Good health — like wealth — does not trickle down the economic ladder.

That’s one conclusion that should be drawn from a widely covered paper published this past week in the Proceedings of the National Academy of Sciences, albeit for reasons that may not be immediately apparent.

The paper — authored by Anne Case and recent Nobel Prize winner Angus Deaton, both of Princeton University — presents an alarming finding: middle-aged white (non-Hispanic) Americans experienced a stunning reversal in mortality trends in the early twentieth-first century, unique among demographic groups. Between 1999 and 2013, this group saw a major — and previously unnoticed — rise in mortality.

The magnitude of this phenomenon becomes clear in the investigators’ conclusion: “The mortality reversal observed in this period bears a resemblance to the mortality decline slowdown in the United States during the height of the AIDS epidemic.”

By their estimation, had the death rate for middle-aged whites simply remained at its 1998 level, 96,000 fewer people would have perished during this period. And if the rate had, as might be expected, continued to decline at its historic pace, some 488,500 deaths could have been averted.

This staggering die-off was accompanied by a general deterioration of health status: more middle-aged whites reported “fair or poor health,” various chronic pains, and “serious psychological distress.” Rates of heavy drinking and abnormally high liver enzymes — a marker of liver injury — also rose. Much of the uptick in mortality was due to various “external causes,” such as poisoning (by alcohol or drugs), liver disease, and suicide.

In sum: a demographic and social disaster has been quietly — almost invisibly — unfolding in America.

The story is incomplete, however, without the critical element of class. While the study provides evidence of more acute suffering within the demographic as a whole, the increase in overall mortality was restricted to those with no more than a high school education. Those with more than a high school diploma also saw an increase in mortality from poisonings and suicide, but still managed to enjoy an overall decline in death rates (as typically happens in modern societies when there isn’t a war, famine, or major epidemic).

As a result of such differentials by class, the ratio in mortality rates between those with a high school degree or less compared to those with a bachelor’s degree or more increased from 2.6 (in 1999) to 4.1 (in 2013). This spike was so pronounced it dragged down the rate for the white middle-aged cohort as a whole.

Case and Deaton’s study complements a larger literature demonstrating that health inequality along class lines — approximated by socioeconomic indicators like income and education — is rising. While it’s long been known that class affects health, decades of cross-class declines in mortality make this dynamic less obvious. Recent research, however, paints a different picture: with the historic rise in economic inequality, there’s been an attendant increase in health inequality.

This year, a publication from the National Academies of Sciences, The Growing Gap in Life Expectancy by Income, confirmed earlier reports in finding that the difference in life expectancy after age fifty between top- and bottom-income groups was much higher for those born in 1960 than those born in 1930. The analysis was limited by a number of assumptions and extrapolations, but its bottom line remains striking: despite three decades of medical progress and economic growth, there’s been no projected improvement in life expectancy for males in the lowest income quintile — and females in this income group have seen a projected decline in life expectancy.

Rising health inequality helps explain how overall health measures can improve even as those of particular populations stagnate — or in the case of Case and Deaton’s study, greatly deteriorate. Such inequities may also be partially driving the US’s lagging health metrics in relation to other countries.

As health researcher James House points out in Beyond Obamacare: Life, Death, and Social Policy, America went from being a health leader to a straggler in the second half of the twentieth century. Despite rising health expenditures,

our standing on a number of indicators of population health — such as life expectancy, infant mortality, and, more recently, maternal mortality — has been declining from among the best in the world in the 1950s to worse today than virtually every developed nation in the world, as well as a number of developing ones. We are even beginning to see some evidence of absolute declines in the health of portions of the American population.

Case and Deaton’s study demonstrates this is precisely what we’re witnessing in the US today.

What are we to make of the factor of institutional racism in this study? It is not that being white has suddenly become bad for one’s health; indeed, despite the dynamics described in the study, black Americans still suffer worse health and lower life expectancy. But as the health policy scholar Vicente Navarro argued in a 1990 Lancet article, understanding racial disparities in health requires the addition of a class-based framework:

The stark fact is that these [racial] differentials cannot be explained merely by looking at race. After all, some blacks have better health indicators (including mortality rates) than some whites, and not all whites have similar mortality indicators. Thus we must look at class differentials in mortality in the US, which are also increasing rather than declining.

This isn’t to suggest that race doesn’t matter outside of class. In a 2005 article, social epidemiologist Ichiro Kawachi and colleagues emphasize the importance of confronting inequalities in each of these realms.

Racism — both past and present — is a crucial determinant in producing worse health among blacks as compared to whites. But within demographic groups and US society as a whole, class is an increasingly critical determinant of health. It’s easy to miss the underlying dynamic of this chilling reality when we focus on aggregate statistics.

To reverse the trends Case and Deaton document, it’s undoubtedly important to fight the various “upstream” causes of mortality, such as substance abuse or mental illness. But if we narrowly focus on specific risk factors for death — say by controlling the opioid epidemic, which has rightly become a major public health priority— we risk missing a larger, more daunting reality: class-based health inequality preceded the opioid epidemic, and will no doubt survive it.

Without more fundamental change, the larger structure of health inequity will persevere, and possibly worsen. Population health is deeply interwoven into the economic structure of society. Those who seek to improve the former must, it is increasingly clear, join in the contest to transform the latter.