Data Analyses
A total sample of 803 mother newborn dyad was recruited in the study. The outcome variable of interest was GBS colonization among pregnant women delivering term or near term infants 35+ weeks gestational age. The primary independent variable of interest was active tobacco smoking throughout pregnancy as determined by the number of cigarettes smoked per day. In the univariate analysis, tobacco smoking was categorized as 0 cigarette/day, 1 cigarette/day, 2 cigarettes/day, and 3 + cigarettes/day respectively for trend analysis. Test of trend was also explored for maternal age. Initially maternal age was empirically categorized into 2 groups (< 20 year of years versus 20+ years). A trend analysis for this variable was then conducted to determine whether GBS rates varied by the various maternal age groups. When a clear trend emerged, maternal age was subsequently empirically expanded to 5 groups (<20, 20-24, 25-29, 30-34, and 35+) to explore whether the trend continued. Tests of trends were also explored for the other independent variables of interest (birthweight, gestational age) but were not significant so these variables were left dichotomized in the univariate analysis. In the multiple logistic regression analyses, all the independent variables of interest were left dichotomized for simplicity of analyses.
Descriptive statistics were conducted. Univariate analysis was conducted to determine if each of the selected variables was associated with maternal GBS colonization using the X2 analysis of comparisons of proportions with Fischer’s exact test. Next, a multiple logistic regression analysis was conducted using only the significant variables in the univariate analysis to determine if each of these factors was still predictive of maternal GBS colonization after controlling for the other potential confounders. The SPSS statistical software (IBM, Armonk, NY 2018) was used in the analyses. A p value of <0.05 was used as test of significance in all cases.