Corona Virus Models and More Numbers

Corona Virus Models and More Numbers

I’m an old engineer.  Give me some numbers, and I’ll give you some estimates.  In addition to reported cases and deaths from the corona virus in the U.S., the world and localities of interest to me, I’ve been looking at the IHME (Institute for Health Metrics and Evaluation) models that Dr. Deborah Brix has referenced in the daily national corona virus task force meetings.  These meetings can otherwise euphemistically be known as “Trump Health Celebrity Apprentice”.

The IHME model is known as the Dr. Christopher Murray University of Washington model after its originator.  The First incarnations of this model at the end of March, were projecting 82,141 deaths in the U.S. by August 4th^t, with an uncertainty range of 39,174 – 141,995.  It was explained that these first models contained Chinese data of dubious accuracy, and also a lot of incomplete data from the U.S. and other countries.  All of these models were developed for hospital planning, and not to identify worst case scenarios.

As more and complete data became available in late March, the model predicted higher numbers of deaths through early August.  The April 1st model showed 93,765 predicted deaths by early August with an uncertainty range of 41,399 – 177,381.

IHME’s explanation: “Our model’s increase in nationwide deaths since the March 31 release is primarily driven by increasing death tolls in states that previously had very few COVID-19 deaths. States with more COVID-19 deaths, such as New York and Washington, show far less fluctuation across daily updates. If the number of cumulative COVID-19 deaths rise by state, we expect increasing accuracy for predictions across states.”

As additional and more accurate data continues to inform the model, and as the number of virus cases seems to be peaking and starting to flatten in some major areas like New York City and New Jersey, The IHME estimate of deaths in the U.S. by August has decreased.  The April 4 estimate was 81,766, and the April 7 estimate is 60,415, with an uncertainty range of 31,211 - 126,703.  IHME advise that their latest model includes better and more complete data, improved modelling techniques and incorporates the fact that the virus is peaking in some major areas.

Having bounced all of these modelled numbers around, it may be time for some historical perspective.  U.S. deaths from plain old flu was 61,100 in 2017-18, and 34,200 in 2018-19.  We’ve lost 12-18,000 people from swine flu in 2009, some 100,000 from influenza A or H3N2 in 1968 and 116,000 from Asian flu or H2N2 in 1957.  We did not shut the country down for any of these past pandemics.  We just “rode them out”.

But we could not “just ride this corona virus out” because of its high mortality rate, rapid human to human spread and lack of vaccines.  While mortality rates for both past influenza pandemics, and for the corona virus are somewhat statistically hazy, indications are that this corona virus is some 20 – 50 times more lethal than other recent influenza pandemics.  We could have lost over two million people if we did not curtail close-in activity, sheltered in place and socially distanced.  Even with this mitigation we are losing large numbers of people as the disease is at its apogee in some areas, but models like the ones being discussed are indicative of the tools we have to control this scourge.

What would really help our sheltering in place situation more is if the politicians and the mainstream media adopted a more positive attitude and stopped trying to tear down everything being done by the experts and the government to handle this thing.

Ray Gruszecki

April 8, 2020