Characterizing Sleep Patterns And The Validity Of Wearable Sleep Metrics In First-year College Students

Wearable technology allows for an unbiased look at physical activity and sleep, which are useful in determining performance outcomes. While previous studies have validated wearable sleep data in experimental settings, few studies have examined the validity of these devices in an observational cohort. PURPOSE: Characterize the sleep habits of a first-year college student cohort and correlate wearable device data to a validated sleep instrument, the Pittsburgh Sleep Quality Index Short (PSQI-S) survey. METHODS: 121 first-year college students were recruited as a cohort within a larger study using wearable and digital tools to examine the correlation of sleep and physical activity to academic performance. Participants were asked to wear a wearable device throughout a semester and to complete the PSQI-S upon study completion, reflecting estimations of sleep duration (Q4) and sleep disturbance (Q5b) over the past month. From the month reflecting PSQI-S data, wearable device-derived total sleep minutes and minutes awake during sleep were compared to Q4 and 5b respectively via linear regression. RESULTS: 75 subjects met a 75% compliance requirement and completed the survey. There was a moderate relation between total minutes asleep reported by wearable device and self-reported sleep hours via PSQI-S (R2 = 0.4103). Regression indicated a tendency to over-report sleep duration via PSQI-S compared to objective wearable data (regression slope 0.41 + 0.12, p < 0.0001, Fig A). There was no association between the sleep disturbance component of the PSQI-S and wearable device for minutes awake (Fig B). CONCLUSIONS: Wearable data was moderately accurate at depicting total sleep time when correlated with validated predictions via PSQI-S. Understanding differences between self-report surveys and wearable-derived real time data are important to better predict health outcomes and performance in context with everyday physical activity habits. University of Michigan Bioscience Initiative