The effects of poverty on health in early childhood have been recognised for many years and have generated a prolonged and often acrimonious causal debate.¹ Explanations have broadly fallen into two groups: those that focus on characteristics of the poor themselves such as the inferiority of their “genetic stock”, their poor hygiene, their poor childcare practices, or their health related behaviours such as smoking and those that focus on the effects of poverty itself and the societal structures that generate it.

Given the longstanding interest in poverty and child health, it is surprising that, as Séguin et al² state in their paper in this issue of the journal, the aspects of socioeconomic status that are important to children’s health remain unclear. Part of the explanation for this continuing lack of clarity lies in the tendency to treat different components of socioeconomic status as adequate proxies for the complex constellation of factors that constitute socioeconomic status. For example, UK studies have been heavily reliant on the registrar general’s social class, a measure based on the occupation of the head of the household, which has been shown to be of limited value in the study of maternal and child health.³ US studies, using the official poverty line to create a poor/non-poor dichotomy, have been shown to underestimate the effects of severe and long term poverty on child health and development.⁴

A component of socioeconomic status that has been extensively used in studying child health, especially in less developed countries but also in developed countries, is maternal education. Education is seen as a powerful promoter of child health and as a means of breaking the so called “cycle of deprivation”.⁵ The resultant research and policy attention on promoting parental skills through education^6–8 shifts the focus away from the societal determinants of family and child poverty towards the perceived failings of poorly educated mothers. However, educational attainment is strongly correlated with the socioeconomic status of family of origin⁹ and educational opportunity depends as much on family income as cognitive ability.¹⁰ Cognitive ability itself is sensitive to adverse social circumstances in early childhood.^11,12

The paper by Séguin et al is an important addition to the literature on poverty and health in early childhood.² It is methodologically sound and is based on prospective data collection in a large cohort with a very small attrition rate. The paper investigates the impact of material hardship, inadequate income to meet needs, not simply income and the analysis accounts for important potential confounders such as maternal education. Material hardship has not been widely used as a measure of the financial aspects of socioeconomic status and its use in this study overcomes some of the problems of misclassification bias associated with current income and other income proxies such as occupational class and housing tenure. The longitudinal methodology also allows the authors to distinguish short term and long term material hardship. The results suggest that, at least in the Quebec context, longstanding material hardship has an effect, independent of maternal education, on a range of health related outcomes in early childhood.

The findings in the paper of Séguin et al raise important questions for future research into the social determinants of health in early childhood and, if confirmed in further studies, have major policy implications.² Further studies in different countries will be required to confirm the independent impact of material hardship on child health. The differential impact of material hardship and maternal education may vary with the health outcomes studied. Séguin et al² studied the impact on acute illness, asthma, hospitalisation, and growth; however, maternal education may account for more of the influence of material hardship on other outcomes, for example, child mental health. Further studies of a range of different health outcomes are needed to resolve these questions.

The issues raised by this paper are not simply arcane problems of research methodology. They have potentially far reaching policy implications. If the findings that material hardship has an impact on child health independent of maternal education are confirmed then policy responses based primarily on educating mothers rather than alleviating material hardship are likely to be unsuccessful. As Séguin et al suggest, social policies would need to ensure that all families with a young child have sufficient income.² In the UK, US, Canada, and elsewhere, this represents a considerable challenge to the policy directions of current governments.