Objective: To examine whether routinely measured variables explained the increased risk of preterm delivery in some UK ethnic groups.
Design: Cross sectional study of deliveries recorded in the Child Health Record System.
Setting: North Birmingham, UK.
Population: All North Birmingham women delivering singletons, 1994-1997 inclusive.
Method: Logistic regression.
Main outcome measures: Odds ratio (OR) and 95% confidence interval (CI) for preterm delivery, defined as less than 37 weeks, less than 34 weeks and less than 28 weeks, unadjusted and adjusted for maternal age, an area-based socio-economic status measure, and marital status, year of birth, fetal sex and past obstetric history.
Results: For Afro-Caribbean women, the ORs (95% CIs) were: for delivery less than 37 weeks, 1.44 (1.26-1.64) unadjusted and 1.22 (1.07-1.41) adjusted; for delivery less than 34 weeks, 1.55 (1.25-1.92) unadjusted and 1.29 (1.02-1.61) adjusted; for delivery less than 28 weeks, 1.66 (1.08-2.55) unadjusted and 1.32 (0.84-2.06) adjusted. For African women, the risk of delivery less than 37 weeks was not significantly raised; for delivery less than 34 weeks, the OR (95% CI) was 1.88 (0.99-3.58) unadjusted and 1.78 (0.93-3.40) adjusted; for delivery less than 28 weeks, the OR (95% CI) was 4.02 (1.60-10.12) unadjusted and 4.10 (1.66-10.16) adjusted. In Afro-Caribbeans, deprivation and marital status explained the differences between the unadjusted and adjusted ORs. There was a linear relation between deprivation and preterm delivery for all ethnic groups, except for Asians.
Conclusions: Factors associated with deprivation and marital status explain about half of the excess of preterm births in Afro-Caribbeans, but not Africans. The risk of preterm delivery might not be related to deprivation in Asians.