Socioeconomic status and birth weight: comparison of an area-based measure with the Registrar General's social class.

OBJECTIVE To compare the relation of birth weight with socioeconomic status measured by an area-based measure of material deprivation and by the Registrar General's social class. SETTING West Midlands Health Region 1991-93. STUDY DESIGN Retrospective cohort study. METHOD Birthweight data by enumeration district deciles ranked by Townsend Deprivation Index based on 1991 census data for all live births in the West Midlands Health Region were studied in three consecutive whole year birth cohorts, 1991 to 1993 and by Registrar General's social class in a 10% sample of live births (within marriage and jointly registered, provided by the Office of National Statistics) in the same region for the same period. Estimated proportions of births < 2500 g and < 3500 g "attributable" to social inequalities were compared for both socioeconomic status measures. The proportion of infants in each birthweight group were calculated for both measures. Relative risk (95% confidence intervals) of birth in each birthweight group for lowest versus highest socioeconomic status groups were calculated. RESULTS The estimated proportions of births < 2500 g "attributable" to social inequalities were 30% using the area-based measure and 27% using the Registrar General's social class. For births < 3500 g, the estimated proportions were 12% for the area-based measure and 7% for social class. There was a positive linear relation between the proportion of babies weighing > or = 3500 g and increasing socio-economic status measured by either method. Gradients in the opposite direction were noted for the proportion of babies born in the other birth weight groups. Relative risk of birth weight < 3500 g was 1.30 (95% CIs 1.28, 1.32) for most versus least deprived decile and 1.17 (95% CIs 1.10, 1.25) for social class V versus I. For birth weight < 2500 g the risks were 1.99 (95% CIs 1.85, 2.18) and 2.04 (95% CIs 1.53, 2.73) respectively and for birth weight < 1500 g, 2.11 (95% CIs 1.73, 2.57) for most versus least deprived decile (numbers too small for analysis in the Office of National Statistics sample). CONCLUSION A substantial proportion of births < 2500 g and < 1500 g are statistically "attributable" to social inequality. The results demonstrate that, using either socioeconomic measure, the likelihood of being born weighing > or = 3500 g, the most advantageous group, is substantially greater in the socially advantaged. Using the area-based measure, an estimated 12% of births < 3500 g could be ascribed to social inequalities whereas the same figure using social class was 7%. These findings suggest that this proxy measure of socioeconomic status may be a better discriminator in the study of pregnancy outcomes in this population than classification by occupational social class. Another advantage is its almost universal availability in routine records and its universal population coverage.