For neonatal deaths, length of union, housing type, maternal age, previous low-birth-weight infant, vaginal bleeding, antenatal care, sex, congenital anomalies, and presence of problems during delivery remained in the multivariate analysis as significant risk factors. The association of these deaths with DWTD adjusted for covariates exhibited a pattern similar to that in the univariate analysis—that is, an increase in risk of neonatal deaths for mothers in the higher exposure categories. However, the trend observed in the univariate analysis was no longer clear. The adjusted OR in the higher category of exposure was 1.47 (95% CI, 0.67–3.19) (Table 4).