Spanish flu v. COVID-19: Comparing the Numbers and Forgetting the Lessons
Plus an insightful article from a prescient author (spoiler: yes, it was me)
You may have seen headlines such as this lately:
U.S. Covid-19 Death Toll Surpasses 1918 Flu Fatalities
The U.S. on Monday crossed the threshold of 675,000 reported Covid-19 deaths, according to Johns Hopkins University, which tracks data from state health authorities. The Centers for Disease Control and Prevention estimates the influenza pandemic killed about that many people in the U.S. a century ago, in 1918 and 1919. Both figures are likely undercounts, epidemiologists and historians say.
There are several differences between the current pandemic and the one that claimed nearly as many lives more than 100 years ago. The U.S. at that time was roughly one-third its current size, so the flu pandemic took a proportionately bigger toll on the population. That pandemic had a devastating effect on young people, including small children and young-to-middle-aged adults, while Covid-19 has hit older people hardest, according to health officials.
Yes, let’s talk about the bit I bolded.
Make meaningful comparisons: mortality is measured in rates
A couple of days ago, I started getting more twitter followers (follow me on twitter!), and a bunch of twitter notifications, and I wondered why.
This is why: COVID-19 death toll surpasses 1918 flu pandemic; Jonah Goldberg says to prepare for a lot of ‘when you adjust for population’ tweets — I bitched at Jonah Goldberg for something stupid he tweeted.
This is the Jonah Goldberg tweet: [with Jake Tapper intro]
And this was my response:
For crying out loud, order of magnitude counts.
Nobody considering mortality numbers looks solely at the total number of deaths. We want to know what percentage of the population it is (and what age, and other aspects, but I’ll get back to that).
The Spanish flu was far deadlier than Covid-19, and the Black Death was far deadlier than either of those.
The Spanish flu, killing about 670K out of 103 million people, which is a 0.6% crude death rate (over the period involved, which was about a year) is obviously deadlier than 675K deaths out of 333 million, which is a 0.2% crude death rate….over a year and a half.
So comparing just crude numbers, the Spanish flu was 3 times deadlier than Covid-19 (so far).
The Black Death, on the other hand, killed about a third of the population of Europe in the 14th century. 33% (or even higher in some places) is way higher than a 0.2% or 0.6% death rate.
Crude death rate is not enough — you want age-adjusted death rates
As I wrote in July, age-adjusted death rates are the best way to compare mortality trends if you just want to collapse mortality information in a single number.
The very short version is:
Age is a huge determiner of mortality, in any period
The age distribution in populations differ a great deal when you’re looking at events decades (or longer) apart
So if you want to compare mortality levels in single numbers, you need to look at age-adjusted mortality rates, which is a weighted average of death rates (where weights are standardized by age group), instead of crude death rates (which is total number of deaths per total population).
Here is an example of the difference between crude death rates and age-adjusted mortality rates:
That covers 1968 to 2020, so it doesn’t capture what happened with the Spanish flu pandemic. That said, we know some effects:
The disproportionate impact of the Spanish flu on young adults
In April 2018, the following article ran in Governing Magazine: Lessons From a Public Health Catastrophe. Let’s look at the age-related differences with the Spanish flu:
A final aspect to consider in preparing to respond to a pandemic is the age group that was hardest-hit by the Spanish flu: young adults, particularly those between the ages of 25 and 34. Excess mortality due to the pandemic was much higher for this age group than any other, over 900 deaths per 100,000 people. [that’s 0.9%]
That’s not an anomaly: People in prime ages may be felled in larger numbers in a pandemic than the very young and the very old, who are the usual sufferers in seasonal flu epidemics. So governments may find the population they generally rely on to provide critical services in a pandemic – such as EMTs, public-health workers and hospital staff — to be the ones most severely affected. Pandemic preparation plans should include the possibility that many core workers may be incapacitated.
That’s what the Spanish flu pandemic did, but not what COVID-19 did, thank goodness. The issues we’re having with healthcare workers right now are more policy-driven than mortality-effects-driven.
