Passive digital monitoring via mobile apps and wearables in multiple sclerosis: A systematic review

Effective management of multiple sclerosis requires precise monitoring to detect disease progression, yet conventional episodic assessments often fail to capture subtle functional decline. This systematic review aimed to synthesize current evidence on passive digital phenotyping via smartphones and wearable sensors as a tool for continuous and ecological evaluation. Following PRISMA guidelines, fourteen studies meeting strict eligibility criteria were included. These investigations analyzed raw data collected unobtrusively from smartphone touchscreens, inertial sensors, and photoplethysmography. The signals were processed to extract digital features, including keystroke dynamics, gait metrics, physiological markers, and behavioral patterns, to test their associations with standard clinical scales. The results demonstrated robust, significant correlations between these objective digital biomarkers and established clinical scores across motor, cognitive, and symptomatic domains. Specifically, processed typing kinematics and gait parameters significantly aligned with upper limb dexterity and walking ability respectively, while specific keystroke latency metrics proved sensitive to cognitive processing speed. Furthermore, multimodal physiological and behavioral markers successfully tracked complex symptoms such as fatigue, depression, and sleep quality. Overall, the findings highlighted that the integration of these multimodal biomarkers into clinical practice offers a scalable, non-invasive solution for personalized management. This approach could facilitate the early detection of "silent" symptoms and optimize therapeutic strategies, effectively bridging the critical gap between sporadic clinical visits and real-world functioning in multiple sclerosis.