Post a reply: How to Maintain the Quality of CSAS Method

As we all know, coverage is an internationally recognized indicator of performance for selective feeding program (SFP) and CSAS is the current gold standard for assessing coverage of SFP. Since the information from the coverage survey help implementing organizations to strengthen programme performance and impact as well as for better future programming, I believe that the quality of the assessment procedure is the important facet of CSAS method. With this in mind, I have few concerns regarding CSAS procedures during field data collection

1. I came across with coverage reports that validation of case finding (capture-recapture study) was done by Active Case Finding vs. Central Location Screening, while the sensitivity of central screening is assumed to be low. This exercise will apparently gives a high estimate of sensitivity for case finding and subsequently over or under estimate the program coverage results. The capture-recapture technique guesses that house-to-house screening has higher sensitivity (90-100%) and similarly the field trial results in Malawi (Myatt et al, 2005) indicated that house-to-house method is more efficient than others. My question:- Is it acceptable to conduct validation study for case finding using central location screening while it yields less sensitivity than active case finding procedure?
2. Is it acceptable to take the typical sample size for CSAS (80 cases) in areas, where it is difficult to get updated malnutrition rates?
3. What is the minimum requirement of the area to be covered for the coverage survey in comparison with the total area of the district? For example, if the area covered by the survey (No. of quadrats X Area of each quadrat) is 75% of the total district area, is this design acceptable?
4. Is there any recommended way to decide the number of villages to be visited in each quadrat during case finding when the number of villages estimated in the sample size calculation is higher than the total number of quadrats? For example, if we get 30 quadrats during mapping, and we require visiting 35 villages, how can we decide the number of villages within each quadrat? Based on the number of villages in each quadrat; Population size; Physical content???
5. One of the attributes of CSAS as a replacement for the outmoded 30X30 method is that the active case finding is rapid and 6 communities can be sampled by one team in a day (http://www.fantaproject.org/ctc/myattPPT1.pdf). However, I believe that in most districts where SFP was established/operated, we expect to get more cases and/or covered (As rationale to establish SFP is high GAM & SAM). Therefore, if we request the team to cover 6 villages a day, we will compromise the exhaustiveness of finding cases (difficult to maintain the principle 'find all or nearly all') and likely to miss few cases/covered and as a result under or over estimate program coverage. Though, it helps to reduce cost of assessment, is it wise to plan for a team to cover 6 villages a day?

Hope to hear an expert opinion