The Journal Gazette
Tuesday, March 17, 2020 1:00 am

From a distance

Lessons of America's 1918 flu epidemic could save lives today

Christer Watson

It is impossible to discuss anything besides the virus now.

Good advice is common through trustworthy sources, such as our local health care professionals. Although I tend to treat anything on Facebook with extreme skepticism, I have seen an interesting graph that was frequently shared.

It showed the effect of the 1918 flu on St. Louis and Philadelphia. The difference was dramatic and, the posts argued, was evidence for the social distancing we are all doing now. I looked up the paper for a more in-depth read. As promised, there were some valuable lessons.

The study was published in 2007 and written by a group led by a scientist from the National Institutes of Health. One justification for this study, which now seems quaint, was the then-recent H5N1 avian flu outbreak.

The authors explain that mathematical models predict that social distancing, such as closing schools and canceling public events, reduces the fatality rate of epidemics. Experimental tests of this prediction, however, are limited.

The authors exploit a natural experiment during the 1918 flu. That flu was incredibly deadly. U.S. cities responded dramatically differently to the outbreak. The scientists analyzed how 17 different U.S. cities responded using 15 different criteria and the resulting differences in flu-related fatalities.

The graph I noticed, which is published in this study, shows the weekly fatality rate in St. Louis and Philadelphia during the final months of 1918. Philadelphia had a large spike of cases. St. Louis had a low level of cases, with a small spike in mid-December.

There was a general pattern in the United States that the flu outbreak started on the East Coast and spread westward. As a result, many Eastern cities were slow to respond with serious public health measures. More Western cities, by observing these problems to their east, were quicker to respond.

For example, Philadelphia had its first reported case on Sept. 17, 1918. Large public gatherings were allowed to continue. One key event was a citywide parade on Sept. 28. Social distancing, including school closures, started on Oct. 3.

St. Louis had its first reported case on Oct. 5. Social distancing started on Oct. 7.

The difference in impact was dramatic. Philadelphia had an excess weekly death rate peak at 257 (per 100,000 people). The total excess deaths were 719 (per 100,000 people). For St. Louis, those numbers were 31 and 347.

That is, the peak weekly death rate was much lower in St. Louis. The total deaths were also fewer. That is, the social distancing did more than just spread out the cases. The social distancing saved lives.

The more sophisticated analysis, involving all 17 cities and measuring when each city started its policies, show an effect of about 20%. That is, the total fatalities were decreased by about 20% because of the social-distancing measures. There is good reason to believe we can do even better than that today.

The authors observed that cities that experienced a small first wave of the flu – probably because of the interventions – tended to have a stronger second wave. The authors observe, however, that these second waves only occurred after cities removed their interventions. No city had a second wave during its social-distancing intervention.

St. Louis is again a good example of this effect. In St. Louis, the social-distancing restrictions were lifted between Nov. 7 and Nov. 9. The excess weekly death rate peaked in early to mid December, probably as a result.

The authors argue the total death rate reduction would have been better if intervention had lasted longer. They write, “In practice, and until emergency vaccine production capacity increases, this means that in the event of a severe pandemic, cities will likely need to maintain [social distancing] for longer than the 2–8 weeks that was the norm in 1918.”

Let that sink in for a couple of minutes.

I have read many people claim we need to evaluate events and reassess the situation day by day. That can sound reasonable, but I am deeply concerned these people are literally concentrating on the daily reported infections. Leaders in St. Louis appeared to do that exact thing in 1918, removed restrictions in November and weekly deaths peaked in December. We can do better.


Christer Watson, of Fort Wayne, is a professor of physics at Manchester University. Opinions expressed are his own. He wrote this column for The Journal Gazette, where his columns normally appear the first and third Tuesday of each month.

Sign up for our Opinion newsletter

Sent daily