Investigating the presentation of uncertainty in an icon array: A randomized trial

Abstract

Clinicians are often advised to use pictographs to communicate risk, but whether they offer benefits when communicating risk imprecision (e.g., 65%-79%) is unknown. To test whether any of three approaches to visualizing imprecision would more effectively communicate breast and ovarian cancer risk for BRCA1 pathogenic variant carriers. 1,300 UK residents were presented with a genetic report with information about BRCA1-related risks, with random assignment to one of four formats: no visualization (text alone), or a pictograph using shaded icons, a gradient, or arrows marking range endpoints. We also tested pictographs in two layouts. Analysis of variance (ANOVA) and regression was employed. There was no effect of format. Participants shown pictographs vs. text alone had better uptake of breast cancer risk messages (p < .05, η 2 = 0.003). Pictographs facilitated memory for the specific amount of risk (p < 0.001, η 2 = 0.019), as did the tabular layout. Individuals not having completed upper secondary education may benefit most. We found weak evidence in favor of using simple pictographs with ranges to communicate BRCA risk (versus text alone), and of the tabular layout. Testing different ways of communicating imprecision within pictographs is a novel and promising line of research.