Prevalence of breastfeeding at four months in general practices in south London

Background: Successive quinquennial National Infant Feeding Surveys have pr ovided a valuable picture of national and regional variations in infant fee ding practices within the United Kingdom. Social variation in breast feedin g had been recognised to be an important source of health inequalities in c hildhood by the Independent Inquiry into Inequalities in Health Report. Aim: To determine the prevalence of breastfeeding at birth and at four mont hs in a sample of women from urban general practices, its variation between practices, and relation to practice population deprivation scores. To repo rt the timing of introduction of solid feeds. Design of study: Cross-sectional questionnaire survey based on a random clu ster sample Setting Women with infants aged four months in general practice s in South London. Method: Mode of infant feeding at birth and four months, and rime of introd uction of solids. Jarman score as a measure of practice population deprivat ion. Housing tenure, maternal ethnic group, and maternal age at leaving ful l-time education. Results: Twenty;five general practices were sampled. Median practice Jarman score was 15.0 (interquartile range [IQR] = 12.6-21.9). Responses were rec eived from 1053 out of 1532 mothers approached (69%). Of these, 87% (897) h ad breastfed at birth, while 59% (609) were still breastfeeding their babie s at four months. Mothers in rented accommodation were less likely to breas tfeed than owner-occupiers (odds ratio [95% CI] = 0.52 [0.37-0.74]), as wer e women of white, compared with chose of black, ethnic origin (odds ratio [ 95% CI] = 0.55 [0.36-0.82]). Those who completed up to two years and more t han two years education after the age of 16 were 2.94 (95% CI = 1.85-4.66) and 9.25 (95% CI = 6.02-14.21) more likely to breastfeed at four months; re spectively, than mothers whose formal education was completed at or before 16 years. Practice-specific rates of breastfeeding ranged from 71% to 100% at birth (median 87%; IQR = 79-93%) and 22% to 83% at four months (median 6 1%; interquartile range = 47-66%). The intra-practice correlation coefficie nt for breastfeeding at four months was 0.052 (within-cluster variance = 0. 23 between-cluster variance = 0.013). There war no association between brea stfeeding at four months and practice-specific Jarman score. Median age of starting solids was 16 weeks (IQR = 15-17 weeks). Conclusions: Housing tenure, maternal education, and ethnic group are signi ficantly associated with breastfeeding prevalence at four months Between-pr actice variation in breastfeeding prevalence is nor associated with measure s of practice population deprivation, ac. assessed by Jarman scores. Consid eration should be given to including information on maternal ethnic group a nd housing tenure in future National Infant Feeding Surveys. Current weanin g practices fall short of the recommendation of the World Health Assembly.