evidence-based blog of Filippo Dibari

Middle-upper arm circumference (MUAC) for nutritional surveillance in crisis-affected populations: development of a method

In Under-nutrition on April 30, 2012 at 3:43 pm

This is the PhD proposal of Séverine Frison (London School of Hygiene and Tropical Medicine). She undertakes her research together with Francesco Checchi and  Claudine Prudhon

Summary – Timely, sensitive, population-representative nutritional surveillance is crucial to detect nutritional emergencies in crisis-affected populations, and responding appropriately. Current approaches to nutritional surveillance mostly rely on regular anthropometric surveys, the results of which are interpreted alongside other crisis indicators such as mortality. However, there is little evidence on which to predicate the design of surveillance systems, and a variety of methods are employed in the field, with questionable impact, partly due to the infrequent nature of surveys.

Middle-upper arm circumference (MUAC) is known as a good predictor of mortality in children, and is increasingly adopted as the criterion for screening and admission to treatment programmes. Furthermore, MUAC is easier to measure than weight for height (WHZ), and may be more sensitive to changes in nutritional status. Despite this, it has not been used prominently in surveillance to date.

Here, we propose to develop a new method for nutritional surveillance in crisis-affected populations, based on measurement of the mean MUAC or the mean MUAC-for-age. Estimating the mean of these indices would entail lower sample size requirements than for prevalence surveys, improve the feasibility of data collection on the field, and allow for greater frequency and spatial resolution of surveillance. While mean MUAC trends could be interpreted separately, we wish here to study a potential method to infer GAM or SAM prevalence from the mean, thereby enabling quantification of programme needs. The proposed method is based on an assumption of normality and prior information about the population standard distribution of these indices, and relies on the ability of MUAC to capture oedema (kwashiorkor) cases. These assumptions need to be explored thoroughly.

The present project’s objectives are therefore to:

  1. Assemble a large dataset of surveys from a variety of settings that can be used to explore the statistical assumptions underlying the proposed method;
  2. Identify appropriate geographic strata into which to classify the surveys, and, more generally, regions of the world where nutritional surveys are undertaken, such that the variability of the SD of MUAC and MUAC-for-age within any given geographic stratum is minimised;
  3. Examine the association between MUAC or MUAC-for-age and oedema (as a sign of kwashiorkor), so as to investigate whether MUAC-based cut-offs for GAM or SAM capture oedema cases, and, if not, whether corrections can be applied to estimates based on MUAC alone in order to account for the prevalence of oedema;
  4. Examine the normality of MUAC and MUAC-for-age distributions at the population level and in small samples, and if necessary apply transformations to the data in order to achieve normality;
  5. Quantify and describe the SD of MUAC and MUAC-for-age across all surveys and within geographic strata, and assess how variability in SD would affect the precision of the proposed method;
  6. Investigate empirically the feasibility of using SDs of MUAC and MUAC-for-age from small sample size surveys directly by testing the stability of the SDs using a bootstrap method.
  7. Compare the appropriateness of MUAC versus MUAC-for-age cut-offs, by considering the degree of precision expected with either index if the proposed method is applied, as well as the relevance of either index for field operations.

We propose to accomplish the above objectives through extensive data analysis of at least 500 previously performed surveys from various areas of the world, livelihood zones and body shape strata. If successful, we envisage a second phase of development, consisting of defining sampling designs and sample size requirements for the proposed method to infer prevalence of GAM and SAM based on mean MUAC or MUAC-for-age.

The project will last 12 months, and is a collaboration between the London School of Hygiene and Tropical Medicine and the Health and Nutrition Tracking Service.

(Severine gave me the “green light” to publish this here)


– – –

NB: You wish to follow up this or other topics from this blog? Type your email in the rectangle at the bottom/right side of this page.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: