Diogo Nogueira, Celia Azevedo Soares
The selection of data visualization formats for cancer risk communication is a routine decision in clinical cancer education, yet it currently lacks an empirical basis. This pilot study examined format preference and comprehension across educational levels in a medical genetics clinic, and whether stated preference corresponds to the format consulted under cognitive demand. In this cross-sectional pilot study at a medical genetics clinic in Porto, Portugal (January-June 2024), 63 adults stratified by educational attainment (low, n = 25; medium, n = 16; high, n = 22) evaluated five visualization formats (bar graph, table, pictograph, pie chart, clock graph) presenting identical risk scenarios. Preference and comprehension were assessed via semi-structured interview using non-parametric tests. Bar graphs (30.2%) and Table (28.6%) were most preferred; comprehension was highest for bar graphs (87.5%) and Table (87.0%). Clock graphs were least preferred (7.9%) and least understood (25.0% for Question 1; OR = 0.064, 95% CI 0.011-0.617, p = 0.023). Comprehension differed by educational level (54.0% low, 78.1% medium, 97.7% high; p < 0.001, Cramer's V = 0.44). Preference and consulted format diverged: tables were used by 36.5-39.7% of participants despite being preferred by only 28.6%. Format selection in cancer risk education should not rely on patient preference alone. Bar graph and table use were associated with higher comprehension; clock graph use was associated with lower accuracy. For patients with lower educational attainment, no format achieved adequate comprehension. This supports multimodal education strategies. These associations are observational and cannot be causally attributed to format itself, as format was self-selected.