hello, We observed the same phenomena in Somali region of Ethiopia in 2000: we opened a project for nutritional emergency and IDPs due to drought with very high mortality and SAM. After few months of intervention, the MAM ratio versus SAM increased due to decrease of SAM. this was correlated also with a decrease of mortality. I think different factors explained that we could decrease SAM but not MAM: 1) context of chronic food insecurity 2) lower coverage of MAM cases because active screening done with MUAC, not detecting a large number of MAM in W/H: at the opening of the program we had a very good coverage because the nutrition program was one of the only way to get food ration, so everybody was trying to be admitted in the program, but when food availability improved, spontaneous admissions of MAM decreased. Hope this help

Some hand-waving ... You may like to think of this in terms of (1) fat-tails and thin-tails in the distribution of MUAC or WHZ and (2) the distribution of MUAC and WHZ being a "finite mixture" (see [url=http://www.en-net.org.uk/question/747.aspx]this thread[/url]. In this "model" we could interpret a reducing SAM:MAM ratio as a change in one of the distributions (e.g. by treatment, reduction of risk, or death). Any help?