cdc_z: Generate Sex- and Age-Standardized Weight, Height, and BMI...
In cdcanthro: Standardized Metrics Based on the CDC and WHO Growth Charts

View source: R/cdc_z.R

cdc_z

R Documentation

Generate Sex- and Age-Standardized Weight, Height, and BMI Metrics From the CDC Growth Charts

Description

Generate z-scores, percentiles, and other sex- and age-standardized metrics for weight, height, and BMI based on the 2000 CDC growth charts. For children with obesity, the extended BMI z-scores and percentiles are calculated. The calculations are restricted to children with ages between 24.0 (inclusive) and < 240 months. For younger children, use who_z. Note that cdc_z is the preferred name; the previous name, cdcanthro, will work but is deprecated.

The BMI metrics include z-scores and percentiles, along with newer metrics such as extended BMIz and percent of the 50th and 95th percentiles. The extended BMI metrics are used for BMIs above the 95th percentile.

Usage

cdc_z(data, age = agemos, wt = wt, ht = ht, bmi = bmi, 
            sex_code = 1:2, all_vars = FALSE)
cdcanthro(data, age = agemos, wt = wt, ht = ht, bmi = bmi, 
            sex_code = 1:2, all_vars = FALSE)

Arguments

`data`	a `data.frame`, `data.table`, or tibble (`tbl_df`) containing the input data. The class of the returned object matches the class of `data`.
`age`	Age in months, specified as accurately as possible. Required.
`wt`	Weight in kg. Optional. If omitted, WAZ-related metrics will be missing.
`ht`	Height in cm. Optional. If omitted, HAZ-related metrics will be missing.
`bmi`	BMI in `kg/m^2`. Optional. If omitted but both `wt` and `ht` are supplied, BMI is computed as `wt / (ht / 100)^2`. If all three of `wt`, `ht`, and `bmi` are omitted, the function stops with an error.
`sex_code`	Numeric, single-digit coding for sex if other than 1 for boys and 2 for girls. Optional.
`all_vars`	logical. If `TRUE`, returns additional intermediate variables (L, M, S values; sigma; and the distance-from-median metrics from Freedman et al.). Default is `FALSE`.

Details

Argument values can be quoted (e.g., age='agemos') or unquoted (age = agemos).

The function expects the child's sex to be named sex; either upper- or lower-case is fine. You cannot have both SEX and sex in your data. Values can be coded as boys/girls, b/g, male/female, m/f, or 1/2. Character values can be upper- or lower-case, and only the first character is considered. Be aware that a code of 'missing' will be interpreted as a boy.

If sex is numeric and not coded as 1/2, specify this in the optional sex_code argument; the code for boys is the first number. For example, 0/1 coding would be specified as sex_code = 0:1 or sex_code = c(0,1).

Weight is in kg, height is in cm, and BMI is in kg/m^2.

At least one of wt, ht, or bmi must be present. Metrics that cannot be calculated from the supplied variables are silently omitted from the output. If a supplied column name doesn't exist in data (a likely typo), a warning is issued.

Age in months should be given as accurately as possible because the function linearly interpolates between completed months of age. If ages are given as the completed number of months (as in NHANES), add 0.5 because values such as 36 represent the interval [36, 37) months. If the function detects integer numbers for all ages, it will add 0.5 to all age values during the calculations; the input ages are not changed.

If age is in days, divide by 30.4375 so that a child who is 3672 days old would have an age in months of 120.641. For additional information on age, see the ‘See Also’ section below.

If all_vars = TRUE, all variables in the Freedman et al. paper will be output, along with the L, M, and S values for each child and the value of sigma for the half-normal distribution.

For children without obesity, BMI z-scores are calculated as

Z = ((BMI / M)^L - 1) / (L \times S)

where BMI is the child's BMI, L is the Box-Cox transformation for normality for the child's sex and age, M is the median, and S is the coefficient of variation. Reference data are the merged LMS files from the CDC (see ‘See Also’).

For children with obesity, BMI percentiles are calculated as

90 + 10 \times \Phi((\mathrm{BMI} - \mathrm{p95}) / \sigma)

where p95 is the sex- and age-specific 95th percentile, \Phi is the CDF of a N(0, 1) distribution (pnorm()), and sigma is the scale parameter of the half-normal distribution. Values of sigma for children with obesity are based on formulas in Wei et al.

