MUAC-for-age more useful than absolute MUAC for nutritional surveillance in Somalia: results from nineteen cross-sectional surveys (2007-2016).

BACKGROUND

Somalia is affected by a civil war and a protracted humanitarian crisis for more than two decades. The international community has put in place nutrition surveillance systems to monitor the situation and inform decisions. However, the indicators commonly used to identify acute malnutrition, weight-for-height Z-score (WHZ) and mid upper arm circumference (MUAC), do not always converge in their estimations of acute malnutrition, creating challenges for decision making. Furthermore, the divergences are not consistent across livelihood populations within the country. We explored the MUAC-for-age Z-score (MUACAZ) as an alternative indicator in Somalia to minimize the discrepancy.

METHODS

We analyzed data from nineteen cross-sectional surveys conducted in Somalia between 2007 and 2016. We compared the acute malnutrition prevalence estimates by each of the indicators and the degree of overlap in the individual diagnosis of acute malnutrition between the WHZ and the MUAC-based indicators. We performed multivariate regression analysis with sex, age and stunting as independent variables and acute malnutrition as the dependent outcome, defined by WHZ, MUAC or MUACAZ. We performed all the analysis in the population overall and in each of the livelihood populations separately.

RESULTS

A total 255,623 measurements of children 6-59 months of age were analyzed. The overall prevalence of global acute malnutrition by MUACAZ (15.8%) was similar to the one obtained using WHZ (16%), whereas prevalence based on MUAC was much lower (7.8%). These patterns of divergence were sustained throughout the nineteen surveys and the livelihoods studied, with only few exceptions. However, the proportion of overlap in the individual diagnosis of children as acutely malnourished was low between WHZ and absolute MUAC diagnosis (18.1%) and also between WHZ and MUACAZ (28.3%). Results show that age, sex and stunting status of the child affected the likelihood of being diagnosed as acutely malnourished to varying degrees, depending on the indicator used.

CONCLUSIONS

The MUAC-for-age (MUACZ) indicator yielded acute malnutrition prevalence estimates convergent with those obtained by WHZ indicator. However, the degree of overlap between these two indicators for individual diagnosis of acute malnutrition is low. Further studies of MUACAZ as an alternative indicator for nutrition surveillance are needed.