Sample Selection in Managua

To carry out the Camino Verde trial baseline measurement, we randomly selected 60 neighbourhoods from a total of 187 in the municipality of Managua. These 60 “measurement neighborhoods” contain a population of some 400,000 inhabitants. .

We drew this sample from the data accompanying a recent piece of municipal legislation (Ordenanza No. 03-2009 del Consejo Municipal, publicada en LA GACETA, Diario Oficial, el 28 de Octubre de 2009). which subdivided Managua into 7 districts. The first step was to identify each neighbourhood’s population in 2010, the district to which it belonged and the neighbourhood’s level of social vulnerability — more vulnerable or less vulnerable according to the municipality’s own criteria. This gave us 14 segments (seven districts by two vulnerability levels). The total population of the segments amounted to 1,038,782 residents, equivalent to 89% of Managua’s total population. The remaining 11% resided in sites where the trial implementation was less feasible: some sites were in suburban areas too insecure for easy access by our facilitators and others were higher income residential neighbourhoods.

The segments’ specific population weight determined the number of sites per segment. For instance, in the attached chart, the higher vulnerability population of district 3 constituted 8% of the total population, which led to it being assigned five of the 60 neighbourhoods in the overall sample.

Randomisation

Once the baseline measurement was complete, we randomly assigned the 60 neighbourhoods to intervention or control groups. The randomisation was stratified according to three key factors for which binary classifications were taken from the baseline measurement results: higher or lower degree of neighborhood organisation; presence or absence of larvas or pupas in household water receptacles; and evidence, or none, of recent dengue infection based on the paired saliva sample results.