Estimation and feasibility of correction modelling for mother-reported child height and weight at 2 years using data from the Australian CHAT trial

Correction modelling using reported BMI values has been employed in adolescent and adult populations to improve the accuracy of self-reporting. This study aimed to evaluate the feasibility of establishing correction modelling for mother-reported child height and weight at 2 years using data from an Australian trial in 2019. Correction modelling for BMI was conducted using mother-reported and objectively measured height and weight of 2-year-olds. Mother-reported height, weight and BMI values of 2-year-old children were adjusted based on objectively measured anthropometric data using linear regression models. ‘Direct’ and ‘indirect’ corrections were applied to the correction of BMI values. We defined the direct collection as using corrected BMI values that were predicted directly by the model and indirect correction as using corrected weight and height values to calculate corrected BMI values. Corrected BMI values via the indirect correction showed higher sensitivity or similar specificity in predicting overweight status, compared to the direct correction, and also showed higher agreement with measured values compared to the mother-reported measures. Corrected self-reported measures via an indirect correction had a better accuracy and agreement with the objectively measured data in the BMI values and classification of overweight, compared to the mother-reported values.