You need to be clearer about what you mean by "LQAS". LQAS classifies (e.g. into low, moderate, high prevalence) rather than estimates (i.e. give a point estimate with a confidence interval) and will typically use small sample sizes (e.g. SLEAC uses a sample size of n = 40 ... less for small SDUs or in low prevalence settings). LQAS is (when done properly) robust, accurate, and useful. LQAS is used in many fields. In nutrition we use it in ... SQUEAC : LQAS is use to test hypotheses about (e.g.) the spatial distribution of program coverage. SLEAC : LQAS is use to classify program coverage at the level of the service delivery unit (SDU). GAM : LQAS is used to classify prevalence over wide-areas (e.g. health districts). M&E : The original use of LQAS was industrial quality control. This means that LQAS is well-suited to program M&E. The use of LQAS for GAM has not caught on. I have not used LQAS for this purpose because I do not think that it offers significant cost savings over methods such as SMART. This is because the bulk of the costs in cluster sampling are in sampling clusters. Once you arrive at a cluster it is only a little more expensive to sample 30 children than it is to sample 3 children. The LQAS method proposed for GAM estimation uses a similar number of clusters to SMART (sometimes more). This means that the cost savings "promised" by the small overall sample size do not translate into significant cost-savings. Since LQAS may cost as much as SMART but yield less information it seems sensible to use something like SMART for this application. It should be noted that LQAS may be very much cheaper than SMART in (e.g.) camp settings where cluster sampling costs are low. I hope this is of some use.