Understanding socioeconomic disparities in breastfeeding in the UK:Exploring the role of environmental quality

Abstract

Guided by the evolutionary framework of life history theory, which predicts lower parental investment in lower-quality environments, my PhD research explores socioeconomic differentials in breastfeeding behaviour in the UK with a particular focus on local environmental quality. My research is quantitative, and I use advanced statistical techniques to analyse two large UK cohort datasets: the Millennium Cohort Study (MCS) and the Born in Bradford (BiB) cohort. The thesis is comprised of three papers. The first paper compares objective and subjective summary measures of the local environment, using factor analysis to pull together both physical and sociocultural aspects. Using multi-level modelling on nationally-representative data (the MCS), I isolate the effects of the local environment above and beyond that of individual socioeconomic status (SES) and wider-scale deprivation and other ward-level factors such as ethnic composition. I find that objectively-assessed environmental quality is a more robust indicator of breastfeeding initiation and duration than subjectively-assessed environmental quality and also that higher individual socioeconomic status provides a buffer, protecting the breastfeeding chances of those with more resources, even in low quality environments. With Paper 1 providing a UK-wide picture of the relationship between environmental quality and breastfeeding, my second paper zooms in on one geographical region in particular. In Paper 2 I use the BiB dataset, with its largely bi-ethnic Bradford population to look at the influence of the physical environment (e.g. air and water pollution) on breastfeeding outcomes and whether there are differences between White British and Pakistani mothers. I use structural equation modelling to explore whether associations are mediated by birth outcomes. I find the predicted negative association between SES and breastfeeding, but no strong or consistent evidence for the same relationship when physical measures of environmental quality are used. Paper 3 uses both datasets to situate breastfeeding within a wider suite of reproductive behaviours and demographic and health traits including other parental investment measures, menarche and age at first birth and uses latent class analysis to test whether these characteristics cluster together to form identifiable life history strategies along an environmental quality continuum. This thesis examines the local environment in detail, operationalising environmental quality in different ways, to see whether environmental quality is an important driver of the SES-breastfeeding association in the UK. Overall, the thesis findings suggest that individual SES is a stronger predictor of breastfeeding than environmental quality, but that the two are strongly linked, and exert their own independent effects. The effects of the local environment are however complex and depend on which indicators of environmental quality are used. Associations are not driven by individual environmental perception as much as they are by more objective measures of environmental quality. Less perceivable, more physical measures of environmental quality do not explain the SES-breastfeeding association, suggesting that in the UK context at least, sociocultural environmental factors are likely to have an important influence on breastfeeding outcomes. Although exact pathways and mechanisms remain unclear, intervening at the environmental level has the potential to improve breastfeeding behaviour, as well as other health and reproductive outcomes.