Purpose: Informed patient decision-making is the central goal of patient decision support interventions, but research is needed on optimal presentation of risk and outcome data. We examined how men 50 years or older (N=188) interpret data from graphical presentations in a state-of-the-art videotape decision aid about benign prostatic hyperplasia (BPH) treatment.

Method: Quasi-experimental descriptive design. Pre/post knowledge and decision survey reported previously. Audio-taped “think-aloud” semi-structured interview during and following viewing of videotape. Following viewing, men reviewed paper copies of three graphs sampled from the video (bar, line, and stick figure graphs) and asked, “What does this graph mean to you, if anything?” and “What did you like about the graphs?”.

Results: There were differences in the accuracy of answers for the three types of graphs (%correct, (N)): bar graph 33.9 (177), line graph 50.3 (169), stick figure 75.8 (165). College educated men were more accurate for each graph; largest difference in bar graph. Black men were more accurate than white men for the bar graph. Stated liking (or not) for graphs was unassociated with accuracy.

Conclusion: Accuracy varied with seeming complexity of graph: bar > line > stick figure. Stick figure graph was most frequently understood, perhaps due to single concept presented. Education was positively associated with accuracy of interpretation. Race has less impact but Black subjects did better on the most complicated (bar) graph. This observed difference is not due to education. Presenting risk comparisons is essential to informed decision-making but appears to be compromised by available graph formats