Could stunting be a contributing factor to the differences? Am equally interested to learn more from the experts on this issue.

At the ENN, we have become increasingly aware that these sort of programming issues around MUAC and WHZ are a real challenge for those working in the field. We are currently exploring a pragmatic and urgent review of the situation, approached by a few agencies to review evidence and negotiate understanding around these. The aim would be to provide some evidence-informed 'stop gap' guidance options for exactly the sorts of scenario you are outlining. We will keep the en-net forum informed of the process once up and running and most certainly of the findings, that will also feature in Field Exchange.

WHO is currently finalising systematic reviews on MUAC and wt-for-height with the Tropical Institute in Antwerp (for admission and for discharge from programmes on SAM and MAM). They have used these systematic reviews and the discussion held at the WHO Nutrition Guidance Expert Advisory Group (NUGAG) meeting earlier this year to update the guidance of the joint statement. They will be sending out a call for comments on the updated guidance on SAM (including admission/ discharge criteria) before the end of the month.

@ Marie: It is indeed very concerning and worrying. The challenge is we use WHZ to determine cut-off and severity of the situation and we use (In most cases) MUAC to admit children. And this creates a challenges. One of the reason for a consistent low prevalences in MUAC based proxy-GAM against WHZ is there is a continuous community screening using MUAC and If you do a survey in these areas the prevalence of WHZ is higher and that of MUAC is comparatively lower. @Bekka: There is a difference in the two prevalences but the difference is very much exaggeration in some communities and the differences are narrow in other places. But we should always consider what is going on regarding community screeningt before comparing the two prevalences. What is the tool for community screening (admission)? I think it affects the results. Anyways a help that address this issue is very much useful.

Thanks all for your reply. @Marie: I’ve seen a couple of field exchange articles on the subject from Guatemala and Philippines and the use of both MUAC and W/H for admission seems a fair compromise until the influence of body shape, morbidity and other factors on the W/H and MUAC relationship fully understood. So, additional evidence on field exchange will help us to understand more. @Florence: Stunting could also affect the relationship, but from some data set’s i’ve noticed that whenever the level of stunting is very high (>55%), the prevalence of MUAC is closer to W/H (although not sure about the exact explanation). @Melaku: community screening for MUAC in an intervention area could also be one interesting explanation for low MUAC prevalence assuming the screened children by MUAC have been included in the programme (so well nourished?). In a discussion with colleagues, it was also mentioned that due to the time consuming nature of W/H measurement in community surveys (note that SMART recommends as minimum of 15 HH per team per cluster in a day even in a challenging environment), in a topographically challenging set-up, teams might take biased measurement in order to complete on time in comparison to MUAC which is easy to take (and hence accurate measurement?). However, despite MUAC being an easy measurement, during standardization tests the teams accuracy and precision is somewhere similar with that of W/H. So, it would also interesting to know the sensitivity and specificity of MUAC measurements in comparison with W/H. As Tamsin outlined, the WHO review could give us an evidence based explanation in to this

Would it be possible to get the reference where '...note that SMART recommends as minimum of 15 HH per team per cluster in a day even in a challenging environment..'? Thank you.

SMART Methodology. Version 1 (April 2006). Page 56 “In practice, even under difficult conditions, teams can manage at least 15 households per day”

Does it mean you need to have at least 15 HH per cluster? I am not sure. I know the discussion topic is totally different but could anyone provide some advise? I am aware of the minimum no. of clusters you need to have in a SMART survey but minimum number of HH is new to me. I have done surveys with less than 15 HH per cluster and am worried now.

In my understanding, after we conduct SMART survey, we expected that the results will lead to response because there is little point of doing a survey if we already know a response will not be possible or at least the results should be used in advocating for a timely response by other partners. So, we are expected to implement and report the results in a timely manner for decision making (usually a week for data collection in my experience). The most challenging survey (due to the pastoralist nature of the community, difficult road, etc) i remember was in somali region of Ethiopia, but we have planned 15 HH/cluster/day and it was successful. So, I would be interested to know how many households you have surveyed in one cluster (if

The surveys were conducted in similar contexts described above. We aso did household listing and used simple random sampling to selct households. So, the amount o time we had onthe ground was further limited. We were only able to cover about 12-13 households a day. 7 teams were used for 7 days to collect data; 1 cluster/day. Thank you for posting the discussion on the SMART forum.

Hi Beka, This is just to respond one of the question you raised i.e "Is there any survey that shows MUAC prevalence is higher than WHZ?" In Northern Nigeria (Eight Staes), UNICEF conducted 30 surveys between August 2010 and March 2012. Of these, the MUAC prevalence was higher in 12 of them. However, in all surveys, except one survey (Zamfara Feb/March) there was no significant statistical difference between MUAC and WHZ. With regard to Zamfara, the MUAC prevalence was 11.7 (9.3 - 14.6 95% CI) whereas WHZ was 6.7% (5.1-8.8 95% CI).

Addressing the original questions (I have just got back from the field) ... As for "the prevalence of SAM based on WHZ <-3 SD of the WHO standards and those based on a MUAC cut-off of 115 mm are very SIMILAR" ... this is true in some contexts but not in others. W/H is strongly influenced by body shape. This means (in general) that MUAC prevalence will tend to be higher than W/H prevalence in cold climates and at high altitude. The opposite will tend to be the case in hot climates at low altitudes. Milk consumption is another issue. I look at some of these issues in [url=http://www.brixtonhealth.com/MyattBodyShape.pdf]this article[/url]. I have something on using MUAC for case-detection and admission [url=http://www.brixtonhealth.com/FNB_27_3_2006_B.pdf]here[/url]. I hope this is of some use.