Background There is an unmet need for reliable and sensitive measures for better monitoring people with multiple sclerosis (PwMS) to detect disease progression early and adapt therapeutic measures accordingly. Objective To assess reliability of extracted features and meaningfulness of 11 tests applied through a smartphone application (“dreaMS”). Methods PwMS (age 18–70 and EDSS ≤ 6.5) and matched healthy volunteers (HV) were asked to perform tests installed on their smartphone once or twice weekly for 5 weeks. Primary outcomes were test–retest reliability of test features (target: intraclass correlation [ICC] ≥ 0.6 or median coefficient of variation [mCV] < 0.2) and reported meaningfulness of the tests by PwMS. Meaningfulness was self-assessed for each test on a 5-point Likert scale (target: mean score of > 3) and by a structured interview. ClinicalTrials.gov Identifier: NCT04413032. Results We included 31 PwMS (21 [68%] female, mean age 43.4 ± 12.0 years, median EDSS 3.0 [range 1.0–6.0]) and 31 age- and sex-matched healthy volunteers. Out of 133 features extracted from 11 tests, 89 met the preset reliability criteria. All 11 tests were perceived as highly meaningful to PwMS. Conclusion The dreaMS app reliably assessed features reflecting key functional domains meaningful to PwMS. More studies with longer follow-up are needed to prove validity of these measures as digital biomarkers in PwMS.

Background Cognitive impairment occurs in up to 70% of people with MS (pwMS) and has a large impact on quality of life and working capacity. As part of the development of a smartphone-app (dreaMS) for monitoring MS disease activity and progression, we assessed the feasibility and acceptance of using cognitive games as assessment tools for cognitive domains. Methods We integrated ten cognitive games in the dreaMS app. Participants were asked to play these games twice a week for 5 weeks. All subjects underwent a battery of established neuropsychological tests. User feedback on acceptance was obtained via a five-point Likert-scale questionnaire. We correlated game performance measures with predetermined reference tests (Spearman’s rho) and analyzed differences between pwMS and Healthy Controls (rank biserial correlation). Results We included 31 pwMS (mean age 43.4 ± 12.0 years; 68% females; median Expanded Disability Status Scale score 3.0, range 1.0–6.0) and 31 age- and sex-matched HC. All but one game showed moderate–strong correlations with their reference tests, (|rs|= 0.34–0.77). Performance improved in both groups over the 5 weeks. Average ratings for overall impression and meaningfulness were 4.6 (range 4.2–4.9) and 4.7 (range 4.5–4.8), respectively. Conclusion Moderate–strong correlations with reference tests suggest that adaptive cognitive games may be used as measures of cognitive domains. The practice effects observed suggest that game-derived measures may capture change over time. All games were perceived as enjoyable and meaningful, features crucial for long-term adherence. Our results encourage further validation of adaptive cognitive games as monitoring tools for cognition in larger studies of longer duration. Study Register ClinicalTrials.gov: NCT04413032.

No abstract