In a climate of international conflict and turmoil, a disease begins to spread across the globe. It hits the United States on the way to becoming a worldwide pandemic. While most infected survive, the fatality rate is well above that of an ordinary flu. Eventually, one in three humans on earth is infected. At least 17 million people, and maybe as many as 100 million, perish.
This is not a prediction — it’s a description of how the influenza pandemic of 1918-’19 (which came to be erroneously known as the “Spanish flu”; more on that below) became one of the deadliest, if not the deadliest, disease outbreak in modern history.
Comparisons to the 1918 outbreak have been rife since the novel coronavirus outbreak began in January. We recently marked the centennial of the 1918 pandemic, and fear of a repeat pervades discussions of the current outbreak.
But even granting that we are early in the history of the coronavirus outbreak, there are important differences between our current predicament and the Spanish flu. The underlying diseases are different, and from what we know at this point, the case fatality rate of the coronavirus is lower, by some estimates, than that of the Spanish flu. We also enjoy a much more robust public health infrastructure in 2019; in 1918, as Laura Spinney documents in her pandemic history Pale Rider, medical experts still hadn’t agreed that the flu is caused by a virus.
Other differences, like the advent of widespread passenger air travel and much deeper global supply chains, put us at greater pandemic risk now than in 1918.
But many experts caution that we should avoid comparisons to the Spanish flu. As flu expert Jeremy Brown writes in the Atlantic, “What’s most striking about these comparisons … is not the similarities between the two episodes, but the distance that medicine has traveled in the intervening century.”
Spinney concurs. “The Spanish flu is one of the worst, if not the worst, pandemic humanity ever went through, and it’s really anomalous in the history of flu pandemics,” she explains. “We’ve had 15 flu pandemics in the last 500 years, and the last five since the 1890s have been properly measured in a scientific way. None [but the Spanish flu] has killed more than 3 million people maximum.”
The “Asian flu” of 1957 and the “Hong Kong flu” of 1968, for instance, were both met with more modern tools of disease surveillance and had death tolls in the range of 500,000 to 2 million: big numbers, for sure, but hardly 1918 levels.
A very, very brief history of the 1918 flu
As Spinney notes in Pale Rider, it can be difficult to pin down the exact origins of the 1918 influenza outbreak.
One theory has it starting on American soil, in Kansas, where it migrated from birds to humans. Albert Gitchell, an Army private and mess cook based in Fort Riley, Kansas, is sometimes identified as the first victim, reporting his symptoms on March 4, 1918. Unfortunately for the world, US soldiers at Fort Riley were at that point preparing for deployment to the Western Front of World War I.
A month later, Spinney writes, “the flu was epidemic in the American Midwest, on the cities of the eastern seaboard from which the soldiers embarked, and in the French ports where they disembarked.”
There are other theories, however. Spinney told me in a phone call that one theory tries to explain the unusual virulence of the 1918 flu by positing that it developed first in Europe’s trenches. Normally viruses decline in their deadliness over time because they need living hosts to keep spreading. That didn’t happen with the 1918 flu — perhaps, Spinney says, “because the virus might have got started in the trenches filled with young men who weren’t very mobile. They were stuck in the trenches like sardines for days, weeks, months. There was no evolutionary pressure for it to moderate its virulence.”
Whether it began in the trenches or ended up there after the arrival of American troops, the virus spread quickly to German soldiers and to neutral Spain. News of the flu was censored in most countries with war censorship regimes, leading authorities in Spain to erroneously think that it was alone in enduring such a brutal outbreak — hence the name “Spanish flu.” Russian POWs returning from Germany spread the disease to the newly created Soviet Union, and by May and June, various countries in Africa, as well as India, China, and Japan, all had outbreaks.
This is sometimes called the “first wave” of the flu, because while it had significant effects (particularly on World War I, where it weakened troops on both sides), it was not the debilitating crisis that we now remember as the Spanish flu.
The second wave, Spinney writes, began in August 1918 almost simultaneously in ports in Freetown, Sierra Leone; Brest, France; and Boston, Massachusetts. British imperial ships spread it around the country’s African holdings, and from South Africa it spread to the rest of the continent. In a matter of months, the flu was slowing down combat in Europe, spreading back to India, China, and Japan, and circulating through mass public celebrations of the war’s armistice on November 11.
Basically the only place not affected was Australia, but a “third wave” of the flu in late 1918 eventually hit there, too.
The second wave of the flu, in particular, had more brutal effects than typical influenza, not least because it was likelier than the ordinary flu to be joined by bacterial pneumonia. This, subsequent research has suggested, caused most of the deaths in the 1918 flu outbreak.
When all was said and done, the flu had killed between 17 million and 100 million. That’s a wide range. While today countries keep detailed records of testing and diagnosis for new outbreaks (as they’ve been doing during the coronavirus crisis), there was no such record-keeping capacity in the late 1910s. As such, research has to rely on estimates comparing actual mortality to a “baseline” level of mortality that would have occurred without the flu. Accurate actual death rates are hard enough to cobble together, given the unreliability of death records from the period, but estimating counterfactual deaths without the flu is harder still.
