For each cluster, 13 households (eight households and five extra households) were randomly selected for each survey collection round. The extra five households were visited only if one of the first eight houses was found vacant, the inhabitants refused further study, the household was not found, or the household was not visited because of a bad road. Four consecutive collection rounds in year 1 and five collection rounds in year 2 were done. A fifth round of collections was added in year 2 and year 3 to compensate time when collections were suspended (March and May) because of the SARS-CoV-2 pandemic. LLIN=long-lasting insecticidal net.