Communicating risks with interactive visualisations and reflective tasks a mixed-methods evaluation of a mammography screening decision aid

Abstract

Static graphs of statistics are established visual aids in risk communication and decision support. Interactive information visualisations (InfoVis) and reflective tasks are supposed to enhance active processing, but the evidence is scarce and mixed. This mixed-methods research investigated the effectiveness and user experience of InfoVis and tasks in the context of mammography screening. In a web-based experiment prospective invitees of the screening program (N = 338; aged 30-49) tried a pre-tested web-based decision-aid with risk information either as text, static graph, or InfoVis with or without reflective tasks. The main outcomes were informed choice and risk knowledge, the latter operationalised according to the fuzzy-trace-theory. The accompanying qualitative evaluation with seven participants applied think-aloud protocols and focused interviews. There was no experimental evidence that InfoVis support risk knowledge or informed choice better than text or static graphs. There were even minor detrimental effects. The qualitative results showed problems with the InfoVis presenting risk of overdiagnosis, and negative reactions towards the-tasks. InfoVis processing was easy when the underlying concept was easy. While reflective tasks seem not advisable in this target group, limited and well-considered application of InfoVis with a low cognitive load can be an alternative, attention-directing visual aid format.