Abstract
The gold standard for categorisation of weight status is clinically measured body mass index (BMI), but this is often not practical in large epidemiological studies. To determine if a child's weight perception or a mother's perception of a child's weight status is a viable alternative to measured height and weight in determining BMI classification. Secondary outcomes are to determine the influence of a mother's BMI on her ability to categorise the child's BMI and a child's ability to recognise his/her own BMI. Cross-sectional analysis of the growing up in Ireland cohort study, a nationally representative cohort of 8568 9-year-old children. The variables considered for this analysis are the child's gender, BMI (International Obesity Taskforce grade derived from measured height and weight) and self-perceived weight status, and the mother's weight perception of the child, BMI (derived from measured height and weight) and self-perceived weight status. Cohen's weighted-kappa was used to evaluate the strength of the agreement between pairwise combinations of the BMI variables. Cumulative and adjacent categories logistic regression were used to predict how likely a person rates themselves as under, normal or overweight, based on explanatory variables. Mothers are more accurate at correctly classifying their child's BMI (κ=0.5; confidence intervals (CI) 0.38-0.51) than the children themselves (κ=0.25; CI 0.23-0.26). Overweight mothers are better raters of their child's BMI (κ=0.51; CI 0.49-0.54), compared with normal (κ=0.44; CI 0.41-0.47) or underweight mothers (κ=0.4; CI 0.22-0.58), regardless of whether the mother's BMI is derived from measured height and weight or self-perceived. The mother's perception of the child's weight status is not an influencing factor on the child's ability to correctly classify him/herself, but the child's self-perceived weight status influences the mother's ability to correctly classify the child. A mother's BMI classification of her child is a viable alternative to BMI measurement in large epidemiological studies.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have