Inequalities in health are well documented and have been the subject of much academic investigation and extensive prior evidence and data reviews in the UK over many years (e.g. Marmot et al., 2010, 2020; Department of Health and Social Care, 2020). So, when the IFS Deaton Review panel began its deliberations, it was immediately clear to us that our review of inequalities would not be complete without consideration of inequalities in the health domain. These inequalities are some of the most important and salient inequalities in modern societies. They are also systematic – with health varying by place, by ethnicity and race, by education, family background and socio-economic status.
Our aim in commissioning the evidence in this theme was not to duplicate, or even summarise, the extensive and authoritative pre-existing evidence that is out there. Instead, we wanted to commission evidence that would allow the panel to position what is known about health inequalities in the context and framework of all the other various inequalities we are documenting and investigating, and therefore to be able to bring health (both as an outcome of interest and as a cause or consequence of other outcomes of interest) into our broader understanding of the underlying forces that are driving inequalities in Britain.
Often, issues of health are considered narrowly in terms of income, education or, in the case of the UK, area-level deprivation. One common factor in all themes of the Deaton Review is that we aim to go beyond this, incorporating issues of political economy, labour, discrimination and geography, as well as framing inequalities in terms of social, economic and political relations and power. This approach seems especially valuable for understanding health inequalities. It permits a fuller discussion of the socio-economic issues beyond what we might call the ‘usual suspects’, which are often discussed in a way that is isolated from the context in which they occur or the upstream factors that caused them.
One thing is clear: poor health outcomes and early death rarely happen suddenly or out of the blue. For the most part, they emerge slowly over the life course. Models of this process tend to emphasise one or more of three potential mechanisms: 1. sensitive or critical periods during development or early childhood; 2. cumulative exposure in which continuing disadvantage takes an additive or interactive toll on health; and 3. a social trajectory model in which previous disadvantage sets one up for toxic exposures later in life that then take a toll in other areas (for instance, early educational experiences then shape exposure to toxic work conditions that are detrimental to health). All three mechanisms will plausibly play a role in determining a wide range of health consequences and health inequalities. Thinking of the evidence this way is important in that it allows a focus on the potential role of different areas for policy change, and for an understanding of what sorts of policies might be aimed at different subgroups of the population.
The main evidence chapter for this theme is by Anne Case and Lucy Kraftman, who begin by discussing key inequalities in life expectancy and mortality, including by cause of death. They document the long-standing differences in life expectancy among regions and countries in the UK and the persistent and large risks in mortality among those with less education and those who live in more deprived areas. While gaps in early childhood mortality between more and less deprived areas have been closing, progress in closing these gaps for adults has largely stalled since 2010.
Case and Kraftman go on to present new evidence on how health inequalities develop and persist across the life course. For the most part, they focus on early childhood, which they show is clearly an area where there are critical exposures. These early conditions set the course for later adult health. Using the most recent data on adult outcomes from the birth cohort studies, they document that socio-economic inequalities at birth and in early childhood development have strong consequences for health well into adulthood – for example, mental health or malaise at age 33, early death (before age 50) or self-reported health at age 55. This theme of how health inequalities develop across the life course is taken up, in different ways, by the authors who provide commentaries and perspectives on the Case and Kraftman chapter or on the issue of health inequalities more generally.
Daisy Fancourt and Andrew Steptoe use the life-course framework but, rather than focus on early life, they discuss the importance of adult experiences and exposures in driving subsequent health inequalities, which they broaden to include psychological well-being. This is important not only in its own right but also because a consideration of the role of multimorbidities shows that mental health problems early in the life course lead to serious non-communicable disease in later years. They highlight the role of ‘non-traditional’ risk factors such as work stress, divorce, bereavement and family conflict, and loneliness and social isolation, and they point out some ways in which policy could address health inequalities resulting from the differential patterning of these risk factors. Finally, they also argue that positive psychological characteristics, including emotional well-being, optimism and purpose in life, are all malleable and could well offset disadvantages experienced in early life.
Two commentaries consider the role of institutions in health inequalities. Janet Currie discusses a range of economic factors that could drive the morbidity and mortality differences that develop between rich and poor over the life course. She highlights the role played by safety nets and public health infrastructure, and places special emphasis on the importance of safety nets in reducing inequalities in childhood health. The commentary highlights the specific impacts of material deprivation as well as exposures to toxic environments that are patterned along economic lines. Currie suggests that perhaps ‘the main reason that life expectancy is low in the US relative to other rich countries is not because there is more income inequality per se, but because the US lacks an adequate social safety net’, which chimes with the issues (also discussed in the Case and Kraftman chapter) of whether austerity measures imposed in the UK following the financial crisis of 2008 can help explain the health inequalities we find in the UK today and the stagnating life expectancy and widening inequalities in mortality at older ages that have since occurred.
