Figure 3.

Dose-response relationships of the cumulative number of indictors for residential traffic within 200 m of the residence with odds of childhood pneumonia, stratified by the child’s sex. (a) Boys; (b) Girls. Adjusted factors in model 1 (multivariate logistic regression analyses) included age, residence-located area, family history of atopy, residence ownership, breastfeeding duration, household dampness-related exposures, household ETS, and household renovation during early lifetime. Adjusted factors in model 2 (two-level logistic regression analyses) included age, family history of atopy, residence ownership, breastfeeding duration, household dampness-related exposures, household ETS, and household renovation during early lifetime. The supplemental Table S5 presented the detailed data for odds ratios (ORs) and their 95% confidence intervals (CIs). The p-values for interaction in crude model, adjusted model 1, and adjusted model 2 were <0.001, 0.002, and 0.006, respectively.