Measuring malnutrition in famines: are weight-for-height and arm circumference interchangeable?

Data from two surveys in Sudan have been used to examine whether weight-for-height (WFH) and mid-upper arm circumference (MUAC) can be used interchangeably at a population level to define the proportion of children aged one to four years that are malnourished, whether they identify the same individual children as malnourished, and whether the relationship between WFH and MUAC varies with age. A MUAC cut-off of 13.0 cm consistently defined approximately the same proportion of children malnourished as 80% WFH in all seven groups of children examined, even though the proportion of children with less than 80% WFH varied between 8.6% and 30.7%. However, sensitivity/specificity analysis showed that many of the children identified as malnourished by the two indices were not the same individuals. Both the MUAC cut-off defining the same proportion malnourished as 80% WFH, and the sensitivity/specificity values, varied substantially with age. Studies of other populations have revealed both different MUAC cut-offs defining the same proportion of children malnourished as 80% WFH, and different sensitivities and specificities of MUAC relative to WFH. We do not recommend the direct comparison of data from surveys using WFH and those using MUAC.