Covid-19 mainly kills old people. So do most other diseases

Covid-19 is a highly infectious disease for everybody over 40, but is less harmful than a disease of similar infectiousness.

(File photo) 80% Covid-19 deaths in US are among those 65 and older | Photo | Bloomberg

Covid-19 is hard on the elderly, with those 65 and older accounting for 80% of the U.S. deaths from the disease for which the Centers for Disease Control and Prevention has released demographic data. But this is true of most illnesses: In 2018, 78% of all U.S. deaths from internal causes (that is, excluding accidents, murders, overdoses and the like) were among those 65 and older.

Assessing the risks posed by Covid-19 has become a morbid global pastime. So has weighing them against the risks posed by other things like seasonal influenza or car accidents or drowning in the pool. A key element in many of these comparisons is the age distribution of the risk. The coronavirus mostly kills people with a limited number of years ahead of them, some have reasoned, so its potential death toll should be discounted relative to that of an ailment or other risk that has a greater effect on the young.

This is not incorrect. But it’s also not a reason to dismiss the threat the disease poses. “When you get old, something kills you,” economist and demographer Lyman Stone wrote on Twitter after sharing some charts that inspired the one above. “But Covid is an extra something. An extra wolf in the pack.”

Seasonal influenza, to which Covid-19 is often compared, is also a disease that preys mostly on the elderly, with the age distribution of its victims often tilted even more toward the very old than that of the coronavirus so far. In 2018, 83% of influenza and pneumonia deaths in the U.S. were among those 65 and older, and two-thirds among those 75 and older. (Many influenza-related deaths are attributed to pneumonia on the death certificates from which the CDC’s detailed age-group numbers are derived, which is why I’ve combined the two here.)

The new strains of influenza that cause pandemics don’t always work this way. Only 13% of the estimated 12,469 U.S. deaths in the 2009-2010 H1N1 pandemic were among those 65 and older, and the elderly’s apparent resistance to that virus was one reason the fatality rate was so low. The earlier H1N1 strain that caused the pandemic of 1918 and 1919 was much more virulent, causing an estimated 675,000 deaths among a U.S. population less than a third the size of today’s, and was especially hard on those in their 20s through 40s.

So no, Covid-19 doesn’t seem to be nearly as bad as the 1918-1919 flu! But with its death toll at 74,000 and rising after just two months of significant spread in the U.S., it is already worse than the worst influenza season of recent years — 2017-2018, with an estimated 61,000 influenza-related deaths. You may also have heard 80,000 cited as the death toll for the 2017-2018 flu season. That was an earlier estimate from the CDC that has since been revised downward, which is a good reminder that all the numbers we have for Covid-19 remain quite preliminary. Given that this pandemic is far from over, though, and that early signs from overall mortality data indicate deaths from the disease have been undercounted so far, the numbers have nowhere to go but up. Their age distribution could change a bit over time, though.

In order to use the age distribution so far to give a sense of the risk posed by Covid-19 relative to other bad things that can happen to people, I’m going to start with 100,000 deaths from the disease over the course of the year as the almost-certain-to-be-exceeded low-end scenario. For the high end, I’ll take the 0.23% of New York City residents who have died from the disease so far and multiply that by the U.S. population, which gets me to a bit above 750,000. That number may seem improbably high, but it’s far from a worst-case scenario in which the disease spreads unchecked (in which case millions would likely die) and it does seem like a useful reference case of what public health officials around the country are trying to avoid. I’ll also run the numbers for 200,000 Covid-19 deaths, a plausible if perhaps optimistic estimate of what might happen if (1) the disease sticks around and has a big resurgence in the fall, as pandemics tend to do, but (2) we get better at treating it and keeping the most vulnerable from catching it.

Here’s how those three scenarios stack up against influenza and pneumonia mortality in 2018. The numbers in the second through fourth columns are simply the estimated mortality (in deaths per 100,000) by age group for each scenario, divided by the 2018 influenza and pneumonia mortality.

