|
Technical expert
10 Dec 2009, 12:48
|
The situations in which (non-elderly) adults are at greater risk of malnutrition than children is very rare. The 6-59 month age group was selected as a "canary in a coal mine" or "early warning" sub-population because of this. The only time I have been forced to focus on the adult population was in Somalia in the early 1990s when the situation was very "advanced" and huge numbers of children had died and there were (mostly) only adults left to treat. There are programs that are targeted at adults outside of severe emergencies (e.g. SFP for pregnant and lactating women) and surveys assessing need will be useful in this context. You should concentrate on the global estimate since, unless you have huge sample sizes, you will have poor precision on a low prevalence indicator such as prevalence of SAM. It is common practice to use women of childbearing age when collecting this data. I suppose you could make arguments about early warning, prevention of low birth weight, targeting at risk households, age-dependence, &c. but there is also a good practicable point that you can get women of childbearing age in a standard nutritional survey (the commonest carer present when you measure the child). You may need to increase sample sizes a little to get enough women, BMI is problematic since it is affected strongly by body shape, oedema, hydration, water retention, and time of day (height reduces during the day, weight goes up after meals and down after defecation). If you use BMI then you should correct for body shape. BMI is quite difficult to collect in surveys (needs different scales and height boards than for children). It is difficult to collect in the old, the disabled, and the weak. It has no meaning in pregancy. The standard MUAC thresholds < 210 and < 185 for moderate and severe. These are used as admission criteria for nutritional support programs for pregnant and lactating women, people living with AIDS, and the chronically sick. These thresholds (as with MUAC in children) are based on risk of negative outcomes. The most evidence available is for risk of low birth weight when used in second and third trimesters (I suppose that first trimester pregnancy is not so easy to detect and you get a lot of loss-to-follow-up as mothers move to their home villages and due to miscarriage). I disagree with Mike. I find the distinction between CED and AED to be useful ways of thinking about malnutrition (akin to wasting and stunting in children but adults can't stunt very much) and there is some epidemiological support for the utility of these concepts from mid-20th century labour and death camps. The two ideas point to, admittedly not well separated, constellations of aetiologies. I have written (with Collins and Duffield) that, just as with stunting and wasting, CED and AED are identified by different indicators (BMI and MUAC) so this should not come as a surprise. The quoted (i.e. by Mike) thresholds seem to me to be very restrictive. I wonder about the evidence-base used to collect them. Some of these thresholds, following Ferro Luzi & James, were developed by regressing MUAC against BMI and selecting a MUAC threshold that corresponded to a BMI threshold which is believed to be associated with a negative outcome (as long as the subject is not old, an athlete, long-limbed, or drank a lot of milk while they were growing up). The problems with BMI means that such thresholds are unlikley to be universal. Just my tuppence. |