The rise of health-related challenges necessitates personalisation in digital behaviour change interventions beyond traditional self-reported assessments. This systematic review analysed 75 studies using the PRISMA methodology to explore how digital biomarkers collected from consumer wearables and smartphones can be leveraged to infer psychosocial factors such as stress, anxiety, and mood, among others. Several key combinations of digital biomarkers, including heart rate variability, electrodermal activity, skin temperature, motion data, phone usage, and GPS-based patterns, were identified for inferring 11 psychosocial factors. The review also examined analytical techniques used for inferring these psychosocial factors, revealing a clear shift from traditional regression models to advanced machine-learning approaches for more accurate and scalable predictions. The findings show that real-time data streams collected from everyday consumer devices hold great untapped potential for continuous and objective monitoring of behavioural health and overall well-being. While new wearable technologies continue to emerge, an equally pressing challenge lies in unlocking the full potential of data already generated daily by smartwatches, smartphones, and fitness trackers. Harnessing this information, from pockets and wrists to real-time inference of psychosocial factors, paves the way for scalable, personalised digital behaviour change solutions, improving early detection, timely support, and preventive care across diverse populations.
Shah et al. (2026) studied this question.