A Few Lessons From 1918

Influenza interventions and impacts

For anyone who has been keeping up with current events, I am sure that you are aware of the emergence of a disease outbreak known as the Covid-19 coronavirus. Aggressive quarantine measures taken by China at first discovery have likely had some positive impact on current extent of spread, but even with those measures infections are now starting to be reported beyond their borders. Although not a definite outcome, it remains within the realm of possibility that large scale transmissions and infections could eventually develop. If such a scenario does play out, there will need to be policy decisions made at different levels of government on mitigation tactics — such as while a vaccine is developed or other treatment options validated.

The goal of this brief essay will be to highlight the findings of a 2007 study that investigated impacts of various metropolitan interventions for the 1918 influenza pandemic. This study, published as “Public health interventions and epidemic intensity during the 1918 influenza pandemic” by Hatchett, Mecher, and Lipsitch, was recently cited in an article offered by Nature, a respected scientific journal, which is how I came across. The study looks at historic levels of infection rates across metropolitan areas in relation to potential non-pharmaceutical intervention measures, and a discussion surrounding the findings of this study will be the sole focus of this post. The comparison to 1918 is not meant to imply an expectation of comparable severity for the current coronavirus, this is merely an illustrative point in history serving as the basis of the study.

Intervention Measures

The study looked at 19 intervention measures that were taken across 17 cities, here (%) indicates the 1918 adoption rate amongst those 17 cities. (Not shown here, the study also looked at timing of the measures based on coincident cumulative excess morbidity rate, an indication of stages of the epidemic for a city).

1. Making influenza a notifiable disease (88%)
2. Emergency declarations (24%)
3. Isolation policies (82%)
4. Quarantine of households where infection identified (29%)
5. School closures (82%)
6. Church closures (88%)
7. Theater closures (88%)
8. Dance hall closures (64%)
9. Other closures (76%)
10. Staggered business hours (to reduce congestion in stores and transit systems) (62%)
11. Mask ordinances (12%)
12. Rules forbidding crowding on streetcars (35%)
13. Private funerals (65%)
14. Bans on door-to-door sales (6%)
15. Interventions designed to reduce transmission in the workplace (0%)
16. Protective sequestration of children (18%)
17. Bans on public gatherings (88%)
18. No-crowding rules in locations other than transit systems (18%)
19. Community-wide business closures (6%)

Discussion

In some cases, these interventions were put in place in the first days of an outbreak detection in a city, in other cases measures were introduced late or not at all. It turns out the effectiveness of early intervention measures were materially impacted by timing of just a few short days — for example just two weeks from first transmissions could represent approximately 3–5 doubling times for an influenza epidemic. One of the findings of the study was thus that those cities who implemented mitigation measures in the earliest days of an outbreak demonstrated measurable improvement to outcomes.

Outcomes in the study were assessed based on a few metrics such as peak morbidity rate and cumulative morbidity rate. The most notable finding of this study was the material benefit to peak rate for those cities that implemented intervention measures, for example the median peak rate of cities who implemented three or fewer interventions was more than twice that of cities who implemented more.

It should be noted though that although still beneficial, cumulative excess morbidity rates were not as dramatically impacted as peak rates, meaning the dampening of peak rates by interventions resulted in a more prolonged progression of infections, with more risk for a heightened “second wave” of infections, such as when interventions were lifted (none of the cities in the study experienced a second wave of infections until after intervention measures were relaxed, and peak intensity of second waves were demonstrably lower than of first waves albeit with an inverse correlation to peak intensity of initial waves). Even though the cumulative rate was not as significantly dampened by interventions as the peak rate, it should be considered that capacity of a city’s medical infrastructure to support peak loads alone may justify taking early interventions. Higher peak rates may also result in an epidemic “overshoot”, contributing to increased cumulative. Dampening the peak rates with interventions did appear to shorten the window between an initial wave and a second wave.

In the interest of brevity, I’ll close with the simple restatement that the timing of a city’s intervention measures appeared to have a material improvement to peak rates, and that a matter of a few short weeks were significant for the potential spread. The intention of this post was merely to draw attention to a few of the key findings of this study based on outcomes for different cities in the 1918 pandemic.

References

Richard J. Hatchett, Carter E. Mecher, and Marc Lipsitch (2007) Public health interventions and epidemic intensity during the 1918 influenza pandemic

For further readings please check out the Table of Contents, Book Recommendations, and Music Recommendations. For more on Automunge: automunge.com