Confirmatory factor analysis of the Sleep Disturbance Scale for Children (SDSC) in a clinical sample of children and adolescents.

The Sleep Disturbance Scale for Children was developed as a parent-report measure to screen for sleep disturbances within the preceding 6 months. Notably, the scale was developed using a sample of typically developing children and children with sleep disorders. The aim of this study was to factor analyse the Sleep Disturbance Scale for Children using a clinical sample of children with a range of neuropsychological conditions and co-morbidities, and determine whether the original six-factor structure was retained. Four-hundred and sixteen children aged 5-17 years were assessed at the Neurosciences Unit (Perth, Western Australia) as part of routine, clinical neuropsychological assessment. Parents and guardians also completed the Sleep Disturbance Scale for Children to rate their child's sleep. Confirmatory factor analysis of the original Sleep Disturbance Scale for Children model (Bruni et al. J. Sleep Res., 1996, 5: 251-261) revealed a less than ideal fit. Three adjustments were made to the model based on factor loadings and modification indices. The sleep hyperhidrosis factor (including items 9 and 16) along with item 10 was removed, leaving a five-factor Sleep Disturbance Scale for Children model. The five-factor model (Sleep Disturbance Scale for Children-R) was factor analysed, and examination of model fit statistics indicated that this new model produced good fit. Additional analyses revealed that older children had greater difficulty falling and staying asleep, and with daytime sleepiness. However, no significant differences were observed across gender, diagnosis or socioeconomic status. The results of this study suggest that the Sleep Disturbance Scale for Children-R may be a more appropriate measure when assessing clinical samples. However, further research is required to validate the Sleep Disturbance Scale for Children-R against objective measures of sleep and to determine appropriate t-score cut-offs.