Estimating Adolescent Weight Status from CDC's Age-Adjusted Growth Charts: The Importance of Age Precision in Obesity and Weight-Related Research

Purpose: The recent explosion of research on childhood obesity reflects national concern about the health consequences and costs of early onset obesity/overweight (Healthy People, 2010). However, assessing growing children's healthy weight presents critical and practical measurement challenges. One such issue is the calculation of age in the classification of child/adolescent weight status. Due to ongoing physical development, Body Mass Index (BMI) alone is not a meaningful indicator of adiposity before age 20. To facilitate interpretation of BMI for children/adolescents, the Centers for Disease Control (CDC) developed age-adjusted growth charts to determine weight status according to BMI, sex, and age in months. Despite these recommendations, studies still leave ambiguous the level of age precision (months vs. years) used to estimate weight status; and to date no studies have empirically addressed the comparative validity of weight status based on BMI-for-age in years vs. months. The purpose of this study is to use national data to assess the importance of age precision in determining unbiased estimates of adolescent weight status.

Methods: Using data from the National Longitudinal Study of Adolescent Health, responses from 6,501 adolescents (age: M=16.04, SD=1.75, 48% male) in grades 7-12 were examined. BMI was calculated from weight/height²; and age was calculated from dates of birth and interview. Age was coded at two levels of precision (months vs. years). Using BMI, gender, and the precise age in months, subjects were assigned to one of the four CDC weight status categories for children and adolescents, e.g., overweight (OW), risk for overweight (RW), healthy weight (HW), and underweight (UW). Non-parametric tests of significance were performed to examine the null hypothesis that no differences between weight statuses based on age in months vs. years exist.

Results: The Wilcoxon Signed Ranks test showed that weight status based on age in years resulted in incorrect weight status classification of 7% of girls and 8% of boys. Among the total sample, a Binomial test of significance showed that weight status proportions derived from age in years differed from proportions derived from age in months on three of the four weight statuses : HW (Asymp. Sig. =.006), UW (Asymp.Sig. =.002), and RW (Asymp.Sig. =.000). Three categories showed biased estimates among boys (HW, UW, RW) and two among girls (UW, RW). Compared to age in months, using age in years significantly underestimates the number of girls and boys who are UW, underestimates boys who are UW and HW, and overestimates boys who are RW.

Implications: Findings provide strong empirical support for estimating weight status based on age in months rather than years. Despite the ease of using age in years to estimate adolescent weight status, this method will introduce systematic bias into subsequent analyses, and interfere with the accuracy of baseline and post-intervention weight status assessment. This study adds to the growing literature on measurement issues related to childhood obesity and disordered eating, and supports a standard for age calculation when conducting weight-related population and intervention research among adolescents.