COVID-19 Introduction: The Pandemic with No Interventions

Our purpose is to help people understand the forces at work in a pandemic, both those that drive it and those that can mitigate it. The simulations in this notebook do not model a specific community. Instead, they explore several broad categories of communities to show how a pandemic affects them differently.

As of mid-March 2020, the only good data about COVID-19's death rate came from dividing how many people died of the disease by the number who were sick with the disease. The consensus view was that COVID-19 was a disease that would kill one percent or more of the people infected. It was a deadly killer that was spreading rapidly and had to be stopped at any cost.

Vulnerable individuals become very sick, with long hospital stays, resulting in a fear of overwhelming the hospital system (since played out in several major urban centers). In the United States there is only one hospital bed for every 357 people, so it is easy to see how a pandemic could cause a severe capacity problem.

In this early phase of the pandemic, the estimated percentage of those who were infected but asymptomatic varied widely, but centered around 25% (the number used in Graph 1). COVID-19 is believed to have a latency period of 3 to 14 days or more (the data on this also varied widely at the time). Due to the latency period, the graphs in the simulations show the number of individuals with symptoms lagging behind those with infections (the symptoms curve will be shifted slightly to the right).

We assume that infection gives a person immunity for the duration of the simulation, but this has yet to be confirmed. However, it seems like a reasonable assumption because those who recovered from the SARS virus had immunity lasting an average of three years.

The simulation uses a simplified economic model that assumes that people with symptoms cannot work, depressing economic output. See the FAQ for more details. When more than 20% of the population is sick at one time, as in this scenario, this simple model underestimates the economic impact.

Beginning in mid-April 2020, we saw the first studies based on serological testing of a random sample of the population for COVID-19 antibodies. Researchers aim to better understand how many people have been infected with COVID-19, regardless of whether they had symptoms. Studies of more than 7,500 individuals from Santa Clara County, CA, Los Angeles, CA, Westchester/Rockland, NY, Long Island, NY, New York City, NY, and Robbio, Italy, show that 70% to 98%+ of all COVID-19 cases are asymptomatic. (See the FAQ for source citations.) Given the worldwide scale of the disease, these studies represent a small sample, and researchers are debating many aspects of the tests. Nonetheless, it is clear that most people who are infected with COVID-19 won't even know it, and the death rate will be lower than originally estimated.

Due to the uncertainties, we will be conservative and use 80% as the default value of asymptomatic cases. As confirmation comes in, we will adjust accordingly. Of course, you can change this or any other parameter in the Sandbox at the bottom of this page!

Even though these numbers mean that COVID-19 is not the killer originally thought, it still is extremely dangerous among populations with some pre-existing medical conditions and has the potential to overwhelm the hospital system if not managed appropriately.

As in Graph 1, the simplified economic model underestimates the economic effects of this scenario.

In the Sandbox you can press the Customize Settings button to explore the effect of any parameter in the simulation! You can also save the results and share with others. For guidance on things to try, keep reading.

Some forces driving a pandemic can be altered with the actions we take as individuals and as a society. These are some of them. Change only one variable at a time so you can see the effect it causes.

• In Customize Settings—Disease Characteristics, reduce the "Basic Reproduction Number, R₀" from 3.0 to 2.0. Press Done, then press Run Simulation. Notice how decreasing R₀ makes the Infections curve broader and moves the peak out in time. This happens because each infected person transmits the disease to fewer individuals.
• In Customize Settings—Disease Characteristics, increase the "Basic Reproduction Number, R₀" from 2.0 to 4.0. Press Done, then press Run Simulation. Notice how increasing R₀ makes the Infections curve narrower and moves the peak inwards in time. This happens because each infected person transmits the disease to more individuals. Now change it back to 3.0.
• In the real world we can control R₀ using things like social distancing and personal protective equipment (PPE).
• To control a pandemic by reducing the number of individuals who seed it (the number infected at the very beginning), action must be taken extremely early. Travel restrictions are one possibility. In Customize Settings—Population Statistics, change "Number already infected (Individuals on day 1)" from 1 to 100 (this is an extremely large seeding, but we're doing it just to show the effect). Press Done, then press Run Simulation. Notice how the Infections curve increases because more seeds are starting the infection. Now change it back to 1.

We cannot control the intrinsic properties of the virus and the disease it causes. Unfortunately, many of the disease characteristics are poorly understood, so it's good to discern which of those uncertainties might cause us the most trouble as we formulate a response. This is one of them:

• Increasing the Duration of the Latent Period broadens and moves the infections peak out in time. You can modify this parameter in Customize Settings—Disease Characteristics, just like you did for R₀ above.
• Decreasing the Duration of the Latent Period narrows and moves the peak in.
• Explore other parameters on your own.

NOTE: If you press Re-run, you will notice (mostly) small changes to the output from run to run. This is because just like in the real world, many events are based on chance (e.g., who infects who when). See more in the FAQ.