HEALTH POLICY

Disparities in Health Appear Preventable

Policy Shift Might Be Effective Treatment

By the winter of 1984, reports of an alarming discrepancy in the rates at which white and black Americans were getting sick and dying had reached the ears of politicians. Concerned by the news, the Reagan administration established a task force to look into racial disparities in health. The 19-member group, appointed by then Secretary of Health and Human Services Margaret Heckler, confirmed in stark terms what had been suspected. It found that 60,000 excess deaths occurred each year in minority populations—deaths that would not have occurred if the persons had been white.

Graham Ramsay

Nancy Krieger and her colleagues used a variety of statistical techniques to analyze disparities in mortality rates. “We wanted to make it clear that we weren’t cherry-picking our approach to measure what the inequities are,” she said. From left, Krieger appears with co-authors Pamela Waterman and Jarvis Chen.

Political Connections
The idea that health disparities are tied directly to politics was not one that many entertained when Krieger and colleagues embarked on their study. Researchers, at least in this country, were preoccupied with linking disparities to an overall decrease in mortality. Some argued that as overall mortality goes down, inequities would increase because the gains in longevity will largely be made by the more affluent and educated sectors of society. Others claimed that a different phenomenon was at work, that improvements in health are driven by a “leveling-up process, where the worst off begin to approach and become more like those better off,” Krieger said.

Meanwhile, researchers in Britain had been looking at mortality rates in light of political changes, the rise of the Labor party in particular, and were finding interesting correlations.

Intrigued by these reports, Krieger, David Rehkopf, then a graduate student at HSPH, and colleagues wanted to see if American mortality rates could be pinned to political trends. “I had very strong priors that we would see a different pattern before and after 1980,” said Krieger. To get a broad enough historical sweep, they decided to begin their analysis in 1960. No one had looked that far back; nor had anyone taken into account socioeconomic as well as racial and ethnic data. “We realized we could take it back further if we had county mortality rates and economic data,” she said. “David probably can recite the names of all the U.S. counties because he got to know them very well.”

Using median income as their guide, they assigned each of the U.S. counties—all 3,073—to an economic quintile. They calculated the average mortality rate of the counties in every economic quintile for each of the 40 years. Using the broad racial categories available in the 1960s—essentially Negro (or black), white, and other—they then teased the quintiles apart into two groups, white and populations of color, and calculated average mortality for each of these more narrowly defined economic quintiles.

The researchers used a variety of statistical techniques to analyze the data, including one called joinpoint regression. “This allows you to look at the shape of mortality changes and see where there are inflection points—where the slope of the line changes,” Krieger said. The results were striking. Though the lines representing mortality rates of the American population as a whole exhibited a steady decline, as did those for more affluent whites, the lines representing mortality rates for minority and poorer groups, including whites, exhibited two inflection points: one around 1966, when the line begins to plunge and another around 1980, when the line begins to plateau (see figure).

Courtesy Nancy Krieger

Picture of health disparities. Premature mortality rates for populations of color (solid lines) undergo a steep decline, beginning around 1966, and a leveling off, beginning around 1980. White populations (dashed lines) exhibit a general decline in mortality, though the rate after 1980 is variable. The poorest white quintiles exhibit a leveling off while mortality rates continue to drop in the most affluent white quintiles. Interestingly, mortality also continues to decline in the most affluent minority quintile.

Using a technique similar to the one employed in the Heckler report, they calculated the number of excess deaths for each year—those that would not have occurred had the four lowest income quintiles experienced the same age-specific death rates as the most affluent quintile. For the period from 1980 to 2002, 18 percent of deaths would not have occurred. For populations of color, the percentage was 30.

It is still not clear that policy changes are to blame for the widening gap. “Here we’re showing what the actual tableau is, what has actually occurred,” Krieger said. “We do not have data to test any specific hypothesis. That is the work of future grants to more directly link changes in policy and circumstances to these kinds of outcomes.”

Nor is it clear how the gap will be narrowed, though Krieger is resolutely optimistic. “The point of this kind of research is not to throw your hands up and say, ‘Oh, it’s so hard, we can’t do anything,’ because the whole point of our paper is that there was a sustained period when the inequalities were shrinking. And there were real policies going along with that,” she said. “So that’s what the reminder is. It’s not inevitable, it’s not intractable, and it requires major mobilization and commitment of resources. But that’s fine because we have a major mobilization of resources with other priorities right now.”

—Misia Landau

top