The Issue. In this article, I address 3 questions about research in the area of health inequities for children: why continue to do research, where should research efforts focus, and how best can we facilitate the development of research in the area of health equity? I was prompted to consider these issues by colleagues who think that doing research into child health inequalities is a waste of time. They say that not because they come from a right-wing perspective; on the contrary, they say, “Listen, we know about this stuff, why don’t you just get on and do something about it and stop worrying about the research?” Therefore, we must rethink the relevance of the why and the what before we consider the how.
Why should we research the relationship between socioeconomic status and children’s health? Given that the relationship between poverty and poor health has been demonstrated so often and for such a wide range of conditions, why should we continue to carry out these investigations?1 The first reason relates to child advocacy. Although the effects of poverty are uncontroversial, what is sometimes less clearly appreciated is that there is not a simple differential in health outcomes between those who are poor and those who are wealthy. There is virtually a straight-line relationship between relative wealth and adverse health outcomes, even within wealthy societies.2 The magnitude of the effect of socioeconomic status generally is underestimated by policy makers. In the United Kingdom, for instance, our research group has reported that almost 30% of newborns who weigh <2,500 g3 and perinatal deaths4 are statistically attributable to social inequality, yet it is only in recent years that professionals and politicians have begun to recognize the importance of this issue.
If we are to be effective advocates for children who are disadvantaged, then we need greater clarity about the nature of the link between socioeconomic status and children’s health. Many scientists have responded to observed health differentials by attempting to find explanations in biological characteristics or individual behaviors rather than structural factors. This is an essential endeavor and reflects a laudable desire to disentangle the proximal causes of social differentials in health. Unfortunately, it sometimes may lead either to failure to understand the role of causal pathways or to the use of analytic strategies that result in an unsophisticated interpretation of data.
For example, a study of social differentials in children’s height in the United Kingdom published in 1991 concluded the following:
“Height was negatively associated with social class, but the association was not significant after allowing for biological variables… . A social class gradient in height is accounted for by associations with biological factors, particularly the parental heights; environmental attributes are weakly associated with height after allowing for biological factors.”5
Treating parental height as a biological variable of course ignores the extent to which this is itself strongly socially determined and hence underestimates the effects of socioeconomic status in any analysis.
In addition to the risk of misunderstanding the nature of links between socioeconomic status and health, a preoccupation with behavioral or biological factors has led some authors to reach what I regard as untenable conclusions about the implications for interventions. This second quote is from a scholarly work dealing with the prevention of perinatal mortality and morbidity:
“There is a feeling that because there are strong trends in perinatal and neonatal mortality with social class, political action to equalise wealth or the housing of the lower classes would automatically result in an equalising of the death rate. From available evidence, this is unlikely to happen. As we have shown, a major determinant of the differences in death rates is the maternal smoking history—give the family more money, and it is possible that their consumption of tobacco (and alcohol) would increase. Rehousing schemes produce tension and stress, which may well have an adverse effect on the fetus or young infant.”6
Although these scientists may be driven by a desire for a better understanding of mechanisms, there remains no shortage of politicians who are prepared to use their results for less praiseworthy ends. There is a strong tendency to attempt to shift the responsibility for action away from society onto individuals, seeing them either as the agents of their own problems through unwise behavior or as the victims of their genes, about which the rest of us can do nothing.
The second reason for continuing this examination of the relationship between socioeconomic status and health is that it may shed light on etiology. At the simplest level, the distribution of putative causal mechanisms must be consistent with the observed social patterns of health. The most obvious example is perhaps the relationship between age of first pregnancy and breast cancer, which is thought to explain the reverse social gradient seen with this condition. Similarly, the strong relationship between socioeconomic status and prematurity and low birth weight has helped to concentrate attention on potential etiologic pathways that include factors such as low dietary folate, psychosocial stress, smoking, and vaginal infections, which seem to share similar social distributions in our societies.7
Where should research on inequalities focus? Three main areas seem important to pursue: the documentation of inequality, the investigation of the pathways that lead from social circumstances to health outcomes, and the evaluation of potential interventions.
It seems clear that we should continue to investigate and document the extent of health inequality to maintain awareness of the problem among politicians and policy makers. The undoubted general improvement in material prosperity and children’s health that has occurred in developed societies should not obscure the continuing discrepancies between the life chances of those born in different parts of the socioeconomic spectrum. There are many places in the world where large segments of the population actually have become poorer, and the discrepancies in children’s health between those countries and our own remain a source of shame.
