(Bloomberg Opinion) -- Covid-19 is especially hard on the old and frail, with 79% of the deaths attributed to the disease in the U.S. among those 65 and older and 94% among people with at least one “comorbidity” such as diabetes, dementia, obesity or hypertension.
This has led some people to argue that most of the country’s 200,000-plus deaths so far from the disease somehow shouldn’t count, which is ghoulish and awful. But I don’t think it’s ghoulish or awful to point out that a lot of the people dying from the coronavirus this year would have died from something else in the not-too-distant future, albeit likely in more pleasant circumstances than alone in a hospital room hooked up to a ventilator.
By “a lot” I do not mean most, and by “not-too-distant future” I do not mean next month. Going purely by the age distribution of U.S. Covid-19 victims, one recent study estimated that on average they would otherwise have had an additional 11.7 years to live. Attempts to adjust for underlying medical conditions among the victims bring that number down somewhat but not radically; a study based on data from Italy found that doing this reduced the lost life expectancy by about a year. Meanwhile, a recent estimate of the loss of quality-adjusted life years, which fully count years of good health and discount those beset by infirmity, put the average for U.S. Covid victims at about seven.
Still, these averages do imply that tens of thousands of Covid-19 victims in the U.S. had pre-pandemic life expectancies of less than a year or two, meaning that in the epidemic’s wake we should expect to see the phenomenon that is known in mortality studies by the definitely ghoulish term “harvesting.” As in, after the Grim Reaper makes an especially big harvest of souls, he’ll have fewer to reap in the immediate aftermath.
“You would expect a decline in deaths following an epidemic, arising from the fact that epidemics, by targeting frail individuals, make the population suddenly more ‘robust’ on average, which can then lead to a deficit of deaths once the epidemic is over,” explained mortality-statistics expert Michel Guillot, a professor of sociology at the University of Pennsylvania and senior researcher at the French National Institute of Demographic Studies.
For a straightforward example of the phenomenon, Guillot pointed me not to an epidemic but to an August 2003 heat wave that killed an estimated 45,000 to 50,000 Europeans, the vast majority of them elderly. In its immediate aftermath, deaths among French people ages 85 and older reverted to more or less normal, with the aftereffects of heat stress perhaps canceling out any harvesting effect. But in 2004, they were below normal all year long, ending down 13.5% from the 2000-2002 average. Together, 2003 and 2004 saw a 6.6% drop in 85-and-older deaths from from the earlier period, probably attributable to mild winter flu seasons as well as harvesting effects. It was as if the heat wave had never happened (except that it had).
In other words, optimists hoping for a post-pandemic mortality decline that cancels out at least some of the death toll from Covid-19 aren’t necessarily wrong, but they may have to wait a while. This is especially true in the U.S., where the disease has never stopped spreading and weekly deaths from all causes have remained well above normal since late March. Because U.S. weekly mortality data trickle in with long and variable lags from the individual states, and the people at the Centers for Disease Control and Prevention don’t do a great job of either communicating the incompleteness of the data or successfully adjusting for it in their excess-death estimates, some of the aforementioned optimists have repeatedly fooled themselves and others into believing that the Covid mortality wave is over and the happy time of reduced deaths is upon us, but so far they’ve invariably been proved wrong when the numbers get revised upward over subsequent weeks.
The CDC’s numbers for New York City do appear to be up-to-date, though, and as one of the places in the U.S. hardest hit by Covid-19 in the spring and most successful in restraining it since, it also seems the likeliest to see below-normal mortality among the elderly. And yes, total deaths among those 85 and older have been a little lower than the 2017-2019 average since mid-June — but only 0.7% lower. From late March through early May, they were 324% higher.
Several European countries were even more successful in driving down Covid-19 cases and deaths over the summer than New York City was, although there’s been a lot of backsliding lately. So did deaths among the elderly fall below normal during the hiatus? They did, here and there. Among the countries I looked at, the effect was most apparent right after the epidemic waned in Italy, but it wasn’t very big and with data available only through midyear we don’t know yet if it lasted.
Also apparent in the chart is that deaths among elderly Italians were well below the previous three years’ average in January, the result most likely of a milder-than-usual flu season. This also shows up in other European mortality charts. In Germany, the large European nation that has so far been the most successful in fending off Covid-19, deaths during the worst weeks of this year’s coronavirus epidemic were actually lower than those during recent flu seasons, and since then deaths have been slightly above normal.
Deaths among those 85 and older were also running well below normal in France before Covid-19 hit, but that hasn’t been the case since.
In Spain, meanwhile, deaths among those 85 and older have since mid-July been consistently higher than in past years — perhaps offering an early indication of the second Covid-19 wave now sweeping the country.
And in England and Wales deaths are running 4% below the past years’ average since mid-June, but I wouldn’t make too much of that just yet.
There will almost surely come a time in the aftermath of the Covid-19 pandemic when markedly fewer elderly people are dying than usual. We just haven’t made it to that aftermath yet, and I’m guessing that we won’t in the U.S. and most of Europe until sometime next year.
In the meantime, 2020 is likely to see big increases in overall mortality. As of Sept. 5 the CDC’s estimated increase in deaths over expected levels so far this year in the U.S. was about 254,000, or 12.8%. Even if deaths revert to average for the rest of the year, which seems unlikely, this would represent the biggest increase in the U.S. mortality rate (in deaths per 100,000 residents) since the influenza-pandemic year of 1918.
Because of the age profile of the disease, though, the life expectancy decline would be less historic. In the same study that estimated 11.7 average years of life lost by U.S. Covid victims, University of California at Berkeley demographers Joshua R. Goldstein and Ronald D. Lee also estimated that 250,000 additional deaths due to the disease would decrease average U.S. life expectancy at birth by about 10 months (0.84 years) — the biggest one-year drop since World War II but smaller than decreases in 1943 and several years in the 1920s and 1930s (mostly due to bad flu seasons), and just a fraction of the staggering 11.8-year decline in 1918.
In France, Guillot and Myriam Khlat of the National Institute of Demographic Studies estimated in June that the country’s average life expectancy decline this year — assuming no second wave in fall, which at the time seemed like a reasonablish assumption — would be just one-tenth of a year for women and two-tenths for men. Lower Covid-19 mortality in France is part of the explanation, as is that milder-than-usual 2019-2020 flu season and an age profile of Covid deaths that is skewed even more to the very old than that of the U.S. A much higher share of U.S. Covid-19 victims are under 65 than in other affluent countries, the result of a somewhat younger population in the U.S. but also of a much less healthy population. Only 4.8% of French people ages 20 to 79 have diabetes, for example, while 10.8% of Americans do.
Maverick French doctor and hydroxychloroquine booster Didier Raoult soon picked up on Guillot and Khlat’s estimate to argue that it meant Covid-19 just wasn’t that a big a deal. I don’t think that’s the right conclusion. But it is a very different sort of a deal than, say, a war or a disease that targets the young.
To be sure, life expectancy at birth isn't a great gauge of the impact of a pandemic either, given that it effectively assumes that the pandemic will keep happening every year. But it at least reflects the impact-by-age in a way that mortality rates don't.
This column does not necessarily reflect the opinion of the editorial board or Bloomberg LP and its owners.
Justin Fox is a Bloomberg Opinion columnist covering business. He was the editorial director of Harvard Business Review and wrote for Time, Fortune and American Banker. He is the author of “The Myth of the Rational Market.”
©2020 Bloomberg L.P.