Carol Propper considers inequalities in healthcare through the lens of a model that thinks about healthcare as one element of the ‘production’ of health over the life course, where other elements are individuals’ circumstances and behaviours. The picture that emerges is that healthcare has an overall ‘pro-poor’ distribution if one considers quantity. But this is driven by the fact that poorer individuals have worse health and hence more healthcare needs. Once one takes this into account, there is generally a slight ‘pro-rich’ distribution of quality, experience and access to services including waits for treatment. There is, however, some evidence that the impact of austerity on the NHS budget has perhaps worsened these ‘pro-rich’ inequalities in recent years. Propper also points to important gaps in our understanding of this topic. First, there is an absence of much research linking healthcare utilisation inequalities directly to inequalities in health outcomes. Second, in order to better design policy, we need to understand more about lifetime rather than contemporaneous usage of healthcare, as well as the extent to which any pro-rich inequities are attributable to the behaviour of healthcare suppliers as opposed to demand-side socio-economic differences in healthcare seeking and self-care behaviour.
Finally, a commentary by James Nazroo on health and healthcare inequalities and how they intersect with ethnic inequalities adds a critical dimension not covered elsewhere, and at the same time connects to the other work in the Review on race and ethnicity. He documents large ethnic inequalities in health that differ by ethnic group and by type of disease and argues that, while these are best understood as the product of social and economic inequalities, they are not just a reflection of generalised class or socio-economic group specific processes. Instead, he argues that the issue of racism is fundamental. In the US, much attention has been focused recently on the degree to which structural racism – by which we mean institutional and structural properties that shape housing, labour market and educational opportunities – ultimately has the effect of shaping economic and social resources along racial and ethnic lines. In the UK context, it is important for us to reconcile evidence related to both social class and measures of educational and economic advantage/disadvantage with the patterning of health by race/ethnicity. Nazroo’s commentary also sets this evidence alongside related issues of individual and institutional racism.
In addition to the huge amount of important specific evidence and narrative that is presented in the evidence here, two crucial more general points emerge, which are either explicitly or implicitly made by all of the authors and commentators.
First, a crucial factor that limits the analysis of health inequalities that can be carried out for England or the UK is the need to rely on area-level indicators as measures of exposure when using whole-population data on mortality or hospital records. Many of the area-level measures were developed to indicate where health and economic resources should be allocated in terms of health and social services. Thus, they were designed for allocation and descriptive purposes as opposed to for understanding the ‘causal’ ingredients in the economic environment that impact health outcomes so deeply. As Case and Kraftman put it, ‘It is not possible to understand the upstream mechanisms responsible for health outcomes using geographic indices of multiple deprivation. Such indices, which transform and combine social and economic health inputs at a small-area level, are not helpful in understanding the mechanisms linking health to income, class, employment and education’. The more detailed individual-specific measures used in a number of longitudinal studies, such as the birth cohort and ageing studies, which also record health outcomes and link to mortality and hospital records, lend real strengths and can provide important perspective. But analyses using such data can only get such depth by trading off sample size and losing the ability to look at small population subgroups or study very specific conditions. This data gap, which could potentially be filled by better use of record linkage, along the lines that began to take place during the COVID-19 pandemic, should be viewed as a priority for policymakers seeking to understand health inequalities and what to do about them.
A second broad theme is that in order to better understand the mechanisms and forces driving health inequalities, one needs to focus on specific outcomes such as mortality rates at particular ages, cause of death, or specific diseases and comorbidities. While period life expectancy may be a good summary statistic for the contemporaneous health of a population, it does not actually represent the health or mortality prospects for any well-defined group of individuals. As geographical, ethnic and socio-economic inequalities in health are becoming more salient, the use of life expectancy to document health inequalities across such dimensions has become common, since it is a statistic that is readily available. But not only does period life expectancy not allow for cohort effects, it also implicitly assumes no one ever moves into or out of different groups, an assumption that is particularly strong when there is non-negligible interregional or international migration. The evidence provided here demonstrates how much further one can go by considering more specific indicators of individual health and population health inequalities and then linking these to the potential processes that drive them.