Convert measurements into Z-scores

Converts measurements into age- or hgt-conditional standard deviation scores (SDS, Z-score) using one or more external reference distributions.

y2z(
  y,
  x,
  refcode,
  pkg = "centile",
  verbose = FALSE,
  dec = 3L,
  rule = 1L,
  tail_adjust = FALSE,
  ...
)

Arguments

y: A numerical vector with measurements`.
x: A vector containing length(y) values of the numerical covariate (typically decimal age or height) at which conversion is desired. A scalar x will be expanded to length(y).
refcode: A character vector with length(y) elements, each of which is the name of a reference. Scalar refcode expands to length(y). Alternatively, refcode may be an object of class reference, typically produced by import_rif(). If NULL the procedure returns NA's.
pkg: The package containing the references in the R/sysdata.dta object. The package needs to be loaded. The default pkg = "centile" searches in the centile package.
verbose: Set to TRUE to turn on warnings
dec: A scalar value indicating the number of decimals used to round the value. The default is 3 decimals.
rule: an integer (of length 1 or 2) describing how interpolation is to take place outside the interval [min(x), max(x)]. If rule is 1 then NAs are returned for such points and if it is 2, the value at the closest data extreme is used. Use, e.g., rule = 2:1, if the left and right side extrapolation should differ.
tail_adjust: Logical. If TRUE the procedure applies the WHO method for tail adjustment under the LMS distribution. We do not recommend this. The default is FALSE.
...: Not used.

Value

A vector with length(y) elements containing the Z-scores.

Note

The type of reference distribution is defined by the distribution field of the study attribute in the reference. The function executes transformations specified by the tx and ty fields before calculating the Z-score.

Functions y2z() and z2y() functions are the inverse of each other.

Author

Stef van Buuren, 2021

Examples

# Weight SD 1 month old boy and girl of 4 KG using built-in WHO references
y <- c(4, 4)
x <- c(1/12, 1/12)
refcode <- c("who_2006_wgt_male_", "who_2006_wgt_female_")
y2z(y, x, refcode)
#> [1] -0.821 -0.332

# using external reference, for females only
fn <- system.file("testdata/nl_2009_wgt_female_nl.txt", package = "centile")
myref <- import_rif(fn)
head(myref)
#> # A tibble: 6 × 4
#>        x     L     M     S
#>    <dbl> <dbl> <dbl> <dbl>
#> 1 0.0192  1.20  3.66 0.115
#> 2 0.0383  1.17  3.83 0.115
#> 3 0.0575  1.15  4.01 0.115
#> 4 0.0767  1.13  4.18 0.115
#> 5 0.0958  1.11  4.35 0.115
#> 6 0.115   1.09  4.52 0.115
head(attr(myref, "study"))
#>         name         year        yname          sex          sub distribution 
#>         "nl"       "2009"        "wgt"     "female"         "nl"        "LMS" 
y2z(y, x, myref)
#> [1] -0.486 -0.486