Use and misuse of stunting as a measure of child health

The term “stunting” has become pervasive in international nutrition and child health research, programme and policy circles. Although originally intended as a population-level statistical indicator of children’s social and economic deprivation, the conventional anthropometric definition of stunting (height-for-age z score <-2SD) is now widely used to define chronic malnutrition. Epidemiologists often portray stunting as a disease, making inferences about the causes of growth faltering based on comparisons between stunted (i.e. undernourished) and non-stunted children, and stunting is commonly used to monitor public health and nutrition programme effectiveness.

However, there is no biological basis for the -2SD cut-off to define stunting, making it a poor individual-level classifier of malnutrition or disease. In fact, in many low- and middle-income countries (LMICs), children above and below the threshold are similarly affected by growth-limiting exposures. The authors of this paper argue that the common use of stunting as an indicator of child linear growth has contributed to unsubstantiated assumptions about the biological mechanisms underlying linear growth impairment in LMICs and has led to a systematic underestimation of the burden of linear growth deficits among children in low-resource settings. Moreover, because nutrition-specific, short-term public health interventions may result in relatively minor changes in child height, the use of stunting prevalence to monitor health or nutrition programme effectiveness may be inappropriate.

In commenting on this paper in his “Development Horizons” blog², Lawrence Haddad emphasises that there is nothing inherently wrong with being short unless it is due to growth faltering; the key limitation of stunting as a summary indicator of lost human capital is that it ignores any growth faltering that occurs elsewhere along the height-for-age distribution, not just <-2SD HAZ; hence the magnitude of the problem being understated using current definitions.

Lawrence suggests that we should look at mean and standard deviation of HAZ in addition to stunting and at the slope of change in HAZ by age. In countries with evidence of growth faltering, rightward shifts of the entire HAZ distribution will reflect positive improvement in growth of all children. In addition, programme investors must recognise that the prevention of growth faltering requires many factors over a significant time period and that the right programme indicators are those that are achievable but that have a clear path towards ultimate outcomes, such as improved growth.