It’s hard to say when, exactly, the pandemic outbreak receded. The third wave, beginning in the winter of 1918-’19, subsided by that summer, and the virus likely lurked around for years, not causing pandemics because most survivors had been exposed and developed antibodies. The long-term health and economic costs were substantial. Economist Douglas Almond has estimated that people exposed in utero to the flu in 1918-’19 received less education, earned lower incomes, and were likelier to have disabilities than people who missed the pandemic in the womb.
Similarities and differences between the 1918 outbreak and now
The Spanish flu is frightening because it demonstrates that in a reasonably modern society, a pandemic killing tens of millions of people is very plausible. But that “reasonably modern” society was still much more primitive when it came to medicine and public health than the world of today.
Here are a few facts about public health in the year 1918:
- We did not know that influenza is caused by a virus, and in fact the scientist Richard Pfeiffer had convinced most of the medical community that it was caused by bacteria; it wasn’t until 1933 that researchers proved conclusively that the flu is a viral infection.
- Antibiotics capable of treating flu-related pneumonia infections (which are typically caused by bacteria) were 10 years from being discovered.
- Antiviral drugs were many decades from being developed; the first came out in 1963.
- There was no World Health Organization, and efforts to surveil and track the outbreak of new diseases were incredibly rudimentary.
- Most countries in Europe were under war censorship regimes that limited the spread of accurate, lifesaving information about the flu outbreak.
For all the advances we’ve made since, what’s striking is how some of the measures authorities instituted at the time look very much like the ones we’re seeing with the coronavirus outbreak. Spinney told me, “They had the kind of social distancing measures that we’re still using today: isolation, quarantine, masks, hand-washing, staggering rush hour so you don’t have massive crowds in the metro and the streets. Those are techniques that are very ancient. People have always understood you have to keep the healthy and the sick separate.”
We must also weigh the massive strides in public health made since 1918 against the advent of global supply chains and passenger air travel. “We have a global population that is four times the size, and at least in the industrialized world, the populations are much older with respect to 1918, and old age weakens immune systems,” Spinney continues. All of that makes us more vulnerable, not less, to a pandemic like this. While antivirals are useful against coronavirus, we do not have a vaccine and will not for at least 18 months, somewhat limiting the public health value of our scientific advances over the last 100 years.
One way we can compare the two outbreaks is by looking at case fatality rates: the share of infections that lead to death. This is always difficult to estimate because there are likely more infections than have been identified by medical authorities. According to Johns Hopkins researchers, as of this writing there have been 111,363 cases of Covid-19 and 3,892 deaths, for a case fatality rate of about 3.5 percent.
But you should take that number with a grain of salt. Countries’ testing protocols vary widely and many, including the US, have only tested a few thousand people while others, like South Korea, have tested hundreds of thousands. That means the denominator for the case fatality rate — the total number of infections — is uncertain, and might be undercounted due to lack of testing. In South Korea, the case fatality rate is currently 0.7 percent, suggesting that better testing might yield more accurate, lower rates. At the same time, as Julia Belluz explains, accurately measured fatality rates are also going to vary significantly from country to country due to differences in health system capacity, low- and middle-income people’s access to health care, etc.
The case fatality rate of the Spanish flu is often cited as 2.5 percent, but this is likely a dramatic underestimate, as science writer Ferris Jabr has written.
The most frequently cited death statistics for the Spanish flu come from Niall Johnson and Juergen Mueller’s 2002 study, which estimated the death toll at 50 million and warned that this might be as much as a 100 percent underestimate, implying a total toll of 100 million. A more recent 2018 paper by Pete Spreeuwenberg, Madelon Kroneman, and John Paget gets a much lower estimate of 17.4 million. If the frequently cited estimate of 500 million infections globally is correct, then the latter death toll implies a case fatality rate of 3.5 percent, but using a higher death toll of 50 million, the fatality rate rises to 10 percent.
This is a huge range of uncertainty. Given how imprecise our counts of total global infections for both the Spanish flu and coronavirus are and how imprecise our estimates for the former’s death toll are, it is hard to say anything definitive about how they stack up against each other in terms of case fatality.
The diseases also differed in whom they infected. The Spanish flu, unusually for an influenza, was less lethal for older people, perhaps because a similar 1830s flu outbreak granted older people still alive in 1918 some limited immunity. The coronavirus, by contrast, has had its most devastating impact in China on older people.
Both Covid-19 and the Spanish flu have already had massive effects outside of their immediate health consequences. The Spanish flu, many World War I historians agree, sped up the end of the war by weakening the ability of each side to field armies; it may even have affected the outcome, though the evidence there is weaker.
Covid-19’s economic effects might outstrip the Spanish flu’s, even if the health effects turn out to be milder, due to the economy’s move toward in-person services, hospitality, and globalized supply chains, all of which are vulnerable to an outbreak like this. Indeed, one effect in 1918 that is not likely to be repeated is an increase in wages in some countries due to a shortage of workers.
It’s natural to want to compare the two outbreaks, and it can be responsible to do so if the comparison is done with care and nuance. But it’s important to keep in mind just how severe the Spanish flu outbreak was, and that while Covid-19 could get much worse, it would have to infect several thousand times as many people as it has to date to match the Spanish flu’s reach. The Covid-19 situation may only get that bad if we fail to adequately adopt measures like social distancing, aggressive testing, and quarantining, and let it get that bad.