"No weight for height" case detection strategies for therapeutic feeding programs: sensitivity to acute malnutrition and target composition based on representative surveys in humanitarian settings.

BACKGROUND

One newly proposed approach to determining eligibility of children aged 6-59 months for therapeutic feeding programs (TFPs) is to use mid-upper arm circumference (MUAC) < 115 mm, bilateral oedema or Weight-for-Age Z-score (WAZ) < - 3 as admission criteria (MUAC+SWAZ). We explored potential consequences of this approach on the eligibility for treatment, as compared with the existing WHO normative guidance. We also compared sensitivity and specificity parameters of this approach for detecting wasted children to the previously described "Expanded MUAC" approach.

METHODS

We analyzed data from 558 population representative cross-sectional cluster surveys conducted since 2007. We retrieved all children classified as severe acute malnutrition (SAM), moderate acute malnutrition (MAM), and those who are both wasted and stunted (WA + ST), and calculated proportions of previously eligible children who would now be excluded from treatment, as well as proportions of non-malnourished children among those who would become eligible. We also analyzed the expected changes in the number and demographics (sex, age) of the selected populations of children according to the different admission approaches.

RESULTS

Both MUAC+SWAZ and Expanded MUAC case detection approaches substantially increase the sensitivity in detecting SAM, as compared to an approach which restricts detection of SAM cases to MUAC< 115 mm and oedema. Improved sensitivity however is attained at the expense of specificity and would require a very large increase of the size of TFPs, while still missing a non-negligible proportion (20-25%) of the SAM caseload. While our results confirm the sensitivity of the MUAC+SWAZ case detection approach in detecting WA + ST (over 80%), they show, on the other hand, that about half of the additional target detected by using SWAZ criterion will be neither SAM nor WA + ST.

CONCLUSIONS

These results suggest that recently promoted approaches to case detection inflate TFPs' targets through the allocation of treatment to large numbers of children who have not been shown to require this type of support, including a significant proportion of non-acutely malnourished children in the MUAC+SWAZ approach. Considering the scarcity of resources for the implementation of TFPs, the rationale of abandoning the use of WHZ and of these alternative case detection strategies need to be critically reviewed.