Are stunted child - overweight mother pairs a real defined entity or a statistical artifact?

ABSTRACT

In a methodological contribution, Dieffenbach & Stein (DS) (The Journal of Nutrition, 142(4), 771-773.) concluded that the double burden of malnutrition (DBM), represented by stunted child - overweight mother pairs (SCOM), is a statistical artifact, meaning that SCOM does not describe a unique phenomenon because the observed rates of SCOM across a number of countries were not strongly different from the product of observed rates of maternal overweight (OM) and child stunting (SC), which DS referred to as the expected rate of SCOM. However, a growing literature continues to use SCOM as an indicator of the DBM. This study shows that the analysis by DS is not sufficient to conclude that SCOM can be explained by the co-occurrence of OM and SC due to chance alone because the analysis by DS was conducted at the country level, but applied to SCOM, which is a household-level variable. Using Demographic and Health Surveys data from 202 country-year data sets, we do not confirm important implicit assumptions that are required for the claim by DS to be supported. We also outline that comparing the expected to the observed rate of SCOM is primarily informative when putting it in relation to factors that influence the supply and demand of food consumed by households. When considering these factors, we find further evidence that it is misleading to consider SCOM as a statistical artifact, as the difference between the observed and the expected rate of SCOM significantly differs by household wealth. Recognizing that SCOM is a distinct phenomenon is important for policymakers who develop double-duty strategies that address malnutrition, and for researchers who need useful indicators to study the determinants of malnutrition at the household level.