Figure 3.

Study enrolment strategy. Health and demographic surveillance system census data were used to select eligible compounds based on the number and age of occupants. Eligible compounds were enrolled and groups of three enrolled compounds were combined with three non-enrolled compounds, generating 60 groups. Within each group the three enrolled compounds were randomly assigned to each of the three study arms. This strategy is used to ensure (1) An even spread or compounds in each arm over the study area. (2) That ≤50% of all compounds in the study area are enrolled in the study, so overall transmission intensity remains unaffected.