WHO Validation

This repository provides utilities that use RCPCH-modified WHO growth packages to compute anthropometric measurements corresponding to requested z-scores.

Packages used (RCPCH forks/branches):

• anthro — WHO growth standards for children under 5 years (uses age in days).
• anthroplus — WHO growth references for children 5–19 years (uses age in months).

Both packages include RCPCH changes used here:

• z_precision support to control z-score decimal precision.
• anthro_measurements / anthroplus_measurements inverse-LMS functions to compute measurements from a requested z-score, with optional extreme-value correction (correct_extreme).

Script helpers

The file who-validation.R exposes three helpers:

• ensure_latest_packages(force = FALSE)
  • Installs the anthro and anthroplus packages from the z-to-measurement branch on GitHub when they are missing. Pass TRUE to force reinstall.
• load_random_dates(rd_path = "data-files/random_dates.csv")
  • Reads random_dates.csv and returns a data.frame with computed age_months (integer) and age_years.
• compute_measurements(df, measurement_method = "length", sex = NULL, requested_z = 2.5, measurement_precision = 2, correct_extreme = TRUE)
  • Given a dataframe produced by load_random_dates(), computes the requested measurement for each row. Uses anthro::anthro_measurements for ages < 61 months (anthro expects age in days) and anthroplus::anthroplus_measurements for ages >= 61 months (expects months).
  • Returns the dataframe with a new column added:
    • measurement_method = "length" -> height_in_cm
    • measurement_method = "weight" -> weight_kg
    • measurement_method = "bmi" -> bmi
    • measurement_method = "headc" -> headc

save_measurements_csv(df, file_path = "data-files/random_dates_with_measurements.csv", na_string = "", compress = FALSE)

• Writes the dataframe produced by compute_measurements() to CSV. If compress = TRUE the file is written gzipped and file_path will have .gz appended if not present.

Example

Run this example in an R session (after installing the modified packages or calling ensure_latest_packages(TRUE)):

# install or update packages if needed (uncomment to force reinstall)
# ensure_latest_packages(TRUE)

# show examples
head(result[c("start_date", "age_days", "age_months", "height_in_cm")])

# load and prepare the random dates dataset
df <- load_random_dates()

# compute heights corresponding to z = 2.5
head(result[c("start_date", "age_days", "age_months", "height_in_cm")])

Notes

• The helpers call the measurement functions directly from the package namespaces and will warn and return NA values if the functions are not available.
• If you install or update the packages while an R session already has their namespaces loaded, you may need to restart the R session for the new code to be used. ensure_latest_packages() will install the packages but will not automatically restart R.

Data

Provide data-files/random_dates.csv with at least three columns: start_date, end_date, and age_days. load_random_dates() expects this format.