Is Life Expectancy Overestimated?

Sept. 15, 2003 -- We may not live as long as we think. And this time, it has nothing to do with obesity, smoking, or the other usual life-threatening culprits.

It has to do with the way the life expectancy calculations are made, suggests a new study.

"We should subtract two years from the U.S. life expectancy estimate," says researcher John Bongaarts, PhD, of the Population Council in New York. "It's not that we're not living longer. We're just not living as long as we think."

In this week's Proceedings of the National Academy of Sciences, Bongaarts challenges the formula typically used to calculate life expectancy, which estimates it's 79.5 years at birth for women and 74.1 years for men.

Those estimates are made under a formula that Bongaarts says has been used for nearly 200 years. In theory, it tracks the life span of a hypothetical group of people from birth at a particular year until the last person dies.

"But you would have to follow those specific people for 100 years, which you cannot do," he tells WebMD. "So what the statistician markers do is to observe the number of deaths that occur at every age in people born in a particular year, and we know that X percentage of 62-year-olds die and Y percentage of 92-year-olds."

These calculations become the basis of what's known as "mortality life tables" that most government agencies and the private sector use to estimate life expectancy.

Design Flaw

The problem: These estimates provide a summary of life span over nearly an entire century -- until the last hypothetical member dies. And in those 100 years or so, conditions that predict mortality are likely to have changed.

So when the mean death age rises because of advantages in health care, medical technology, and just better living -- as they have been doing in the U.S. and elsewhere -- current life expectancies calculated at birth actually overestimate life expectancy, says Bongaarts.

By his calculations, the typical American actually lives two fewer years than current projections, while Japanese people live 3 1/2 fewer years.

"There is a slight bias in the current formula. Under my theory, we need to shave off a couple of years in most countries with a rising life expectancy," he tells WebMD.

Others Agree

Bongaarts isn't the first statistician to charge that the current method of calculating life expectancy is flawed, says Patrick Heuveline, PhD, of the University of Chicago's Population Research Center.

"Many people in this field are aware of the debate over the way life expectancy is calculated, but there is some disagreement on the best way to change it, if at all," Heuveline tells WebMD.

"Because there is no real way of extrapolating what will happen 60 or 100 years from now, people estimate life expectancy on the safest assumption that can be made -- things will remain the same."