Feasibility, reliability, and validity of ambulatory smartphone-administered cognitive tests in multiple sclerosis

Background: Detection of subtle changes in cognition in multiple sclerosis (MS) may be enabled by ambulatory smartphone-based cognitive tests. Objective: To examine the feasibility, reliability, and validity of ambulatory cognitive tests in people with MS. Method: Adults with MS (N = 255) completed the NIH Toolbox Cognitive Battery, Symbol Digit Modalities Test, Paced Auditory Serial Addition Test-3, and 14 days of 4X/day ambulatory cognitive tests of working memory (dot memory) and processing speed (symbol search). Feasibility (response rates), measurement reliability (overall and by number of days of tests), and construct validity (convergent validity and known-groups analyses) were calculated. Results: Participants completed 80.1% of available testing sessions. Reliability was 0.98 for dot memory and 0.99 for symbol search. Reliability attained ⩾0.90 within 3 days for dot memory and 1 day for symbol search. Symbol search scores correlated significantly with neuropsychological tests of processing speed (rs range = |0.53–0.69|) and differed significantly by disability level (t = −3.95, p < 0.001) and MS subtype (t = −2.81, p = 0.005). Dot memory scores correlated with neuropsychological tests of working memory (rs range = |0.45–0.58|); dot memory scores did not differ by disability level or MS subtype. Conclusion: Smartphone-administered tests of working memory and processing speed demonstrate good feasibility, reliability, and construct validity in MS.