As previously suggested, it is important to attempt to disentangle the mechanisms through which social factors can affect children’s health. Some of us, who see the effects of social disadvantage in structural terms, have been guilty of a tendency to reify socioeconomic status and to treat it rather simplistically as a “cause” of adverse outcomes. This has led us into the somewhat sterile debate about whether it is socioeconomic status or various health behaviors that cause health outcomes. Socioeconomic status is of course not a thing in itself but a marker for various constellations of risk factors related to material circumstances and socially determined behaviors. We need to develop a more sophisticated approach in which we consider the complex causal pathways through which different aspects of being socially disadvantaged can have effects on health (Fig 1), pathways that map proximal and distal causes. If we are to make the most of this information, then we need to work not only with colleagues in anthropology and other social sciences but also with molecular biologists to tease out the important pathways and mechanisms that link socioeconomic status and health. This approach has the potential to advance etiologic understanding and to suggest possible targets for interventions.
Finally, we need to direct increasing attention to the design and evaluation of interventions to address health inequalities. We can think about these interventions as being in 3 “layers”: structural changes, whole-population public health measures, and interventions aimed specifically to address the needs of those most in need. Much dialogue has occurred over the years about the type of structural changes that are needed. In essence, what we want is to change the distribution of resources in society, a deeply unpopular idea. This is not the only approach, although I remain of the opinion that it is the most important.
I was reminded recently of a wonderful talk given many years ago by a health administrator from Costa Rica, who said that when the then-current regime came to power, they said,
“Look, we know that our people are sick, and the reason our people are sick is because of the oppression from the world capitalistic economy. We have got to get rid of the world capitalistic economy.”
But after it had been in power for a while, they began to think that it did not seem likely that there would be a collapse of the world capitalistic economy. So they said,
“OK, well, we still think the world capitalistic economy needs to be overthrown, but in the meantime, maybe we can nibble away at some of these things that are going on.”
Of course, we know that Costa Rica has been one of the great success stories in health. So I think that as much as many of us would like to see structural change, we need to think about other interventions as well.
One reason that researchers often concentrate their attention on health behaviors is that the structure of society is seen as immutable. This is demonstrably untrue. The economic situation in which families find themselves is greatly influenced by government economic and fiscal policy. In the United Kingdom, for example, between 1979 (when the conservative government led by Margaret Thatcher came to power) and 1987, the proportion of children living in households below half the average national income rose from 12% to 26%.8 It is estimated that recent changes in the tax and benefit system in the United Kingdom will improve the life of children by raising 1 million children above the poverty line within the next 3 years.9
However, the effects of such changes alone, welcome as they are, are unlikely rapidly to eradicate current discrepancies in outcomes. Whole-population public health measures are a potentially important approach to improve health disparities. There are many past examples that have had important effects: the addition of vitamin D to staple foods to combat rickets, the near-universal provision of immunization against many diseases, and perhaps the provision of health visitor services. Childhood injuries, a particularly marked area of inequality, for instance, may be susceptible to improvement by interventions such as legislation covering speed limits, use of safe containers, or building design.
Despite the risk of implying an acceptance of the status quo or of stigmatizing the disadvantaged, I suggest that we need to consider interventions that aim specifically at those individuals or families most in need. By this I mean interventions that address areas of particular concern to the disadvantaged and take account of the forces that make them more at risk. For instance, in the United Kingdom, there is a strong relationship between socioeconomic status and the risk of smoking, an important cause of adverse child health outcomes. The major determinant of this differential does not seem to be rates of taking up smoking but rates of giving up. This is not because people who are poor are unaware of the dangers of smoking but because of the greater difficulties they have in quitting, something that presumably reflects the generally more stressful lives they lead. Programs such as those designed by Blackburn and Graham,10 which aim to help women who are poor quit smoking and attempt to take account of the particular difficulties they are likely to face, seem to offer one way forward. One could imagine the design of intervention programs in other areas that take account of the specific needs of the disadvantaged. Such interventions might range from providing access to preventive services outside of working hours and persuading supermarkets in poor areas to offer healthy foods as loss leaders to facilitating access to social support for young mothers.
If we believe that research in this area is important, then how best can we ensure that it happens? Perhaps the single most important thing we can do is refuse to settle for mediocre approaches to research methods. If we are to be taken seriously and really influence policy makers, then we need real improvement in the rigor with which we evaluate interventions and conduct etiologic research. If we are to attract the brightest young researchers, then it is important that this not be seen as an undesirable area of research.
