In growthcharts/yzy: Translate Measurements, Z-Scores and Centiles with the RIF format

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)

Background

The Reference Interchange Format (RIF) is a simple text format for storing and exchanging growth references. Growth references typically consist of a conditioning variable, usually age or height, and a set of parameter estimates at each tabulated age or height. For example, the first four rows of the WHO Growth Standard for lengths of boys are

library(centile)
ref <- load_reference("who_2006_hgt_male_")
head(ref, 4)

In this example, x stands for age in years and the L, M and S columns stand for the estimated parameters of the LMS model (Cole & Green 1992). The study attributes stores some additional information:

head(attr(ref, "study"))

The RIF format is a simple way to store references as text files that can serve as input for different systems.

General specification

Let us first take a look at an example. The following block contains the first 12 rows of the file who_2006_bmi_female_.txt.

name = who
year = 2006
yname = bmi
sex = female
distribution = LMS
citation = WHO 2006
publication = WHO Multicentre Growth Reference Study Group (2006). WHO Child Growth Standards: Length/height-for-age, weight-for-age, weight-for-length, weight-for-height and body mass index-for-age: Methods and development. Geneva: World Health Organization; pp 312. (web site: http://www.who.int/childgrowth/publications/en/)
[data]
x   L   M   S
0   -0.0631 13.3363 0.09272
0.0027  0.0362  13.3185 0.0936
0.0055  0.1355  13.3006 0.09448
...

A RIF formatted files should follow the following principles:

• A RIF file contains the parameters of one reference distribution only. Thus, separate distributions for males and females go into different files;
• The file starts with a header that describe the contents of the reference table. The section below provides the accepted keywords;
• The = sign separates the keyword and its value;
• A special keyword [DATA] separates the header and data;
• The first line after [DATA] contain the variable names. These names adhere to certain naming conventions given below;
• There is one column named x, the conditioning variable;
• Data columns are separated by tabs.

Keywords

Keyword | Required | Interpretation ------------- | ----------- | --------------- name | Yes | two-letter country code, who, eu27, fenton, and so on year | Yes | four number, may be left blank yname | Yes | Outcome: hgt, wgt, hdc, wfh, bmi, dsc sex | No | male or female, may be left blank sub | No | Subgroup distribution | Yes | NO, LMS, BCCG, BCPE or BCT citation | No | Short citation phrase publication | No | Reference to source documents remark | No | Free text for additional remarks tx | No | Transformation of x prior to Z-score calculation xt | No | Back-transform to x after Z-score calculation ty | No | Transformation of y prior to Z-score calculation yt | No | Back-transform to y after Z-score calculation

Any other keywords, each on its own line, can be added and will be saved in attr(ref, "study"). The number of header rows is limited to 24.

Column names

Apart from the column named x two or more column with parameters, depending on the specification of the distribution.

Distribution | Required | Interpretation ------------- | --------------------------- | --------------- NO | mean, sd | Mean, Standard deviation LMS | M, S, L | Location, Scale, Skewness BCCG | mu, sigma, nu | Location, Scale, Skewness BCPE | mu, sigma, nu, tau | Location, Scale, Skewness, Kurtosis BCT | mu, sigma, nu, tau | Location, Scale, Skewness, Kurtosis

Some other distributions can be specified, but no Z-scores will be calculated:

Distribution | Required | Interpretation ------------- | --------------------------- | --------------- MEA | mean | Mean only PCT | pxx, pyy | One or more percentiles (xx, yy numeric) MP | mean, pxx, pyy | Mean + Percentiles (xx, yy numeric)

growthcharts/yzy documentation built on Oct. 28, 2024, 10:32 a.m.