I think that the cut-points need to be decided by mortaility risk rather than by comparison with W/H since W/H is only weakly associated with mortality and the relationship between W/H and mortality varies between populations. The relationship between MUAC and mortality is pretty stable accross different populations. These review articles: [url]http://tng.brixtonhealth.com/node/15[/url] and: [url]http://tng.brixtonhealth.com/node/18[/url] show the relationship between MUAC and mortality in several popualtions. From these reviews. I would used MUAC < 110 mm and 110 mm <= MUAC < 125 mm for severe and moderate cases respectively (which is what you have).

I forgot to mention that you may want to use MUAC as a first stage screen for admission into programs using W/H for admission. I would caution against doing this unless the W/H measurements and calculations are done in the field. Referring on MUAC to a W/H program can destroy coverage through the "problem of rejected referrals" (see CTC literature for more on this issue). The trick to using MUAC in this way is to go for a cut-point that detects (e.g.) W/H < 80% with high sensitivity. The cut-point should not be too high as this will involve you weighing and measuring a great many children which removes the advantage in using a two-stage screen. The common threshold for this use is 135 mm. If you have data from a nutritional anthropometry survey that collected both MUAC and W/H (yet another reason why you should collect both and SMART / ENA should handle MUAC data) you can fine-tune this threshold. There is a (rather old DOS-based) program available from: [url]http://www.brixtonhealth.com/Screen100Setup.exe[/url] designed for this purpose. The analysis is not difficult ... (1) apply a case-definition for W/H, (2) set a threshold for a MUAC case-definition at 125 mm, (3) apply the MUAC case-definition, (4) create a 2-by-2 table (MUAC case by W/H case) and calculate sensitivity and specificity, (5) increase MUAC threshold by 1 mm and if < 145 mm repeat from step 3 (above), and (6) select the MUAC threshold with high sensitivity and reasonably specificity.

Thank for replying and suggestion sincerely Nicholas

I have heard that WHO/UNICEF will recommend that the cut-off point for SAM be increased to MUAC <115mm, and MAM >=115 to <125 for children from 6 months. Would this be to make sure that more children with an increased risk of mortality are identified earlier?

In short ... YES. It will approximately double case-numbers (assuming good case-finding). There will be some trade-off of sensitivity in favour of specificity. Things are not quite as simple as your question seems to suggest. The case-definition can make a large difference to case-finding exhaustivity. This is one of the reasons that MUAC is now preferred over W/H see: [url]http://tng.brixtonhealth.com/node/15[/url] for a full discussion of this issue. Case-finding / case-detection is, however, more than having a handy case-definition. We can see this in the coverage proportions achieved by OTP-style programs. Programs that rely on extension workers, periodic screening, and passive case-finding usually achieve coverage proportions of between 15% and 30%. CTC-style programs with well-implemented community mobilisation regularly achieve coverage proportions of between 70% and 80%. The point of this is that if you want to find case you also need to have a well-designed and well-implemented strategy for case-finding. It is one thing to identify cases and quite another to recruit them. This requires a proximate, responsive, and hospitable program. Effective case-finding is just one element of program coverage.