A recent study highlighted in the New York Times starkly showed how differences in life expectancy between the rich and poor in the United States have grown remarkably over the past few decades. Researchers at the Brookings Institute used data from the University of Michigan's Health and Retirement Study to compare life expectancy by levels of earnings for men and women born in 1920 and those born in 1940. Their analysis showed that for men born in 1920, there was a 6.2 year difference in life expectancy at age 50 between the top 10 percent of earners and the bottom 10 percent. For men born only twenty years later in 1940, that difference had nearly doubled to 11.3 years. The increase was even greater or women: the difference increased from 4.7 years to 10.3 years.

These are striking data not only because of the large increase in disparity over time but also because of the sheer magnitude of the differences themselves: an over 10 year difference in life expectancy between rich and poor for those born in 1940. In men, this increase in the disparity occurred in the context of increases in life expectancy overall, but in women, as reported previously by other research, there appear to have been actual declines in life expectancy in lower income groups.

Differences in life expectancy are summary indicators of differences in the timing of death, but they do not capture differences in morbidity. Given what we know about the burden of chronic illness and disability in the US population, and the relationship between income and chronic illness, income differences in health overall are likely to be substantially greater than what we observe for life expectancy. Moreover, the Brookings study focused on life expectancy at age 50. Differences are likely to be even starker when infant mortality and other deaths before age 50 are taken into account.

The finding that mortality is strongly patterned by indicators like income or education and moreover that disparities in death by social position have been increasing in the United States over the past decades is not new: Kitagawa and Hauser demonstrated this in a landmark study in 1975 and Pappas et al replicated this finding for later years in a study published in the New England Journal of Medicine in 1993. Indeed prior research has shown that mortality differences by social position were actually narrowing in the US during the first half of the 20^th century, but began to increase beginning in about 1960.

A common argument is that these patterns result from reverse causation, i.e. the fact that poor health causes people to lose income. But reverse causation has been ruled out as a major explanation by data showing links between parental income and the health of children and by a large body of work showing that income is related to early markers of disease that could not have an effect on earnings.

Although there is a strong temptation to attribute the income differential to differences to health care access, we know for a fact that access to health care has a very small impact on population-level variations in health and mortality and moreover differences in health and life expectancy by income exist even in societies with universal access to health care. Genetics, another popular explanation for virtually all health differences, can also be ruled out given the sizes of the differences, their variation over time, and the many different outcomes for which the differences are observed (ranging from cardiovascular disease, to mental illness, to injuries).

So what explains this strong patterning? Well, most likely it results from many interrelated factors acting together. Income affects living conditions and resources that in turn affect behaviors we know are strongly predictive of health. Income also affects exposures to built and social environments that have consequences for health (through exposure to things like air pollution or by shaping behaviors). There has been a large and hotly debated literature on the extent to which living in an unequal society is itself stressful. But regardless of whether inequality itself (via social comparisons) is a major source of health damaging stress, clearly low income is strongly associated with greater exposure to many stressors that likely have implications for health as well. These factors probably reinforce each other and operate over the life course and even across generations.

The presence of large differences in mortality by social class (as well as recent increases in these differentials) is not unique to the United States and has been demonstrated in many other countries. However, data also show that this phenomenon is not invariant, indeed as demonstrated by Mackenbach and colleagues, there is quite a bit of heterogeneity in the magnitude of the health inequalities. These differences are not a “natural” and unavoidable aspect of human societies.

One way to intervene to eliminate the disparity in health between rich and poor is to use a targeted approach to block the mechanisms linking low income to adverse health. Reducing smoking among the poor for example would likely have an important impact on differentials in mortality. Improving heath care access could also have an impact although only on diseases where medical care has a major effect on survival. These things are important and should be done. However, as articulated by Link and Phelan, the systemic nature of the differences and the multiplicity of factors through with social position affects health (as well as historical experience) implies that differences often re-emerge in other ways when a specific factor (like smoking) is blocked.

The unavoidable conclusion is that addressing the adverse health consequences of income inequality in a substantial and sustained way will require acting on income inequality itself. This is not only a moral imperative (driven by the profound injustice that a persons income is among the strongest predictors of their life expectancy), but also a public health imperative, because we know that improving population health cannot be achieved without addressing the fundamental drivers of disparities in health within the population.

The pervasive link between income inequality and health is especially relevant today, in the context of high and rising US income inequality and a demonstrated US health disadvantage compared to other nations. Our task in public health is to continue make the connections between income inequality and health visible and to demonstrate the health consequences of social and economic policy broadly defined. This is especially urgent today, when in many circles, discussions of the drivers of health continue to be narrowly focused on health care and the promise of precision medicine.