Digital Executive Function App Performance among Post-Stroke Individuals: When Accessibility Makes a Difference

Background Executive functions (EF) are essential cognitive processes that support goal-directed behavior, including planning, organization, and problem-solving. Stroke survivors often experience EF deficits, affecting independent daily functioning. Digital EF tests offer innovative assessment tools, but their accessibility may influence performance. This study examined differences in performance between accessible and less-accessible versions of three EF digital tests among post-stroke individuals and healthy controls. The tests included two functional apps: the Four-Item Tablet Test (4ITT) for virtual shopping, shortened version of the Medicine Arrangement Test for Occupational Therapy (MATOT) for medication sorting, and a digital Trail Making Test (TMT). Methods A cross-sectional, crossover design was used with 32 post-stroke individuals and 32 age-matched healthy controls. Participants completed all three digital EF tests in both accessible and less-accessible versions. Accessible versions included one of the following features: voiceover, high contrast, or multi-page layout, while the less-accessible versions lacked these features. Performance was measured via task completion time, accuracy, and perceived ease of use. Non-parametric tests with Bonferroni corrections were applied to control for multiple comparisons, with adjusted significance thresholds (e.g., p < 0.006 for the 4ITT, p < 0.012 for the MATOT). Results The post-stroke participants performed better on the less accessible single-page version of the 4ITT (e.g., faster completion, p = .005 and higher efficiency, p = .006) and the less accessible version of TMT (time to complete p = .036). They performed better on the accessible voiceover version of MATOT (improved accuracy, p = .022 and time to first action, p = .001). The control group had no significant performance differences between versions but reported greater difficulty with the less accessible low-contrast version of TMT (p = .005). Conclusions A single-page layout enhanced performance in a virtual shopping task, while voiceover improved medication-sorting among post-stroke individuals. We conclude that specific accessibility features can support or hinder performance post stroke depending on task demands. Notably, performance differences were found only in the post-stroke group, suggesting the differences are stroke-related rather than age-related. This underscores the need to design digital EF assessments that specifically address stroke-related cognitive challenges.