measures that have concrete meaning in a person’s life. A man who hears that his life expectancy may be reduced by 1 year may not be impressed. But you may grab his attention if you translate that into losing a half-hour every day of his adult life. Spiegelhalter calls this half-hour a “microlife” and can enumerate the microlife cost of behaviors like drinking a few extra beers.
Dr. Paul Coates, acting director of the NIH Office of Disease Prevention and director of the Office of Dietary Supplements, poses a question during the seminar.
Sometimes, though, scientists do not have a solid understanding of the risks involved in a particular behavior. In that case, Spiegelhalter advocates finding ways to communicate uncertainty, such as an image that captures multiple possible outcomes or a confidence scale that conveys experts’ level of agreement.
After all, he said, some uncertainty is part of the process. The challenge is “to have faith about the rightness of one’s actions,” said Spiegelhalter, quoting NIH’s Dr. Paul Han, “while affirming the irreducibility of doubt.”
To watch a videocast of this lecture, visit http://videocast.nih.gov.