Pastoralist populations are notoriously difficult to survey. I know that SQUEAC assessments have been done in pastoralist settings (e.g. Turkana, Somalia, Niger). I am not sure that you identify a problem with SQUEAC. It seems to me that the first SQUEAC told you about coverage in the population as it existed at one time and the second SQUEAC told you about coverage in the population as it existed at another time. The population changed. I am not sure that you can criticise a method for not telling you about coverage in a different (future) population. As with all SQUEAC assessments it pays to find out about how the population is consituted because it is often the case that programs fails to accommodate one or more population sub-groups well. It is important to identify these and to investigate coverage with small surveys / studies. Even if they are not present you could report this and make an informed guess WRT the likely effect on future coverage. Transhumant pastoralist populations move but they do tend to move in predictable ways. I think the lesson here is to know the population. BTW : If you have a hypothesis (such as the one you put forward above) then SQUEAC provides tools to test the hypothesis. I would have done some small studies even after the stage III estimate had been done. I would be wary of stuffing questions into the stage III survey since sample size would probably be too small for any meaningful analysis.

Dear Freddy, I will start by commenting on your last question: is SQUEAC suitable for nomad communities? If I “translate” your question, this could be read as “is CMAM suitable for nomad communities”? This is basically what you are asking. The answer is obviously “yes”, it is!! The objective and the challenge of CMAM are to improve access to service and this applies to all sorts of communities and lifestyles. It is up to the programme to develop strategies to achieve this. Serve nomad’s community is not necessarily easy but it can be done. My suggestion would be that you start by knowing your beneficiaries: who migrates? The entire family or only adults? Do they migrate to the same or different areas (sometime women and children migrate not as far as man)? Where do they migrate? For how long? Do they migrate in group or separate families? Etc. If 40% of the population of the programme area are nomads, I believe it is mandatory for you to understand how do they live, their cultural/social/religious characteristics, etc. You cannot support somebody you don’t know; this is already a sign of the failure of the programme. About the 5% coverage decrease: one of the data that you are supposed to analysis is what is called the “response to the calendars” (diseases, hunger, agricultural work, etc.). In your case, the “migration” calendar should be part of the calendars analysed. If you do this, you would have probably remarked that the programme do not respond to the calendar for this element. This information could have been confirmed in the interview with mothers. In addition, SQUEAC advices to undertake “small studies” or “small surveys” to better understand some elements (e.g. nomads vs sedentary population) and see if what are the difference between the two and if the programme accounts for both populations and identify the barriers. If you would have done this, you may be able to say that you had a 5% decrease in your “prior” because the programme do not respond to this population, but I understand you haven’t done this. Therefore, I imagine that the 5% comes from the “likelihood” part of the equation; in this case, the nomad people will not have influenced your data. Don’t forget that active case finding is a “direct method” which means that when you do active case finding in a village or area, you are supposed to find all SAM children, this will be your denominator, but if the children are not there………………they cannot be in the denominator and not in the nominator (the covered children), therefore it will not affect your likelihood. In my opinion, I believe that in the reality your coverage figure is even lower that the figure you got and this because you don’t seem having taken in account the failure of the programme to cover 40% of the population.

Sampling transhumant pastoralists is difficult but not impossible. The main issue is that they may move in small groups setting up camp for a few weeks at at time before moving on to pastures new. The pastoralist survey method developed by ACF (FEX article [url=http://fex.ennonline.net/35/newmethod]here[/url]) provides a useful way of mapping / sampling pastoralist troupes. Without a method like this your stage III sample will systematically exclude children travelling in troupes (as Lio points out ... this is not a universal practice) and you will not have a pastoralist group for stage II small studies and surveys. The direction and magnitude of any bias is difficult to know. Moving troupes have livestock with access to pasture and tend to be food secure as well as having the advantages of mobility WRT infectious disease whilst "pastoralists" who have lost their livestock and are living in or near permanent villages tend to be food-insecure (they are ex-pastoralists) and, being unused to a sedentary lifestyle, prone to infection. This suggests that pastoralists moving into your area will have few SAM cases and a very marginal effect on coverage (assuming you even sampled the moving pastoralists with any thoroughness). An observation ... Surely you have considerable overlap between the 95% CIs on the two estimates? You have only weak evidence that coverage has dropped.