Pacific D
Rationale for the Creation of the Arkansas Child Population BMI Growth Curve (307879)
Anthony Goudie, Arkansas Center for Health ImprovementKanna Lewis, Arkansas Center for Health Improvement
*Joseph W Thompson, Arkansas Center for Health Improvement
Keywords: Severely obese, LMS transformation, BMI growth curve
The Centers for Disease Control and Prevention (CDC) 2000 growth curves, constructed by Lambda-Mu-Sigma (LMS) transformation, do not produce z-score estimates above the 97th percentile, and the CDC warns against extrapolation of the growth curve. Many methods to further quantify the degree of obesity in the upper distribution of the tail have been proposed, but there is no universally accepted standardized metric with precision and statistical robustness. In Arkansas, height and weight measures have been collected on public school children since 2004. Currently, 22% of kindergarten to grade 8 students are obese (= 95th percentile of the CDC curves), and 16% of children are above the CDC 97th percentile. To ascertain the effectiveness of interventions for the severely obese population, we constructed new growth curves using LMS on 10 years of longitudinal BMI measures collected from Arkansas children. The CDC 97th percentile on average translates to the 84th percentile on the new growth curves, with additional distributions available for higher weights. This method enables us to accurately assess interventions for the most severely obese children in Arkansas.