IE 11 is not supported. For an optimal experience visit our site on another browser.

Mortality mystery: Female death rates are rising--and it's not clear why

The sad state of America’s health is no secret. Our food culture makes obesity the default. Our car culture makes walking an act of defiance.
File Photo: Women wait for various medical procedures outside a barn during the  Remote Area Medical (RAM) clinic July 20, 2007 at the Wise County Fairground in Wise, Virginia. Rural families, most with little or no insurance, lined up for hours to...
File Photo: Women wait for various medical procedures outside a barn during the Remote Area Medical (RAM) clinic July 20, 2007 at the Wise County Fairground...

The sad state of America’s health is no secret. Our food culture makes obesity the default. Our car culture makes walking an act of defiance. Our health care system deprives millions of preventive care, and our winner-take-all economy leaves millions feeling a lack of control over their lives. “For many years, Americans have been dying at younger ages than people in almost all other high-income countries,” the Institute of Medicine declared in a recent report titled Shorter Lives, Poorer Health. “This disadvantage has been getting worse for three decades, especially among women.”

How much worse? New research from the University of Wisconsin’s Population Health Institute sheds light on that question, and the findings are as intriguing as they are worrisome. Writing in the journal Health Affairs, epidemiologists David Kindig and Erika Cheng report that 43% of all U.S. counties saw increases in female death rates over a recent 14-year period, while male mortality rose in just 3% of counties. The new study identifies some shared characteristics among the 1,334 counties where more women are dying prematurely. But no one knows exactly how these social and geographic factors shape mortality rates―or why women and men are on such different trajectories. What’s clear is that wide swaths of the country are on a long losing streak.

Not all the news is bad. Nationally, the risk of dying prematurely (before age 75) declined in both sexes during the study period. Of the 3,140 counties that Kindig and Cheng tracked, 97% experienced reductions in male mortality from 1992 to 2006, and more than half saw female death rates fall. Yet 108 counties saw increased mortality among males, and 12 times that number (1,334) saw female death rates rise. What could explain these reversals? To find out, the researchers examined a range of likely contributors to see which ones the backsliding counties had in common. “Our unit of analysis was the county itself,” says Kindig. “We looked at social, behavioral, geographic factors to see how each of them contributed to the changes we saw. We also looked at medical factors, such as the number of doctors per 100,000 people and the percentage of adults without health insurance.”

Health care may seem like a crucial variable, but it had no discernable effect on death rates. “Surprisingly,” the authors write, “none of the medical care factors predicted changes in male [or female] mortality during the study period.” As you’d expect, counties with high rates of smoking and obesity were more likely to suffer increased premature mortality. But the strongest predictors of premature death were neither medical nor behavioral. Here are three key findings:

•Counties with high concentrations of Hispanic residents were less likely to see their death rates rise―no great surprise because Latino immigrants are healthier on average than U.S.-born residents.

•Counties with higher concentrations of college graduates had lower death rates, and less educated counties had higher ones, even when the researchers corrected for differences in income.

•As the maps below make clear, the rise in early death was highly concentrated in the South and parts of the West. In fact geography was the strongest protective factor the researchers identified. Counties situated in the Northeast and the Upper Midwest were largely immune from mortality surges, even when the researchers factored out differences in wealth, education and behavior.

How could location alone exert such powerful effects on health? “We’ve done similar studies at the state level, and region always surfaces as a dummy variable,” Kindig says. “Clearly something is going on. It could be cultural, political, religious or environmental, but nothing in the national data set lets us unpack the question. We need more granular data, and we need anthropologists to look at what’s qualitatively different about people’s lives in the higher-risk regions. This makes you want to jump on a bus and start taking notes.”

Migration patterns may offer a partial clue. Just as healthy people are more likely than their infirm neighbors to leave struggling Latin American countries to seek livelihoods in the United States, the healthiest residents of this country’s depressed rural regions may worsen their health profiles simply by leaving for more economically vibrant areas. Future studies could gauge the “healthy-immigrant” effect by monitoring the flow of people through different parts of the country. While helping to explain the declining health of Appalachia, the Mississippi River Valley and some inland parts of the West, a closer analysis of migration could help explain the tremendous health gains that northern urban areas have enjoyed in recent decades. As the authors observe, the urban advantage may also stem partly from more “government spending on . . . public safety, social and welfare services, affordable housing, and education.”

Since 2010, the Population Health Institute has ranked the nation’s counties on health-related measures ranging from air quality to poverty rates to the prevalence of junk-food outlets. When you look at places like Mingo County, West Virginia, or Owsley County, Kentucky, it’s no great surprise to learn that both have suffered rising death rates in both sexes. College education is a rarity in these parts. Child poverty is the norm, recreational facilities are scarce, more than half of the restaurants are fast-food outlets, and nearly a third of adults lack social support. Here is how Mingo County's death rate compared to state and national averages in the four years following Kindig and Cheng’s study period:

What lessons can we draw from these findings? One is that we’re still surprisingly ignorant about the forces affecting our health. Medical, behavioral, social and geographic factors all surely affect us. But together, the ones we can measure explain less than half of the disparities this study documents. And they tell us virtually nothing about the huge gap between male and female mortality trends. Women still live significantly longer than men in this country. Yet their health is declining measurably throughout much of the country, and no one knows why. “There’s something we’re not measuring,” says Cheng. “The question is: What?”

Smoking trends are probably part of the answer. The surge in female smoking started later, and tapered off later, than the surge among men, and the health consequences can take decades to set in. So the recent rise in female mortality may partly reflect smoking patterns from past decades. While researchers study such questions, policymakers would do well to consider what it has already shown. “None of the medical care variables that we examined had significant relationships with mortality change,” the authors observe. “Meaningful health improvement efforts must extend beyond a focus on health care delivery and include stronger policies affecting health behaviors and the social and environmental determinants of health.” Communities around the country are beginning to treat health as a community asset that involves housing, education, transportation and the food supply. It’s a fundamental shift from the traditional view of health as a personal achievement. And as these findings suggest, it’s long overdue.