Background: Low socioeconomic position is generally associated with increased risk of preterm birth, but it remains unclear whether the inequality depends on the socioeconomic measure used, if the associations differ according to the degree of prematurity, and how individual level risk factors mediate the association.
Methods: The hazard ratios (HR) of preterm birth associated with five different measures of socioeconomic position and three degrees of preterm birth were analysed in a dataset of 75 890 singleton pregnancies (1996-2002) from the Danish National Birth Cohort. This, and the mediating role of selected individual level risk factors (smoking, alcohol consumption, binge drinking, pre-pregnancy body mass index, gestational weight gain) were estimated, using Cox regression analyses.
Results: Mothers with <10 years of education had an elevated risk of preterm birth compared with mothers with >12 years of education and the association interacted with parity, while income and occupation affected the risk to a lesser degree. The adjusted HR for less educated nulliparous and parous women were 1.22 (95% CI 1.04-1.42) and 1.56 (95% CI 1.31-1.87), respectively, compared with women with >12 years of education. For parous women with <10 years of education inclusion of smoking in the model decreased the HR of preterm birth to 1.43 (95% CI 1.19-1.72).
Conclusions: Maternal educational level was the strongest predictor of preterm birth among five socioeconomic measures and the gradient did not differ significantly according to the degree of preterm birth. For parous women smoking explained some of the educational gradient but in general the selected risk factors only reduced the relative educational gradient in preterm birth marginally.