Brian Zikmund-Fisher, PhD, began his presentation with a barrage of numbers with which patients have to deal, including risk percentages, treatment success rates, blood test results, medication dosing charts, and nutrition labels. As an example of an objective numeracy measure, he asked the audience which represents the higher risk: 1-100, 1-1000, or 1-10? (Answer: 1-10.) Subjective measures of numeracy include the questions, “How good are you at working with fractions?” and “How often do you find numerical information useful?” He cautioned that, as with other kinds of literacies, numeracy does not equal education level.
Problems with numeracy — the ability to understand, transform, and derive meaning from quantitative health information — can often come from the way the information is represented. Zikmund-Fisher pointed out that the objective question he asked the audience exemplified this problem. Because people assume that an increase in the denominator means an increase in risk, people may have been confused or doubted the right answer, even if they got it.
Zikmund-Fisher pointed out the dual nature of risk: there is the risk of occurrence, and then there is non-occurrence. Risk calculators tend to focus attention on the numerator (occurrence risk), and patients must do mental math to see the equivalent risk of non-occurrence. It’s also important to ask what we want patients to know about risk. Or, we can ask, what do our patients need to do? Zikmund-Fisher introduced the “harm anchor” as a practical concept: it is the point at which something needs to happen. Harm anchors may sometimes need to be the focal point of numerical information.
In order to overcome numeracy barriers, we need to provide context, know that less is more, and match the format to the need. We do not want to use “might want” or “need later” to figure out what to include in our discussions or materials. The ethical obligation for transparency cannot override the requirement to be actually useful in a given moment for a given individual. We must recognize why we are providing data before we provide numbers.