2.4 Statistical analysis
We used absolute standardized differences (ASD) to measure the effect size of baseline covariates between groups. When the ASD was <0.1, the characteristics between two groups were considered similar. In addition, Kaplan-Meier analysis was used to calculate the cumulative incidence of AD and a log-rank test was used to measure the significance between groups. We used a Cox proportional hazards model to estimate the associations of maternal antibiotic use during pregnancy with childhood AD and associated covariates, which were reported as crude and adjusted hazard ratios (HR and aHR, respectively) and 95% confidence intervals. To consider the effect of infant antibiotic/acetaminophen use on prenatal antibiotic exposure and avoid reverse causation (AD occurred before antibiotic/acetaminophen use), we performed an analysis with exclusion of AD diagnosed before 1 year of age, as shown in the subgroup analysis. All calculations were performed using SAS version 9.4 (SAS Institute, Cary, NC, USA).
Results