Effect of sensitivity on the Type I error when treatment decisions are based on the cluster-specific results, assuming 30 clusters are surveyed and 100 children/cluster.

A specificity of 99.8% was assumed. The critical cutoff was ≥ 2 positive tests in at least 1 village. The calculations were made for a true prevalence of 1%. At least 89% sensitivity is required to have a Type I error of < 5%.