UCSF

We know of no studies comparing parent-reported sleep with accelerometer-estimated sleep in their relation to paediatric adiposity. We examined: (i) the reliability of mother-reported sleep compared with accelerometer-estimated sleep; and (ii) the relationship between both sleep measures and child adiposity. The current cross-sectional study included 303 Mexican American mother-child pairs recruited from Kaiser Permanente Northern California. We measured sleep duration using maternal report and accelerometry and child anthropometrics. Concordance between sleep measures was evaluated using the Bland-Altman method. We conducted zero-ordered correlations between mother-reported sleep, accelerometer-estimated sleep and child BMI z-scores (BMIz). Using linear regression, we examined three models to assess child BMIz with mother-reported sleep (model 1), accelerometer-estimated sleep (model 2) and both sleep measures (model 3). Children had an average age of 8.86 years (SD = 0.82). Mothers reported that their child slept 9.81 ± 0.74 h [95% confidence interval (CI): 9.72, 9.89], compared to 9.58 ± 0.71 h (95% CI: 9.50, 9.66) based on accelerometry. Mother-reported sleep and accelerometer-estimated sleep were correlated (r = 0.33, P < 0.001). BMIz outcomes were associated negatively with mother-reported sleep duration (model 1: β = -0.13; P = 0.02) and accelerometer-estimated sleep duration (model 2: β = -0.17; P < 0.01). Accounting for both sleep measures, only accelerometer-measured sleep was related to BMIz (model 3: β = -0.14, P = 0.02). Each sleep measure was related significantly to adiposity, independent of covariates. Accelerometry appeared to be a more reliable measure of children's sleep than maternal report, yet maternal report may be sufficient to examine the sleep-adiposity relationship when resources are limited.