LAUSR

© 2018 Wolters Kluwer Health, Inc. All rights reserved. www.epidem.com | e45 Web Site and R Package for Computing E-values effects up to 30 days after DST changes with a natural cubic spline with 3 internal knots, placed at equally spaced values in the log scale. Finally, we derived the fraction of deaths attributable to both, summer and fall DST changes, from the previous model estimates.5 The daily number of deaths caused by road traffic accidents averaged 3.5 (standard deviation = 2.8), ranging from 0 to 33. On the day of a DST change, the average number of deaths raised up to 4.3 (standard deviation = 3.3). Risk of fatal road traffic accident increased by 30% on the same day as summer DST change (risk ratio [RR] = 1.30; 95% CI = [1.04; 1.62] at lag 0), to decrease in the consecutive 1–3 days (Figure). Fall DST change showed a similar distributed lag pattern, but with a smaller risk increase on the same day (RR = 1.16; 95% CI = [0.93; 1.45]) and almost no harvesting effect thereafter (Figure). Over the study period, almost 40 deaths can be attributable to DST changes, 23.4 (95% CI = [2.7; 39.9]) and 14.7 (95% CI = [˗7.8; 32.8]) for the summer and fall changes, respectively. Since the DST change always occurs on a Saturday night, the observed short-term increase in fatal road traffic accidents could be explained by the combination of sleep disruption2,3,6 and regular weekend effects, where traffic injuries might be more often related to drug or alcohol consumption.7 The summer change implies 1 hour less of sleep, potentially resulting in augmented fatigue and impaired driving ability on the day of the change. As for the fall DST change, no hours of sleep are lost, but the loss of 1 light hour in the evening, when collision risk is highest,8 can explain the observed lower excess risk. Our findings suggest that, on average, DST changes are associated with a cost of 1.5 lives every year due to road traffic accidents in Spain’s capital cities. Considering that 78% of deaths from road traffic accidents in Spain occur on roads outside of cities,9 this could have an additional cost of another 5 lives every year. Albert Prats-Uribe Preventive Medicine and Public Health Training Unit Parc de Salut Mar Universitat Pompeu Fabra Agència de Salut Pública de Barcelona, Spain