Sunday, March 15, 2020

COVID-19 Update: Developing Countries and Flattening the US Curve


One of the key unknowns about this epidemic is the status of developing countries, for example those in warm equatorial regions such as in Africa and India. If we are fortunate, we will find that warm weather and humidity inhibit the spread of COVID-19. But this is not yet conclusively established.

In the video below Dr. John Campbell describes the situation in the Philippines (population 100M): hospitals are overwhelmed in Manila (population 20M) and drastic measures have been taken by the government. Given the lack of medical infrastructure, the mortality rate there could be well above 1% for cases. This would mean millions of deaths if the virus sweeps the population.

Even worse, we might expect the epidemic to play out similarly among billions of other people around the world. Perhaps climactic protection does not extend to Manila for specific reasons, such as the widespread use of air conditioning, the large malls and food courts, etc. (What is happening in Bangkok and Kuala Lumpur?) Or perhaps Manila is a harbinger of what is to come in many other countries.

The Philippines case study might give important clues as to the seasonality of COVID-19. How much relief will summer provide?


The calculations below are from an email I sent yesterday, describing what is required to "flatten the curve" in the US. The dotted line in the figure above can be interpreted as roughly 0.3 percent infected in any local population region (e.g., NYC or Seattle). For the US as a whole, this is ~1 million total infections in the country at any given time.
By my estimates the US health system will be overwhelmed if we have to deal with ~1 million cases (which is only 0.3% of the population) at any instant of time. This comes from counting free ICU space (~50k beds in US) and the observed fact that ~5% of cases require intensive care. See more here:

If the number of infected at any given time is higher than ~1M the mortality rate will be higher than the ~1% observed under good medical conditions. Even under good conditions, where the 1% applies: 50% of US population infected (i.e., some kind of sweep) means 1.5 million deaths! This could easily be 2 or 3 times higher if our health system is overwhelmed, as in northern Italy and Iran.

Now, assuming an ICU treatment period of 7 days (optimistic?), the amount of time required to deal with 1 million infected at a time on the way to an integrated 150 million (i.e., 50% infected) is 7 x 150 days = 150 weeks... about 3 years! Obviously during this period we would build more intensive care infrastructure, etc. but the point remains that flattening the curve so that we avoid the situation where people who need intensive care are turned away, causing death rate among cases to shoot well above 1%, will be challenging!

I think aggressive measures are absolutely necessary to avoid a catastrophic outcome. Whether, e.g., K-12 closures have a material impact on that is not completely clear to me, but I think we should probably err on the side of caution.

We have to reduce R0 drastically before hitting ~1M cases or we are in for mass panic. In February Italian cases grew by 1000x. That means we have only weeks to get things under control in US before we overwhelm our health system.
The main unknown: How effective are the measures now widely taken in the US (schools and companies moving to remote operation, expanded testing and quarantine) going to be in reducing the doubling rate from 2+ times per week (i.e., doubling time of few days, as observed in Italy in February) to something less rapid?

We are going to find out in coming weeks.

No comments:

Blog Archive