Background Daylight saving time (DST) clock changes have been associated with sleep loss for around a week after the transitions. This is thought to contribute to an increase in the number of heart attacks, strokes, depressive episodes and fatal traffic accidents after the clock changes. However, most studies of DST clock changes on sleep have relied on subjective self-report data, which may be affected by recall bias. Furthermore, estimates from more objective, accelerometer-measured sleep studies have lacked precision due to small sample sizes. We aimed to explore the effects of DST clock changes on sleep duration in a large accelerometer dataset.

Methods UK Biobank participants (n= 3,130; aged 43-77) wore accelerometers for 6 days over Spring and Autumn DST transitions between October 2013 and 2015. Paired t-tests compared people’s sleep on days before and after the clock changes, and on the Saturday and Sunday of the clock change. Linear regression analyses compared weekdays before and after the transitions. Analyses were stratified by sociodemographic characteristics. Results were compared to the normal weekend immediately prior to the clock changes.

Results In Spring, mean daily sleep duration was 20 minutes shorter (95%CI -25 to -15) on days after the clock change than before (normal weekend: 10 minutes longer). On the Sunday of the clock change sleep duration was 54 minutes shorter (95%CI -60 to -47) than on the Saturday (normal weekend: 20 minutes longer). Sleep duration was 18 minutes longer on the Wednesday after the clock change than the Wednesday before the change (95%CI 4 to 31) (normal weekend: no weekday differences). In Autumn, mean daily sleep duration was 37 minutes longer (95%CI 33 to 40) on days after the clock change than before (normal weekend: 15 minutes longer). Sleep duration on the Sunday of the clock change was 60 minutes longer (95%CI 56 to 65) than on the Saturday (normal weekend: 25 minutes longer). There were no detectable differences in sleep duration between the weekdays before and after the Autumn transition.

Conclusion Using a large accelerometer dataset we found that the Spring DST transition is associated with a reduction in sleep duration, whilst the Autumn clock change is associated with an increase in sleep duration. In contrast to previous research, the effect of both transitions on sleep duration was short-lived. Future research should explore ways to minimise the acute sleep loss experienced over the Spring clock change to reduce adverse health effects.