Due to the disproportionate impact on younger adults, the period life expectancy dropped a huge amount in 1918:
(if you want to learn more about period life expectancy, start with this: Mortality with Meep: Cohort vs. Period Mortality Tables or you can read this: Mortality with Meep: U.S. Life Expectancy Fell 2.4% in 2020, and Death Rates Increased 16.1%)
So, period life expectancy dropped about 12 – 13% in 1918 in the U.S., mainly due to the Spanish flu, because there was an outsized effect from young adults being the main group killed by the disease (also, period life expectancy was relatively short — under 60 years!). That was a drop of about 7 years.
But life expectancy dropped only about 1 year in 2020 due to COVID impacts, and that was a decrease of less than 3% compared to 2019.
So if you want to compare the effect of the Spanish flu vs. COVID-19 on the U.S. population, all of these rates —- percentage change in period life expectancy, age-adjusted death rates, or even crude death rate — are all more reasonable choices than simply number of people who died.
More lessons from the Spanish flu pandemic… not followed
If it’s not clear: the Spanish flu was far worse in its effects, in mortality at the very least, in the United States compared to COVID-19.
[so far]
But there are echoes of the screw-ups in 1918 that happened in 2020 and 2021.
Back to the article from Governing Magazine [from April 2018]:
Just as governments prepare for relatively frequent disasters such as hurricanes and ice storms, as well as infrequent emergencies such as major earthquakes and terrorist attacks, the experience of the Spanish flu pandemic shows that governments should plan as well for such infectious disasters. There are two aspects of the Spanish flu pandemic that governments should particularly remember and consider in this centenary year: communications and operations.
With respect to communications, consider how the Spanish flu got its name: Spain was one of the few major countries that did not suppress reporting on the pandemic. Spain was a neutral power during World War I, so it didn’t have the motive for quashing news of something even deadlier than war rushing through its population. Even the United States, the likely origin of the pandemic, kept the information restricted. As a result, some localities suffered more than they might have due to delayed response because they did not realize what was going on.
The problem extended beyond the quashing of information. Too often, governments projected certainty when there was uncertainty. This produced distrust among the public, leading to people not adhering to knowledgeable official advice in preventing the pandemic’s spread.
Okay, author, don’t hit us too hard.
As for operations, let us contrast two cities and their responses: Boston and New York City. Both being major ports, they saw ships with ill military men returning from World War I starting in August 1918. The need was to prevent the disease from spreading to the civilian population. Boston’s health commissioner thought the rapidity of the infection meant that the epidemic would burn itself out quickly. That was not the case. By the end of the pandemic, Boston had an excess mortality of 710 deaths per 100,000 people.
Contrast this to New York’s approach. When notified of sick people arriving in its harbor, ships were quarantined immediately and patients were isolated from the general public. New York City had been well prepared due to its general policy for dealing with epidemics and having in place a system to impose quarantines, curfews and crowd control to limit the spread. Even with ideal policies, this extremely virulent disease did spread. However, due to the efforts of the city government, mortality was reduced. The excess death rate for New York City was 452 per 100,000 people, the lowest for any city on the East Coast.
As severe as the 2017-18 flu season was for the United States, though, the Spanish flu pandemic was a disaster of a different order, taking the lives of an estimated 670,000 Americans at a time when the nation’s population was less than a third of today’s. It may seem hopeless to try to prepare for a public-health catastrophe of that magnitude. But being ready to keep the public informed and to move quickly to contain any outbreak that does occur — to learn from what happened a hundred years ago — is the right place to start.
Yeah, it would have been nice if any of the pertinent officials had read my article (yes, dear reader, I had written this Governing Magazine article) and taken some of the advice.
I was reminded this morning I had done my first draft of the Governing Magazine article four years ago (and yes, it got altered multiple times before it was finally published in April 2018). I had forgotten some of my points and came across….
Too often, governments projected certainty when there was uncertainty. This produced distrust among the public, leading to people not adhering to knowledgeable official advice in preventing the pandemic’s spread.
Ugh.
Just.
Uuuuuuuuugh.
Carry on.