dcyphr | Nonpharmaceutical Interventions Implemented by US Cities During the 1918-1919 Influenza Pandemic


This study aims to see how nonpharmaceutical interventions impacted the 1918-1919 influenza pandemic in 43 US cities. They looked at the excess death rate compared to different interventions taken. The cities that took quick and consistent interventions had the least excess deaths. Those who layered different types of interventions also had lower excess deaths. We can use this data to help plan effective nonpharmaceutical interventions in addition to developing pharmaceutical interventions.


The influenza pandemic in 1981-1919 killed about 40 million people worldwide and 550,000 people in the US. Many pandemic planners think that nonpharmaceutical interventions, like social distancing and business closure, will not prevent the pandemic. But, nonpharmaceutical interventions may lower the total number of cases, lower the peak of cases, and lower the death count. This study uses data from 43 US cities to see if specific nonpharmaceutical interventions of each city were more effective during the three pandemic peaks.


The researchers used the US Census Bureau's Weekly Health Index to assess 43 cities, all over 100,000 population size. They chose cities with most complete data. They also used all of the public health documents on nonpharmaceutical interventions of each city that were available.

Data Analysis

The study covers September 8, 1918 through February 22, 1919. 23 million people died from the 43 cities in this study, 22%% of the total US population. They assumed that each person contracted the flu 10 days before they died. The three main categories of nonpharmaceutical interventions were school closure, public gathering bans, and isolation/quarantine. The final category considered other changes like different work schedules, transportation restrictions, and mask wearing. But this category was hard to keep track of so it was not always included in the data analysis. Based on newspapers and records, this study classified cities as either activated (on) or deactivated (off) for nonpharmaceutical interventions on a given day. Cities were classified as activated when these interventions were enforced by law. This study measured weekly excessive death rates (EDR) for each city, based on the city having death rates about the nation’s average that week. The three main things they looked for were 1) the difference in time it took for the national average first peak and the city’s EDR peak, 2) the magnitude of the first EDR peak, and 3) the accumulative EDR over the entire 24 week period. They also looked at the public health response time (PHRT).


There were 115,340 total excess deaths in the 43 cities after 24 weeks. The most common interventions taken were combining school closure and public gathering bans. Layering the interventions was more effective in preventing excess deaths. Cities that had a faster public health response time had lower mortality rates. If the interventions were longer, this also showed lower mortality rates. This pandemic had three waves. The researchers found that the interventions taken over the first wave had nothing to do with the mortality rates of the second wave, and the same for the second to the third wave. The second and third wave had much higher mortality rates across the nation. None of they city’s size, density, sex distribution, or age distribution changed the mortality rates.


Time of activation, duration, and combination of nonpharmaceutical interventions were important in lowering the mortality rate of the 1918-1919 influenza pandemic. The US Centers for Disease Control and Prevention recommend implementing nonpharmaceutical interventions when the first case is confirmed in an area. Since growth of a pandemic is exponential, it is important to act quickly when implementing interventions. For example, New York City took quick and drastic measures, and ended up having the lowest mortality rates on the East Coast. Pittsburgh, however, had a couple week delay before implementing interventions, and they experienced the highest excess mortalities. The three peaks seen in the pandemic were often seen after the city loosened its nonpharmaceutical interventions. In fact, no city experienced a second peak until after the nonpharmaceutical interventions were lifted. 

History cannot tell us exactly what is going to happen, but we can study the 1918-1919 pandemic in hopes to be more effective against the current pandemic. The society of the 1918-1919 pandemic has some similarities to ours now, but things are still very different. Public support and compliance with medical guidelines has drastically increased in the past 100 years. People now have much better access to information and have been more educated about the pandemic. We have significant advances in medical technology that have also helped us.

Still, this study of the past reflects modern models of the pandemic, which only reaffirms that nonpharmaceutical interventions taken quickly and consistently have better outcomes. This study breaks successful interventions into 9 ideas: strength, consistency, specificity, temporality, dose response, biological plausibility, coherence, analogy, and experiment.