Dear Anonymous, I can see a couple of problems with this Survey assuming it was conducted within the same population group: The first one, is either the first one or the current one has DESIGN, SAMPLING or ANTHROPOMETRIC MEASUREMENT ERRORS at population level. Data from multiple low and middle income countries on stunting over time show us that usually countries doing well on stunting reduction programmes register at best a 15-20% reduction within 3-5 years time. The second one, is there could have been a massive population movement or demographic change in the your survey area since the first survey you are comparing to. However i would be very interested to know where and when the two surveys were conducted, what methodology was employed, how the survey was managed and so on. That would at least give us an idea of where the problem occurred. It is also good to do further statistical tests to check on the statistical validity of the anthropometric data for your survey.

It would be useful if you told us the reduction was within which period of time is it 1,2,3..... for us to be able to judge and see otherwise I hope the data you are using to compare your results with also were within the same population and analysis was done in the same way as the previous finding. But that reduction is possible given time.

Were the two surveys carried out at the same time of the year ? There may be substantial seasonal variations in the prevalence of stunting, as growth in height is seasonal and curiously does not take place at the same time as growth in weight. See: Brown KH, Black RE, Becker S. Seasonal changes in nutritional status and the prevalence of malnutrition in a longitudinal study of young children in rural Bangladesh.Am J Clin Nutr. 1982 Aug;36(2):303-13. Maleta K, Virtanen SM, Espo M, Kulmala T, Ashorn P. Seasonality of growth and the relationship between weight and height gain in children under three years of age in rural Malawi. Acta Paediatr. 2003 Apr;92(4):491-7.

Hi all, The first survey was carried out in May 2014 and the other one in May 2015 (Same season of the year) in the same geographical area (one of the semi arid county in Kenya). The survey was a cross sectional integrated nutrition survey using SMART methodology. ENA for SMART software was used to analyze the data. The procedures, supervision and quality checks were the same for both surveys. Thank you

I am assuming that by "stunting" you are referring to both the process (stunting) and outcome (stuntedness). If you are only referring to the process then a drop like this is possible. Ordinarily we would not see such a drop in prevalence of both stunting and stuntedness. This is because the children aged between about 2 and 5 years who are already stunted will usually remain stunted, tracking but not approaching the H/A reference median, until the pre-adolescent growth spurt. This means that prevalence of stuntedness cannot change very rapidly. I think that you may have a problem with comparable populations. Semi-arid land tends to be associated with pastoralism (livestock are a way of turning scrubby vegetation into meat, fat, milk, skins, &c.). Pastoralism, even here in Wales, tends to have an element of transhumance (seasonally patterned migration). This may be the case in your survey area. Note that seasonality is not calendar seasonality, A one or two month variation is not uncommon - local weather dominates. It could be that in one year your sample was dominated by non-pastoralists and on another year the pastoralists arrived earlier and gave you a different population. Just my tuppence.

Thanks Mark. there was a bit migration especially in the pastoral communities of the North. The area has 3 livelihood zones; pastoral, agropastoral and mixed farming. Thanks also for Andre and other contributors

Well this explains why you have this kind of data. In your region (East Africa) there are a lot of pastoralist communities which point to the case being as i pointed out earlier. This also means a potential change in the demographic characteristics of the community. Also Mark's explanation makes more sense too. Given that the survey area is the semi Arid part of Northen Kenya dominated by pastoralists and agro pastoralists following the river lines there is high tendency for mixed population groups from influx or out/in migration. This might have resulted in the variation. I would imagine, in this case, there was high in-migration of pastorialist communities earlier than much earlier than last year due to seasonal changes (early onset of rains in the area may be). This is particularly not uncommon in the region, a lot of the borderline areas suffer from mixed populations that are constantly mobile there by making it difficult to have comparable surveys to monitor such nutrition trends among the population. One suggestion would be to identify the population nature as to the migration issues earlier and adapt your methodology to the same. There are more appropriate sampling methodologies which can be used in combination with the SMART methodology that capture such issues. UNHCR East Africa regional office might have a good experience in dealing with such cases as they predominantly work in such communities. As for the result I would suggest you consult other similar surveys in the region and/or in Kenya and see how the trend goes. I am sure they would give you a good idea of the trend. But i this region this level of stunting reduction has not been registered so far within such time gaps (1-3 years). It is far too unrealistic if we are considering the same population groups.

As you may have noticed ... I am not a fan of the WGS approach. I remain unconvinced that the absence of the (bourgeois) ideal equates to a pathology. This may be how a bourgeois may see it. The WGS does have some value as a yardstick. We may never see the population median H/A reach the reference median but we should, if we have avoidable stunting and effective programming, see it approach the reference median. Used this way the WGS does allow us to see progress. I think we need to be clear about what we mean by "normal". It we were (e.g.) to build a reference from a population in which avoidable stuntedness were common then we risk seeing avoidable stuntedness as something unavoidable. A good reference probably needs to be criteria-referenced but, I think, less strictly so than the WGS. The older NCHS reference was useful in the sense that it was not a censorious "standard" but a representative sample of children living in a country with conditions predominantly favourable to growth. WRT educational perfomance of stunted children. A lot of work has been done showing performance deficits. Much of this work suffers from a class bias (i.e. children from poorer background, regardless of anthropometry, tend to do poorly in curricula designed for children from wealthier backgrounds and stunted children are usually from poorer backgrounds) and inadequate control for confounding (e.g. as above with socio-economic status as a "grand confounder" but with more specificity so that (e.g.) poverty is associated with lack of electric light which reduces the ability of the child to study at home and poverty and stunting are associated as are parental education, access to pre-school education, disposable income to spend on toys and books, leisure time for study, dietary diversity, meal frequency, school absenteeism due to illness or lack of money to pay fees ... and so-on). Must studies appear to be both shallow and narrow. I think we can and should, in the short to medium term, concentrate on broader universal programming such as deworming and school-feeding as well as (of course) universal provision of quality education.

Oops ... my last message (above) was intended for [url=http://www.en-net.org/question/1755.aspx]this discussion[/url]. Season migration patterns may be tracked by local ministries of agriculture, WFP, FAO, or specific projects. You can often get this information by asking in livestock markets since pastoralism tends to require strong information networks.

From Eucabeth Onyango: Hi, looking at the feedback on the Livelihoods of the communities,maybe its could be too unrealistic for the results in the reduction of stunting rates for the communities. One: how was the intervention meant to manage the infiltration from the communities into the intervention areas. Two: the ages of the targeted populations during the surveys is quite a priority for you to realize the results you have before you today. Thirdly,the reliability and accuracy of the data must be exceptionally worrying based on the results generated from them.