Rates and social patterning of household smoking and breastfeeding in contrasting European settings

To compare rates and social patterning of household smoking and breastfeeding in families with newborn infants in birth cohorts in Coventry, UK and Veria, North Greece.

Infants born in 1996 in Coventry, 1999 in Veria were recruited into birth cohort studies using similar methodologies. In Coventry recruitment was by family health visitor at the primary visit; in Veria, hospital-based paediatricians enrolled infants at the neonatal examination. Data were collected at the initial contact on household smoking, type of feeding, and household socio-demographic characteristics. Rates of initial breastfeeding and household smoking with 95% confidence intervals were estimated and breastfeeding and household smoking regressed on parental education and housing tenure in logistic regression models.

Data were available on 2612 Coventry infants and 773 Veria infants. Rates of household smoking and breastfeeding were higher in Veria compared to Coventry. In Coventry, living in rented accommodation and lower maternal and paternal education were associated with household smoking and bottle feeding. Logistic regression models fitted on initiation of breastfeeding failed to show social patterning in Veria but more educated mothers showed a longer duration of breastfeeding. Only low paternal education was associated with household smoking after adjustment for maternal education and housing tenure.

Smoking and breastfeeding are more prevalent among households with young infants in Veria compared with Coventry. The social patterning of health-related behaviours noted in Coventry is less marked in Veria. The relevance of these findings for public health interventions in the contrasting settings is discussed.