Feasibility of frequent self‐administered mobile cognitive assessments in adults with and without cognitive impairment

Abstract Background Traditional paper and pencil approaches measure cognitive performance at one time point. Early indications of cognitive impairment often appear irregularly, suggesting diagnoses derived from a single session can fail to characterize normal fluctuations and lead to inaccurate conclusions. Mobile applications enable brief cognitive assessments to be completed frequently and remotely. However, the feasibility of older adults adhering to a schedule of repeated self‐administered cognitive assessments is unclear. We tested the repeated use of a mobile assessment among older adults as a potential method for collecting cognitive performance data to detect early signs of cognitive impairment. Method The current study included 33 participants (n = 23 cognitively unimpaired, n = 10 cognitively impaired) recruited from the Boston University Alzheimer’s Disease Center. A remote practice session using the Digital Automated Neurobehavioral Assessment (DANA) mobile application was completed prior to beginning an assessment schedule consisting of three assessments within the first week (segment 1) and one assessment per week during the following three weeks (segment 2). Feasibility was measured via schedule adherence across segments, stratified by age and cognitive status. Cognitive efficiency, a derived measure of speed and accuracy, was compared for the Procedural Reaction Time (PRT) task at the first assessment and longitudinally across segments 1 and 2. Demographics including age, sex, race, education, and mobile device type were also collected (Table 1). Result Across all segments, 90.9% of participants adhered to the assessment schedule with a ±1 day grace period applied. Adherence was similar between each segment, and by cognitive status and age group (Table 2). No significant differences in cognitive efficiency were observed on the PRT task between cognitive groups (Figures 1‐2, Table 3). Conclusion Our findings suggest that older adults, including those with cognitive impairment, are capable of completing self‐administered remote smartphone assessments at a high‐frequency. Traditional analytic methods failed to differentiate cognitive groups based on mean cognitive performance; however, between‐group comparisons could mask within‐person fluctuations characteristic of cognitive impairment. Therefore, further study with a longer longitudinal assessment period and larger sample size will be required to determine whether more advanced analyses can distinguish impairment status based on cognitive performance patterns.