Education, Income, and increased Mortality in Poorly Educated Whites in the US

In 2012, Sabrina Tavernise produced an alarming article for the New York Times: Life Spans Shrink for Least-Educated Whites in the U.S..  She was reporting on a study of mortality data that indicated that non-Hispanic whites with less education than a high school degree were experiencing a sharp drop in life expectancy.  Such a drop did not occur in the case of Hispanics and blacks without a high school education.  This chart was provided’

There was no good explanation for why this was occurring.  Education and income tend to track.  Consequently, there are a number of possible explanations based on less-healthy lifestyles that might be expected among a very low-income population. The drop in life expectancy was significantly greater for women than for men.  Therein may reside a clue as to what might be going on.

Such a large increase in mortality in a developed country is nearly unheard of.

“The five-year decline for white women rivals the catastrophic seven-year drop for Russian men in the years after the collapse of the Soviet Union, said Michael Marmot, director of the Institute of Health Equity in London.”

“By 2008, life expectancy for black women without a high school diploma had surpassed that of white women of the same education level, the study found.”

After a brief burst of publicity on this topic, it seems to have retired to academic circles with apparently no provable explanation in sight.

In the past few months another analysis of mortality data has emerged and provided new insights.  Anne Case and Angus Deaton have produced the article Rising morbidity and mortality in midlife among white non-Hispanic Americans in the 21st century.  They were particularly interested in health and mortality of those in the 45 to 54 age group.  They provided this chart to make sure we realized that something strange and troubling is going on in the United States.

Note that the curve for (45-54) Hispanics (USH) continues to follow that of other countries while only that of (45-54) non-Hispanic whites (USW) goes soaring into space.

The authors also conclude that education is a critical factor in the level of mortality.  All the rise in mortality comes from the cohort with a high school degree or less.  Those with some college education but no degree have slightly decreased mortality.  For those with a college degree or post-graduate education the mortality rate has continued to drop.  Note that the data presented by Tavernise was based on those without a high school diploma, a much smaller group.

Much of the increase in mortality comes from bad lifestyle choices.  Increases in drug use (poisoning) and alcohol abuse are indicated as major contributors, along with a greater number of suicides.

It is interesting that the data on mortality from diabetes (poor nutrition) and lung cancer (smoking) are not contributors to the rise.

The authors also break out the data on mortality from poisoning, suicides, and liver disease by age group for non-Hispanic whites.

If the factors considered here are the dominant causes of increased mortality, then there is a definite peak in the middle years, with lower increases for younger and older groups.  However, it is significant that all age groups indicate higher mortality from these causes.

The authors provide a brief discussion of possible explanations, but as with Tavernise, they can only speculate about changing lifestyle choices.  It seems that their emphasis on the age factor is perhaps more of a diversion than a fundamental clue.  The educational attainment variable seems to be the dominant effect, as identified in Tavernise’s article.

Education is important because it is associated with income, which in turn is correlated with quality of life (lifestyle choices), family stability, and economic security.  Let’s consider a few more pieces of data that provide additional insight.

This source provides an interesting look at how income (education?) affects longevity.  Consider this chart based on Social Security data.

Beginning in the 1970s, the life expectancies of wealthier 65-year-olds began to diverge from those of lower income people. The fact that lower income people have seen little increase in longevity at age 65 is a good counter argument to those who would claim that the Social Security retirement age should be raised.  One can think of reasons why this mortality divergence might occur, but one has to also explain why this effect suddenly began to occur in the 1970s.

There is an age-related phenomenon that might also provide a clue as to what is at work.  An article in The Economist titled Age and happiness: The U-bend of life provided this interesting chart.

When social scientists poll people on how satisfied they are currently with their lives they derive responses as a function of age that produce a U-shaped curve with a minimum in middle age.  If one equates satisfaction with life with happiness, then the younger are happier and the older are happier.  Scientists conclude that this type curve exists in all but a few societies, but the minimum can vary in age.

“….interest in the U-bend has been growing. Its effect on happiness is significant—about half as much, from the nadir of middle age to the elderly peak, as that of unemployment. It appears all over the world. David Blanchflower, professor of economics at Dartmouth College, and Mr Oswald looked at the figures for 72 countries. The nadir varies among countries—Ukrainians, at the top of the range, are at their most miserable at 62, and Swiss, at the bottom, at 35—but in the great majority of countries people are at their unhappiest in their 40s and early 50s. The global average is 46.”

If one returns to the age-grouped chart of Case and Deaton, a mortality versus age curve would look like the inverse of the U-bend curve just above.  This suggests a possible inverse correlation between mortality and happiness.  If the opposite of happiness and satisfaction is anxiety, then one can hypothesize that the stress related to increased anxiety in middle-age has deleterious health effects and increases mortality.

The article provides this input on the correlation between happiness and health.

“Whatever the causes of the U-bend, it has consequences beyond the emotional. Happiness doesn't just make people happy—it also makes them healthier. John Weinman, professor of psychiatry at King's College London, monitored the stress levels of a group of volunteers and then inflicted small wounds on them. The wounds of the least stressed healed twice as fast as those of the most stressed. At Carnegie Mellon University in Pittsburgh, Sheldon Cohen infected people with cold and flu viruses. He found that happier types were less likely to catch the virus, and showed fewer symptoms of illness when they did. So although old people tend to be less healthy than younger ones, their cheerfulness may help counteract their crumbliness.”

There is also this interesting finding that has some relevance to white versus black and Hispanic issues.

“In America, being black used to be associated with lower levels of happiness—though the most recent figures suggest that being black or Hispanic is nowadays associated with greater happiness.”

Finally, the article makes this assertion related to educational attainment, income, and happiness.

“Education, in other words, seems to make people happy because it makes them richer. And richer people are happier than poor ones—though just how much is a source of argument….”

If we are to make sense of all this data, we must identify a mechanism, or mechanisms, that increase mortality for whites but not blacks or Hispanics, and operates mainly on poorly educated people.  It must also be unique to the United States because it is apparently not operative in any other developed nation.  And yet it is even more complicated than that.  We like to think of the United States as a single country and average data nationwide in order to arrive at conclusions.  This averaging process can hide some rather significant excursions.

This source provides data on life expectancy at the age 50.  It tallies how many years one might be expected to live after reaching age 50 depending on which county one lives in.

The darker colors indicate lower life expectancies.  The amount of variation is enormous.  One could drive a hundred miles and find a location where people live twenty years less than they do in the place just left.

This data suggests that there are multiple factors important in determining mortality rates: climate, culture, ethnicity, race, occupation, environment…..  Good luck in sorting all that out!

The only thing we know for sure is that something is going terribly wrong in our society.