For children, the risk of dying from Covid-19 appears to be much less than that of dying from influenza and pneumonia even in the most pessimistic of my three scenarios. From age 15 onward the Covid risk is higher, with the relative difference peaking for those in the 45 through 54 age group, whose risk of dying from Covid-19 in my middle scenario is four and a half times higher than their influenza and pneumonia mortality rate in 2018.

Causes of death other than disease tend to have very different age profiles from the coronavirus and the flu. With transport accidents, which the CDC says caused 41,283 fatalities in 2018, the very old still face the greatest mortality risk, but below age 65 the highest-risk categories are teenagers and young adults.

For those ages 1 through 24, even a wildly out-of-control Covid-19 pandemic that exceeded the worst-case scenarios and killed 5 million Americans would appear to pose less risk than cars and trucks and boats do in a normal year. Then there are swimming pools, which I mentioned above mainly because of talk-show host Phillip McGraw’s headline-making assertion on Laura Ingraham’s Fox News program last month that 360,000 Americans die in swimming pool accidents every year. McGraw (aka Dr. Phil) had added a couple of zeros to the annual number of drowning deaths, which was 3,706 (not counting boating-related drownings, which are considered transport accidents) in 2018. Swimming pools accounted for only 746 of those drownings, natural-water accidents 1,770 and bathtubs 536. I’m not going to bother with a table, but suffice it to say that for toddlers the overall risk of drowning is several times greater than the risk of Covid-19 under all three scenarios, and up through age 34 the risks are in the same ballpark.

What are we to make of such comparisons? Well, young people really truly do not seem to be in a lot of danger from Covid-19. It’s not zero danger: A few kids are dying from the disease. But the risk they face from catching it at school appears to be much smaller than the risk they face from, say, getting to school. This doesn’t necessarily mean the school and university closings of the past couple of months were a mistake — they were aimed chiefly at protecting teachers, parents and the population as a whole, not students. It does mean that opposing school reopenings in the name of protecting the kids doesn’t make a lot of statistical sense.

Dismissing Covid-19 as a disease that only endangers those who were going to die soon anyway doesn’t make a lot of statistical sense either. The mortality risks from Covid-19 start exceeding the risk from influenza or transport accidents well below age 65, and the middle-aged and younger seniors can face bigger relative risk increases than the very old. Here, for example, is the increase in mortality risk that 200,000 U.S. Covid-19 deaths in 2020 would imply relative to the actual all-causes mortality rates from 2018.

For a less-hypothetical look, here’s the increase in mortality rates implied by Covid-19 deaths in New York City through May 3, in this case relative to the city’s 2017 all-causes mortality numbers. That is, I’ve divided the city’s confirmed and probable Covid-19 deaths by age group by the overall deaths per 100,000 in the city in 2017. These aren’t exactly the percentages by which mortality rates will increase in New York in 2020 — some people who died of Covid-19 would have died of something else this year anyway, some have died or will die of other causes because the coronavirus pandemic kept them from getting needed treatment, plus the Covid-19 death toll will continue to rise. But it’s a useful approximation.

These are staggeringly large increases in mortality risk, especially given that all the city’s Covid-19 deaths have occurred in less than two months. And they are risk increases for the entire population, not just those infected by the coronavirus. David Spiegelhalter of the Winton Centre for Risk and Evidence Communication at the University of Cambridge calculated estimates for the latter using U.K. mortality data and found that the resulting increases in annual mortality risk start at about 20% for boys under 10 and 30% for girls, then follow a similar pattern of being highest for those in their 50s, 60s and 70s, for whom being infected with the coronavirus appears to at least double their risk of dying in a given year.

Does this age distribution make Covid-19 less harmful than a disease of similar infectiousness and deadliness that targeted children and young adults would be? Definitely! (I’m 56, so I think I’m allowed to say that.) But it remains a highly infectious disease that for everybody over about 40 is significantly deadlier than anything else they’re likely to encounter during the course of a normal year.-Bloomberg