I do not have much data on admitting WHZ in CMAM programs because I do not tend to work with programs that use W/H (for many reasons but mainly because it limits delivery options and can kill coverage). I have checked and found a dataset for all (963) admission to an NGO-run CTC program running from from September 2003 - April 2004 in the Ethiopian Highlands. Admission was on MUAC (< 110 mm) or W/H (< 70% WHM by NCHS) or oedema. Using the current WGS reference I find about 16.1% had a WHZ <= -4 (this is with a few records with WHZ < -5 censored). Now that admitting MUAC has increased to < 115 mm we would expect that proportion to be much lower (i.e. < 8%) in a program with good coverage. As for mortality, the relationship between WHZ and mortality is complicated by body shape and other factors that vary from place to place. Results from cohort studies will, therefore, be of little use unless your program is in the area where the cohort study was done. The general finding is that is (at these levels of WHZ) mortality risk increases with decreasing WHZ so that a child with a WHZ = -3.9 will likely have a worse prognosis than one with a WHZ = -3.1. So ... it will be quite high. I hope this helps.

Thank you Mark. A related question: How do you establish the proportion of children <-4 when MUAC is raised to 115mm (you say 8.1%)? The relationship between prevalences of SAM according to WHZ and MUAC is tricky to generalize on, no?

It is a bit of a guess ... We know that moving from MUAC < 110 to MUAC < 115 will increase the number of eligible children by about 2 - 3 times. I did an analysis of this (for WHO/UNICEF/WFP) some years ago to investigate the implications of raising the MUAC threshold using a databases of 560 nutrition surveys and it the effect was to increase case numbers between 2 - 3 times. If we assume MUAC is approximately normally distributed then we can do a simple simulation with something like: =NORMDIST(115,145,15)/NORMDIST(110,145,15) in a spreadsheet. This gives 2.32 (i.e. the eligible population at MUAC < 115 mm is 2.32 times bigger than the eligible population at MUAC < 110 mm). The new eligible population is the old eligible population (MUAC < 110 mm) with the addition of the children with MUAC = 110 mm, 111 mm, 112 mm, 113 mm, or 114 mm. It is not unreasonable to assume that all of these additional children with have WHZ > -4. With good coverage we can expect to double our program numbers with about half now above with MUAC > 110 mm and WHZ >= -4. The proportion with WHZ < -4 will, therefore, be about 16.1 / 2.32 = 6.94%. It is a rough and ready calculation. You should check my logic.

Forgot to add ... if coverage is good then we will have early treatment seeking. this will means that we will have very few very severe cases after the first few months of program operation. In this case the proportion with WHZ < -4 should be very small.