Features of the CDC Growth Charts
The data used to construct the 2000 CDC charts included a nationally representative reference population of children and teens from 2 years to 20 years of age. State-of-the-art statistical smoothing methods were used to fit the data from national surveys to create smooth curves.
There are several clinically significant features of the 2000 CDC growth charts that include:
- BMI-for-age charts for children and teens aged 2 years to 20 years.
- Stature-for-age and weight-for-age charts for children and teens aged 2 years to 20 years.
- Set 1 of the clinical growth charts has the 5th, 25th, 50th, 75th, 90th, and 95th percentile lines. In addition, the BMI-for-age growth chart includes the 85th percentile line to identify overweight. These BMI charts are the most commonly used in the United States to screen for underweight, healthy weight, overweight, and obesity.
- Set 2 of the clinical growth charts has the 3rd and 97th percentile lines on specific charts for selected applications. Pediatric endocrinologists and others who assess the growth of children with special health care requirements may wish to use these charts.
- Between the 3rd and 97th percentiles the smoothed percentile curves have corresponding z-scores.
- Smoothed percentile curves and z-scores are used to evaluate the growth of children.
- Percentiles are the most commonly used clinical indicator to assess the size and growth patterns of individual children in the United States. Percentiles rank the position of an individual by indicating what percent of the reference population the individual would equal or exceed. For example, on the weight-for-age growth charts, a 5-year-old girl whose weight is at the 25th percentile, weighs the same or more than 25 percent of the reference population of 5-year-old girls, and weighs less than 75 percent of the 5-year-old girls in the reference population (Kuczmarski et al., 2002).
- Because z-scores have a direct relationship with percentiles, a conversion can occur in either direction using a standard normal distribution table. For every z-score there is a corresponding percentile and vice versa population (Kuczmarski et al., 2002).