* The collection of the stool samples by Project HOPE was discontinued after the baseline survey due to a low response rate, time, and the lack of sensitivity (3 stool samples are required for a definite prevalence rate). Project HOPE is now using the data that the MINSA has been collecting in conjunction with their anti-parasite campaigns.
Sample sizes were calculated with the following formula:
n = Z2 pq/d2
n= sample size
Z= statistical certainty chosen
p= estimated coverage or prevalence level/rate to be investigated
d= level of precision or sampling error
The statistical certainty was chosen to be 95% (Z=1.96). The value of p was defined as the coverage rate that requires the largest sample size (p=0.5). The value of d depends on the precision or margin of error desired, which for this case was set a 10% (d=0.1). Given the above values, the sample size (n) needed was determined to be:
n= (1.96)2 (0.5 x 0.5) / (0.1)2
n= (3.84) (0.25) /0.01
Due to the fact that it would take a great deal of time to randomly select an identified individual from a survey population, and then perform this selection 96 times (for a sample of n=96) an alternative method of cluster sampling was used. Using this method, a minimum of 30 clusters are selected in which several individuals within each cluster are chosen to reach the required sample size.
In order to compensate for the bias introduced by interviewing persons in cluster rather than as randomly selected individuals, experience has shown, given the values of Z, p, and d above, that an average sample of 300 (10 per cluster) should be used. For this project though, 46 clusters were used due to constant migration and new communities being established in all 3 target areas-especially in REDES El Dorado, and to a lesser extent in REDES Lamas and San Martín. Additionally, changes had to be made due to guerilla activities in some parts of the targeted