Using Wearables to Unobtrusively Identify Periods of Stress in a Real University Environment

One of the often overlooked features of wearable physiological sensing devices is the possibility for institutions to observe and manage the health and well-being of a group of people. For example, a school could use wearables to monitor for increased levels of stress or sickness within the student body. The accuracy and utility of current sensing devices make this possible. However, in real-world settings, many users do not use devices consistently or correctly. Moreover, institutional monitoring of individuals raises privacy concerns. In this study, we use real-world data to identify periods of cyclical stress in a cohort of 103 Japanese university students while preserving each student's privacy. We observe heightened waking heart rate (HR) and maximum waking HR, alongside corroborating evidence from sleep HR, sleep heart rate variability (HRV), activity patterns, and sleep phases, during periods coinciding with significant academic and societal events, suggesting clear associations with stress. We show the feasibility of detecting collective changes in stress biomarkers within a cohort using consumer wearables.