Comparison of WHO and Centers for Disease Control and Prevention Growth References for Classifying Height and Weight Status in Young Children in the United States-Affiliated Pacific.

BACKGROUND

Multiple international growth reference systems exist for classifying child undernutrition and excess weight, yet different references yield systematically different prevalence estimates. This has important implications for surveillance in Pacific Islander populations, where both the Centers for Disease Control and Prevention (CDC) and WHO references are used across jurisdictions.

OBJECTIVES

This study aimed to quantify classification discrepancies between WHO Growth Standards (ages 2-5 y), WHO Growth Reference (ages 6-8 y), and CDC Growth Charts (ages 2-8 y) in Pacific Islander children, and to assess implications for population health surveillance.

METHODS

This secondary cross-sectional analysis used 2014 data from the Children's Healthy Living Program (n = 5499 children aged 2-8 y from 11 Pacific jurisdictions). Weight and height were measured by trained, standardized measurers. Children were classified by each reference system using SAS programs provided by the WHO and CDC. Agreement was evaluated using McNemar's test, kappa coefficients, and percent agreement.

RESULTS

Height-for-age agreement was high across both age groups (≥98.3%). BMI-for-age agreement differed substantially by age: poor for ages 2 to 5 y (kappa = 0.26; 85.5% agreement) and high for ages 6 to 8 y (kappa = 0.86; 97.9% agreement). For ages 2 to 5 y, the WHO classified 11 percentage points more children as overweight than the CDC. When identical percentile cutoffs were applied across reference datasets, prevalence differences narrowed substantially, indicating that most discrepancies result from differing classification criteria rather than reference population differences.

CONCLUSIONS

Growth reference selection introduces systematic measurement bias into Pacific Islander child health surveillance. Surveillance systems should maintain reference consistency over time, explicitly report reference criteria, and consider dual reporting to enable valid temporal and geographic comparisons. Future research should evaluate which reference better predicts health outcomes in this population.