RESUMEN
OBJECTIVE: To investigate the validity of the Brief Infant Sleep Questionnaire (BISQ), in assessing sleep quality in childhood. METHODS: This was a validation study with children from the Pelotas 2015 Birth Cohort. BISQ was applied to mothers when their children were 3, 6, 12, and 24 months of age. The poor sleep indicators analysed, as defined by BISQ, were >3 wakings per night, nocturnal wakefulness >1 h and total sleep duration <9/24 h, compared to number of wakings per night and nocturnal and total sleep duration defined by actigraphy taken as the gold standard. The Actiwatch wGT3X-BT device was used by the child consecutively during five days at three and six months and for three days at 12 and 24 months. At each age the prevalence, sensitivity, specificity, accuracy, and positive (PPV) and negative predictive values (NPV) of each sleep indicator was calculated. RESULTS: A total of 586 children were enrolled in the study. Nocturnal wakefulness >1 h was the most frequent indicator at all ages, with higher sensitivity (varying from 27.5% at six months to 54.8% at three) and lower specificity (53.4% at three months to 79.4% at six months), in comparison to the other sleep indicators. Specificity for >3 wakings and total sleep duration <9 h was greater than 85.0% at all the ages. Higher accuracies were observed for total sleep <9 h at 3 (85.6%), 6 (88.2%) and 12 months (73.6%) and for > 3 wakings at 24 months (84.5%). The sensitivity for the presence of at least one indicator decreased with age from 56.0% at three months to 35.8% at 24 months, whereas the specificity increased from 50.6% at three months to 63.8% at 24 months. CONCLUSION: The high specificity of the BISQ sleep parameters supports the validity of parents' reports on sleep-related problems in childhood for use in epidemiological studies.