The class of the returned object matches the class of the input data: a data.frame input returns a data.frame, a data.table returns a data.table, and a tibble (tbl_df) returns a tibble (provided the tibble package is installed; otherwise a data.frame is returned with a warning).

Value

A data.table, data.frame, or tibble containing the original data and various weight, height, and BMI metrics.

If your dataset has ages that are < 24 or >= 240 months, these rows will be present in the output dataset. All calculated metrics will be NA. If this is not what you want, delete these rows before running cdc_z().

Variables in output:

waz, haz, bmiz: Sex- and age-standardized z-scores for weight, height, and BMI. bmiz is based on the 2000 CDC growth charts for non-obese children and extended BMIz for obese children.
wap, hap, bmip: Sex- and age-standardized percentiles for weight, height, and BMI. bmip is based on the extended BMI metrics for obese children.
mod_waz, mod_haz, mod_bmiz: Modified z-scores used to flag outliers; see the CDC SAS program.
orig_bmip, orig_bmiz: Original BMI percentiles and z-scores based solely on the LMS method. orig_bmiz and bmiz are identical for BMIs at or below the 95th percentile.
p95: The 95th percentile for BMI in the reference data for each sex and age.
bmip95: BMI expressed as a percentage of the 95th percentile; 120 percent is the lower threshold for severe obesity.
perc_median: Percent difference between the child's BMI and the median (50th percentile) BMI for the child's sex and age.

If all_vars = TRUE, additional BMI metrics described in Freedman et al. are also returned. These express BMI as distance or percent distance from the median. To convert to percent of the median, add 100 to the values.

Note

Argument names can be quoted or unquoted: cdc_z(data, age, wt, ht, bmi), and cdc_z(data, 'age', 'wt', 'ht', 'bmi') produce the same result.

Author(s)

David Freedman DavidSFreedman@gmail.com

References

Kuczmarski RJ, Ogden CL, Guo SS, Grummer-Strawn LM, Flegal KM, Mei Z, et al. 2000 CDC Growth Charts for the United States: methods and development. Vital and Health Statistics Series 11, Data from the National Health Survey 2002;11:1–190.

Wei R, Ogden CL, Parsons VL, Freedman DS, Hales CM. A method for calculating BMI z-scores and percentiles above the 95th percentile of the CDC growth charts. Annals of Human Biology 2020;47(6):514–521. \Sexpr[results=rd]{tools:::Rd_expr_doi("10.1080/03014460.2020.1808065")}

Freedman DS, Woo JG, Ogden CL, Xu JH, Cole TJ. Distance and Percent Distance from Median BMI as Alternatives to BMI z-score. British Journal of Nutrition 2019;124:1–8.

Examples

data <- data.table::CJ(sex = 1:2, agem = 120.5, wtk = c(30, 65), htc = c(135, 144))
data[, let(bmi = wtk / (htc / 100)^2)]
out <- cdc_z(data, age = agem, wt = wtk, ht = htc, bmi = bmi)
round(out, 2)[1:5]

out <- cdc_z(data, age = agem, wt = wtk, ht = htc, bmi = bmi, all_vars = TRUE)
round(out, 2)[1:5]

d <- data.frame(sex = c(1, 2, 1, 2, 2),
                age = c(141, 54, 217, 155, 52),
                wt  = c(57, 25, 72, NA, 17.7),
                ht  = c(143, 102, 166, 169, 105))
d <- cdc_z(d, age, wt, ht)
round(d, 2)

d <- data.frame(sex = c(0, 1, 0),
                age = c(120.5, 60.5, 180.5),
                wt  = c(40, 25, 65))
cdc_z(d, age, wt, sex_code = 0:1) # note use of sex_code

# tibble input returns a tibble (requires the tibble package)
if (requireNamespace("tibble", quietly = TRUE)) {
  d_tbl <- tibble::as_tibble(d)
  cdc_z(d_tbl, age, wt, sex_code = 0:1)
}

head(NHanes) # age is given as completed months
d <- data.table::copy(NHanes)
head(cdc_z(d, agemos, wt, ht, bmi))[,.(seqn, agemos, waz, haz, bmiz)]
d[, let(agemos = agemos + 0.5)]
head(cdc_z(d, agemos, wt, ht, bmi))[,.(seqn, agemos, waz, haz, bmiz)]
# same Z-scores

cdcanthro documentation built on Oct. 31, 2026, 1:07 a.m.