Wearable, Smartphone and mHealth Approaches to Academic Stress in University Students: A Narrative Review

Abstract
Academic stress is common among university students and is linked to poorer mental health and
academic outcomes. Traditional measurement relies on retrospective questionnaires such as the
Perceived Stress Scale (PSS), which miss day-to-day changes. This narrative review summarizes how
three technologies have been used to measure or reduce stress in students: wearable physiological
sensors, smartphone-based digital phenotyping, and mobile health (mHealth) interventions. It covers
nine key empirical studies and reviews, with every reported figure checked against the original
publication.
In controlled laboratory protocols, multimodal chest-worn sensors reached binary stress-detection
accuracy of about 93% (WESAD, N = 15). Performance was lower in daily life: the cStress model
predicted field self-reports with 72% accuracy, and a context-aware wrist model detected 70% of reallife stress events at 95% precision. Over a 30-day study of 201 students, wearable features classified
high versus low perceived stress with 78.3% accuracy, outperforming phone-usage and behavioral
features. Smartphone sensing showed that conversation, sleep, and activity fall as workload rises across
a term. For interventions, one randomized trial of a mindfulness app (N = 88 analyzed) reported a large
reduction in PSS scores versus a wait-list control, while a systematic review of 89 digital interventions
found that most were effective or partially effective but carried moderate-to-severe risk of bias.
The main gaps are the difference between laboratory and field performance, the inability of
physiological signals to separate distress from positive arousal, small and homogeneous samples, and
low long-term engagement. Future work should test closed-loop just-in-time adaptive interventions with
micro-randomized trials, personalize models to each student, and include students outside North
America and Europe