Intended for healthcare professionals

Rapid response to:

Endgames Statistical Question

Standardising outcome measures using z scores

BMJ 2014; 349 doi: https://doi.org/10.1136/bmj.g5878 (Published 01 October 2014) Cite this as: BMJ 2014;349:g5878

Rapid Response:

Statement (d) in the Statistical Question (4 October): "The greater the [BMI] z score, the heavier a child compared with other children of the same age and sex" is incorrectly said to be true. It presupposes that BMI measures heaviness not fatness. Because BMI is calculated as weight/height squared, it is quite possible for the heavier of two children of the same age and sex to have a lower BMI, and a lower BMI z score, if the child is sufficiently tall.

A simple example makes the point. Take two boys aged 10, heights 130 & 140 cm and weights 43 & 47 kg respectively. So the taller boy is also the heavier. However their BMIs are 25.4 and 24.0, with BMI z scores 2.75 and 2.50 respectively (using the UK90 reference). So the taller and heavier boy is also the thinner, with lower BMI.

Competing interests: No competing interests

15 October 2014
Tim J Cole
professor of medical statistics
UCL Institute of Child Health