Errors in calculating weight-for-height.

Abstract

Weight-for-height is widely used as an index of recent malnutrition and often forms the basis on which major policy decisions are made about nutrition interventions. This article illustrates potential for error in the calculation of weight-for-height. Comparisons are made between results obtained using the international reference figures published by the World Health Organization, the plasticized cards distributed by Teaching-aids At Low Cost and by using the computer software package Epi Info (versions 5 and 6). An assessment is made of the errors introduced by treating supine measurements incorrectly as stature or by treating stature measurements as supine. Comparisons are also made between the use of standard deviation scores and percentages of the median. It is recommended that the international reference figures be published in a more convenient tabular form and that the plasticized cards and percentages of the median should not be used for studies involving international comparison. Investigators who plan to use Epi Info are advised to take into account the way in which it calculates weight-for-height when collecting and analyzing their height measurements.