Trajectory and determinants of agreement between parental and physicians' reports of childhood atopic dermatitis.

Abstract

Parent self-administered reports are commonly used in studies on childhood atopic dermatitis (AD) but data on its validity are sparse. We aimed to examine the agreement between parent- and physician-reported measures of childhood AD throughout early life and identify the determinants. In this prospective cohort study, we used data of 449 infants and their mothers recruited in the Ulm SPATZ Health Study in Germany. Longitudinal data of parental and children's caring physicians' reports were used to assess the point and cumulative agreement of parent- and physician-reported AD diagnoses, AD onset age, and trend of agreement at child ages between 1 and 6 years overall and by child and parent demographics and health conditions. A Generalized Estimating Equation model was fitted to identify factors associated with the sensitivity of parent reports. The point agreement between parent- and physician-reported AD was substantial at the age of 1 (kappa=0.63, 95% CI: 0.51-0.75) but declined with age and became fair after the age of 3 (kappa < 0.40). The cumulative agreement remained moderate at the age of 6 (kappa=0.51, 95% CI: 0.43-0.60). Parents had a bias towards delayed reporting of the AD onset age. The AD severity was the only strong determinant for the agreement of AD diagnoses and largely explained the variance of the sensitivity of parent reports. The disagreement between parent- and physician-reported AD increases with child age, likely due to the change of AD severity. Using parent-reported data might miss a substantial portion of mild childhood AD cases.