It is true that there are difficulties in using the most rigorous research designs for some types of interventions that we wish to evaluate, but this must not be an excuse for sloppy thinking or avoiding the methods most likely to produce valid answers. The enormous influence of the evaluations of Head Start was in part because they were conducted with great methodologic rigor. The wealth of high-quality research in areas such as criminology demonstrates that it is possible to conduct rigorous evaluations of complex social interventions. The establishment of the Campbell Collaboration,11 an international group whose aim is to follow the example of the Cochrane Collaboration in health and to summarize the high-quality evidence in education and the social and behavioral sectors, will help to make this research more easily accessible to people in other fields.
High-quality research needs not only to use rigorous methods but also to address the questions for which practitioners and policy makers need answers. Widespread training in “evidence-based practice” among clinicians is a useful step in this process, helping people to identify their information needs and understand the rules of evidence. Since the establishment of the Centre for Evidence-Based Child Health,12 we have argued that, wherever possible, this training should include people from all disciplines working with children. This will help to ensure that the questions generated are relevant across disciplines and offer the opportunity to develop a common “language of evidence” that can help to span professional boundaries. We hope that this type of training also can encourage practitioners to see the value of research to the work they do and to see examples of high-quality research dealing with social determinants of health and relevant interventions.
Once we attract young researchers into this field, we need to nurture their development. In the United Kingdom, this can be difficult. There are relatively few academic departments working on community child health or on the effects of social factors on children’s health. Even for established researchers, it can be difficult to compete for research funds against those doing molecular biology. This inequity between the importance of social determinants of health and the scarcity of research funding could be addressed usefully by our professional organizations in their contacts with government and other funding sources. In particular, it seems important to address the need for infrastructure funding. What is required is research and training networks that include research methodologists and practitioners from a whole range of disciplines. An excellent example of this approach in a related field is the Can Child Centre for Childhood Disability Research in Ontario, Canada.13 The center is built on a network of academics and clinicians and produces some of the best research in childhood disability in the world. If we are to encourage young researchers into this area, then we have to offer them exciting, research-rich environments in which they can develop their skills.
I will end by suggesting that an understanding of basic epidemiology is important for all practitioners in child health. Epidemiology is one of the true basic sciences for our discipline, forming a bridge between research and its application to policy and clinical practice. The emphasis it gives to population-rather than individual-based perspectives on health can help clinicians and hopefully policy makers to recognize the crucially important links between the social environment and children’s health.
The Royal College of Paediatrics and Child Health-American Academy of Pediatrics Equity Project should consider development of a joint research program in social epidemiology. The program could define priority research questions in social epidemiology and implement joint research methodologies to answer these questions. This could then serve as the basis for preparing new pediatric researchers in social epidemiology and for translating new knowledge into practice.
- ↵Spencer N. Poverty and Child Health. Abingdon, Oxon, United Kingdom: Radcliffe Medical Press; 2000
- ↵Spencer N, Bambang S, Logan S, Gill L. Socioeconomic status and birth weight: comparison of an area-based measure with the Registrar General’s social class. J Epidemiol Community Health.1999;53 :495– 498
- ↵Gulliford MC, Chinn S, Rona RJ. Social environment and height: England and Scotland 1987 and 1988. Arch Dis Child.1991;66 :235– 240
- ↵Golding J, Butler NR. The socio-economic factor. In: Faulkner F, ed. Prevention of Perinatal Mortality and Morbidity. Basel, Switzerland: Karger; 1984:31–46
- ↵Oppenheim C. Poverty: The Facts. London, England: Child Poverty Action Group; 1990
- ↵Child Poverty Action Group web site. Available at: http://www.cpag.org.uk
- ↵Blackburn C, Graham H. Smoking Among Working Class Mothers: Information Pack. Coventry, United Kingdom: University of Warwick; 2000
- ↵University of Pennsylvania Graduate School of Education web site. Available at: http://www.campbell.gse.upenn.edu
- ↵Institute of Child Health, Paediatric Epidemiology and Biostatistics Unit web site. Available at: http://www.ich.ucl.ac.uk/ich/html/academicunits/paed_epid/ebm.html
- ↵McMaster University, Centre for Childhood Disability Research (CanChild). Available at: http://www.fhs.mcmaster.ca/canchild
- Copyright © 2003 by the American Academy of Pediatrics