Objective To examine the temporal-spatial trends of very preterm birth (vPTB) and its association with social class.

Methods From the Aberdeen Maternity and Neonatal Databank we identified all spontaneous deliveries occurring in Aberdeen from 1950 to 2010. Gestation at delivery <32 weeks was defined as vPTB and was compared with a gestation of >32 weeks. Social Class based on Occupation (SC) was the deprivation measure; time trends of vPTB were examined in 5 year intervals. The association with SC was examined using logistic regression. Poisson regression was used to examine change in this relationship over time adjusting for potential confounders. Spatial analysis used post code mapping of vPTB in comparison with the Carstairs index.

Results The odds of having a vPTB in 2005–2010 were 1.35 (95% CI 1.03–1.79) in comparison to 1950–1954. The adjusted Odds Ratio for unskilled social class and vPTB was 1.79 (95% CI 1.28–2.50). When temporal changes were taken into account, the Rate Ratio for vPTB was 2.18 (95% CI 1.42–3.37) in the unskilled social class compared to the professional social class. However, when smoking status, ethnicity and age were included this association reduced to a Rate Ratio of 1.73 (95% CI 1.04–2.90). Spatial analysis did not show any relationship of vPTB with Carstairs’ categories of social deprivation.

Conclusion Deprived social class is associated with risk of vPTB but this is partially explained by confounding. Interventions to prevent vPTB are better directed at risky maternal behaviours such as smoking rather than targeting deprived areas.