Fig. 1.

Multivariate model of variables associated with overall supplement use in (a) early pregnancy (3MP, three months prior to conception; PPA, after conception but prior to pregnancy awareness; T1, from awareness to gestational week 13) and (b) mid to late pregnancy (T2, second trimester; T3, third trimester) among a population-based prospective cohort of women with a viable singleton pregnancy (n 2146), Melbourne, Australia, July 2011–July 2012. Data are presented as individual odds ratios with 95 % confidence intervals represented by vertical bars. Reference groups: age = 30–34 years; ethnicity = Caucasian; parity = 0 previous children; educational attainment = tertiary education; annual household income = > $AU 100 000. Note that all variables are mutually adjusted for each other. Non-significant univariate predictors (3MP: maternal pre-pregnancy BMI, maternal alcohol consumption (yes/no) pre-conception; PPA: maternal pre-pregnancy BMI; T1, post aware: maternal alcohol consumption (yes/no) in T1, maternal smoker status (yes/no) in T1, dietary folate intake, maternal pre-pregnancy BMI; T2: pregnancy planning (yes/no), maternal ethnicity, maternal alcohol consumption (yes/no) in T2, dietary folate intake; T3: pregnancy planning (yes/no), maternal ethnicity, parity, household income, maternal alcohol consumption (yes/no) in T3, dietary folate intake) were excluded from the model