Correlates of weekday sleep duration before and during the COVID-19 lockdown: Evidence from two Austrian national surveys from 2017 and 2020.

The COVID-19 pandemic profoundly disrupted daily life, including sleep behaviour. While several international studies report longer sleep durations during lockdown, most relied on convenience samples and only few influencing factors were considered. The aim of this study was to identify correlates of weekday sleep duration before and during Austria's first lockdown using two population-based cross-sectional surveys. Data were drawn from online surveys in September 2017 (n = 784) and May 2020 (n = 847), weighted to match the Austrian adult population by age and sex. Weekday sleep duration was derived from self-reported bed and wake times. Weighted linear regression models were fitted using backward elimination to select among known and suspected correlates of sleep duration. In 2020, additional COVID-specific covariables were considered. Covariable selection stability was assessed via bootstrapping. Compared to 2017, sleep duration increased in 2020 by 40 min (95% CI 27-55). Several covariables (minimum required sleep, work schedule, dinner time, chronotype, and diagnosed sleep disorder, education level and smoking status) were consistently retained in the selection models in both survey years. In 2020, COVID-specific factors such as changes in sleep quality, emotional burden during lockdown, and attitude toward the future further explained variation in sleep duration. Inclusion of these variables increased the adjusted model fit (R2) from 0.26 to 0.31. Beyond established determinants, COVID-related factors contributed substantially to explain variation in sleep duration during the lockdown. Our findings underscore the importance of incorporating context-specific covariables in future surveys to capture behavioural adaptations, e.g. during societal crises.