When we were working on the original CTC research program which gave rise to CMAM program designs we used W/H in CSAS coverage surveys. We used the active and adaptive (snowball, informant driven) case-finding procedure and then took both W/H and MUAC (at that time MUAC was wrongly considered to be inferior to W/H). Even with the active and adaptive case-finding "filter" we found taking W/H to be slow, expensive, and cumbersome. This was because it is slow and cumbersome and because it takes three people to do W/H well. It was also rather error-prone because of the need to calculate WHZ in the field. We later achieved some savings by adding a MUAC filter. If the suspected case had a MUAC above a given threshold we did not take W/H. The threshold was decided by examining data from local nutritional anthropometry surveys (a DOS-based tool "MUAC screening tool" was developed - you can run it under DOSBOX on Windows - and it is still available from [url=http://www.brixtonhealth.com/software.html]this site[/url]) and picking a high sensitivity and moderate specificity threshold. This did speed up the work so we could sample more villages in a day (needed because SAM is a rare condition). It was still rather expensive, slow, cumbersome, and error prone. My advice is not to bother with W/H as it is a pretty useless measure for detecting children at risk of near term death. Programs using W/H are (IMO) stuck in the past and will (due to the problems with achieving high spatial and temporal coverage of screening) have low coverage. I hope this is of some use.

Dear Mark, Many thanks for your quick reply, much appreciated. Thanks for sharing your CSAS coverage survey experience. I fully agree with your opinion. I share the same view including W/H in coverage assessment will not add much value in survey outcome as you said it will rather slow survey data collection. I’m involved with SQUEAC survey and programme that use W/H as admission criteria along with MUAC, ask this question regularly. Therefore I needed an expert’s opinion to refer to. Thanks again.

Thank you for your kind comments. I am happy to help. I too find myself in this situation. I think that the use of W/H is more "vampirical" than "empirical" (i.e. it cannot be killed by evidence and reason alone). I therefore act "unreasonably" and refuse to use W/H in the coverage survey - we could do it in a research project but it is just not practicable for operational work. I still get asked to do coverage surveys despite being "unreasonable".