A Comparison of Sleep Duration Accuracy Between Questionnaire and Accelerometer in Middle Childhood

Purpose Healthy sleep is vital to children’s well-being, and assessing sleep efficiently and accurately can help understand children's lifestyles. Due to the difficulty in objectively measuring sleep duration using wearable sensors in large-scale surveys of children, self-administered questionnaires are often used in Japan; however, their accuracy is uncertain. We evaluated and compared the accuracy of questionnaire-based sleep times to those of wearable sensors. Methods This observational study was conducted between November 2019 and January 2020. A self-administered questionnaire on lifestyle habits and ActiGraph GT3X+ (ActiGraph, Inc., Pensacola, USA) accelerometer data were collected from 40 fourth-grade elementary school students in Kagawa Prefecture, Japan. We analyzed measurements for 256 days out of 280 days (40 persons × 7 days) after excluding days when the rate of wearing the accelerometer was < 90%. Results The median sleep duration per accelerometry was 453 minutes, and the median time in bed was 519 minutes. Questionnaire-based time in bed was 11 minutes longer, with relatively high inter-individual variability. The difference in bedtime was 26 minutes earlier, and wake-up time was 12 minutes earlier for the questionnaire. The average sleep efficiency was 87.4%, and one-third of the children had sleep efficiency < 85%. Conclusion The difference in sleep duration by questionnaire compared to accelerometry was approximately 10 minutes, suggesting the questionnaire may determine sleep duration with accuracy.