Statistical analysis
Follow-up was initially planned for 5 years but later extended to 8 to
allow more reliable diagnosis of asthma and testing of airways function
and inflammation. To measure effects of geohelminths on asthma
prevalence with >80% power at significance level of 0.05,
we estimated that we would need to follow-up 1,725 children to detect a
difference in asthma prevalence of ≥6% with infection rates of 50%
among mothers and 35% among children. Primary exposures were maternal
and childhood geohelminth infections and primary outcomes were wheeze,
asthma, and SPT to any allergen. Exploratory analyses addressed the
effects of geohelminth species and infection intensities on primary
outcomes, and effects of geohelminths on airways reactivity, and markers
of airways inflammation. Univariable and multivariable logistic
regression were used to estimate associations. Potential confounders to
be considered in the analyses are shown in Table 1. Urban-rural
residence was defined by geographic boundaries. A socio-economic status
(SES) index was created using principal components analysis of 7
socio-economic variables as described.15 A
conservative analytic approach was used for all adjusted analyses in
which potential confounders included were those with P<0.05 in
univariable analyses for any of the primary outcomes. Trends across
infection intensity categories were evaluated using Cochran–Armitage
test for trend. All statistical analyses were done using Stata
11 (Statacorp, College Station, Tex).