Infant nocturnal wakefulness: a longitudinal study comparing three sleep assessment methods

Study Objectives: To examine the convergence between actigraphy, sleep diaries, and the Brief Infant Sleep Questionnaire (BISQ) in the assessment of infant nocturnal wakefulness (i.e. minutes awake after sleep onset [WASO] and number of night-wakings [NW]) in the context of a longitudinal study. Methods: The sample included 226 families, who were recruited during pregnancy. Data were collected at 3, 6, 12, and 18 months postpartum. Infants' sleep was monitored at home for five nights using actigraphy, sleep diaries, and the BISQ. Outcome measures included WASO and NW, as well as sleep latency, sleep duration, and sleep onset. Results: Trajectory analyses demonstrated that all three methods showed declines in NW and WASO from 3 to 18 months. Statistically significant correlations were found between the three methods at all assessment points for all sleep variables. However, agreement rates (using Krippendorff's a and Bland-Altman analyses) between actigraphy and parental reports were poor. For NW, agreement between actigraphy and parental reports at 18 months was lower than that at 3 and 6 months. Diaries and BISQ showed satisfactory agreement for sleep latency. Conclusions: Although the three methods' measures of infant nocturnal wakefulness are significantly correlated during infancy, absolute agreement between these methods is poor overall. The growing disagreement between actigraphy and parental reports (in NW) across development probably suggests that parents become less aware of infants' awakenings, due to the increasing ability of infants to self-soothe. Using both objective and subjective assessment methods seems especially important after the age of 6 months. Statement of Significance The reliable evaluation of infants' night-waking problems is crucial for clinical and research purposes. This longitudinal study demonstrated, for the first time, how actigraphy, sleep diaries, and the Brief Infant Sleep Questionnaire converge in the assessment of infant nocturnal wakefulness from 3 to 18 months. Examining the agreement between different sleep assessment methods in a longitudinal context is necessary, because the convergence may change along infant development. In the present study, all three methods were significantly correlated and clearly demonstrated that infant sleep consolidated over time. However, absolute agreement between the three methods was poor overall, and worsened over time for the number of awakenings. Future studies should examine the correspondence between different sleep assessment methods in clinically sleep-disturbed infants.