RESUMEN
BACKGROUND: The overall purpose of the present study was to attain more insight in month-to-month variation of sleep duration and quality in order to improve design and interpretation of, e.g. epidemiological studies using sleep as outcome. METHODS: The study design entailed monthly self-reports from 38 (26 women/12 men) daytime workers, who completed the Karolinska Sleep Diary (KSD) once a month during one year. A subgroup (n = 16) also wore actigraphs on one day every month during a year. Self-reports of bedtime, time of awakening, sleep duration, individual sleep characteristics, disturbed sleep index (DSI, 4 items) and awakening index (AWI, 3 items) were analyzed together with actigraphy-derived measures. Hours of daylight were used to test for circa-annual variation in statistical models adjusted for intake of hypnotics and alcohol, gender, age and within-person variability. RESULTS: Hours of daylight were found to be associated to self-reported bedtime (p = 0.032) and DSI (p = 0.030), thereby indicating a circa-annual variation. Bedtime was delayed by 1.8 min (95% CI: 0.6-2.9 min) per 1 hour increase in length of daylight. Sleep was slightly more disturbed during the winter. CONCLUSION: Only circa-annual variation in self-reports of bedtime and DSI were observed in a healthy daytime working population, and the effects were small. Therefore potential bias due to circa-annuality in the studied parameters appears to be of limited concern in adult daytime working populations.