Making sense of PROM outcomes: a mixed method study to optimize graphical visualization formats for children

Graphs are often used to increase patient understanding of Patient Reported Outcome Measure (PROM) scores. We aimed to investigate which graph visualization options are best interpreted by children. We conducted a quantitative study to assess children’s interpretation accuracy and perceived clarity of different PROM visualizations graphs through a test, and a qualitative study using ‘think aloud’ interviews about the same visualizations to explore how different visual elements were used for interpretation. Participants included (1) children from the Dutch general population (quantitative) and (2) children (8-18y) treated at Emma Children’s Hospital (quantitative and qualitative).We created sample graphs (e.g. bar, line) based on literature, varied graphical elements (e.g. addition of numerical information yes/no), and presented them in randomized order. Logistic and linear mixed models analyzed interpretation accuracy and clarity ratings. We analyzed interviews with a deductive approach. We included 558 children in the quantitative study (1: 546, 2: 12), and 14 in the interviews. Foundational elements like clear labeling were found beneficial. ‘Heatmap’ and ‘color’ as indicators of concerning scores had higher interpretation accuracy and clarity, and SD-lines lower. Bar graphs had slightly higher accuracy. Radar graphs scored significantly worse on interpretation accuracy and clarity compared to all other graphs. Interviews revealed a preference for single domain graphs with color-coding indicating concerning scores. We recommend using bar graphs incorporating heatmaps or color as concerning score indicators in clinical encounters with children. Application of our results will likely facilitate patient engagement in consultations and potentially improve patient-centered care.