Ethnic differences in the income/child health gradient: the role of antenatal care and maternal health behaviours
Stavros Petrou, University of Warwick
Ron Gray, Maggie Redshaw, Maria Quigley, National Perinatal Epidemiology Unit, University of Oxford
Previous empirical studies in the United States, Canada and United Kingdom have shown a positive relationship between family income and child health. Nevertheless, the channels through which income translates into better child health and the extent to which these potential mechanisms vary between ethnic groups are still poorly understood. Identifying the pathways through which family income operates before and after child birth can have important public health and fiscal policy implications. It can inform government decisions about the optimal timing of intervention programmes aimed at tackling child health poverty-driven inequalities.
This study provides a comprehensive analysis of the ethnic differences in the potential mechanisms underlying the income/child health gradient in the UK. We use data from the Millennium Cohort Study, a nationally representative birth cohort study collecting information on health, wealth, education, employment and family characteristics of almost 19,000 children born in the UK in 2000-2001. Using a child health production function we model two child health outcomes, namely height-for-age and parent-reported health status, at ages 9 months, 3 and 5 years for six ethnic groups. Child health is considered a function of income and parental inputs and characteristics other than income. Our main focus is to determine the extent to which health risks during pregnancy (intrauterine environment) and the use of antenatal care by mothers can account for the observed income gradient in child health and how the impact of these factors varies across and within ethnic groups. We consider also the impact of other potential mediating mechanisms occurring in early infancy such as physical health pathways (i.e. breastfeeding, household smoking, maternal self reported health), psychosocial pathways (maternal depression, child-parent-relationship) and extra-familial contexts (i.e. neighbourhood and childcare quality, cultural environment) and assess the degree to which they contribute to the income/health gradient. We address the problem of potential income endogenity by adopting a “mopping-up” approach in our cross-sectional analyses and a fixed-effects approach in our longitudinal analyses.
We expect our results to (i) increase understanding of the mechanisms underlying the relationship between family income and child health outcomes and (ii) inform the current policy debate on the optimal timing of intervention programmes for targeted populations. If a healthy foetal environment and frequent use of antenatal care by mothers account for most of the income/child health gradient as opposed to the impact of pathways subsequent to child birth, we could argue, according to the emerging “antenatal investment hypothesis” (Doyle, Harmon, Heckman and Tremblay, 2009) that intervention programmes aimed at assisting women at the start of their reproductive life might be more cost-effective than implementing a later policy of remediation. We also expect a differential effect of the various mediating mechanisms across ethnic groups. These results will help policy makers identify the sub-populations at which specific interventions should be targeted.