Tag Archives: moving the curve

COVID-19: Don’t Just Flatten The Curve, but also, do this….

Is Washington State leveling off?

Are mainly Republican states seeing the greatest increases in COVID-19 infections (outside of New York)?

Is the NOLA effect an explainer of the the distribution of COVID-19 outbreaks?

What is the difference between flattening the curve and pushing the curve, and why is the latter what we need to be doing?

These and other questions….

Washington State is especially interesting because until New York muscled it out of the way, it was a US epicenter of Covid-19 infection. Over the last several days, the percentage increase in cases in Washington look like this:

16
11
18
11
11
11
4

Before getting too excited about that “4” we must digress to examine the “ever trust the last datum rule” in epidemiology. If you have been following the progress of COVID-19, you’ll notice that on many day — most days, really — the current situation always looks a bit rosy because the exponential rise in new cases is less for the most recent reading than for the previous readings. This is almost always an artifact of the nature of the data. Ignore the last day.

Having said that, it remains true that over the last several days the state of Washington’s new case number has not gone up as a percent of total cases. Washington is still in trouble, but measures being taken there may be helping.

It has been suggested that Republcian run states are going to have disproportionately more trouble from COVID-19 than Blue states, because Democrats pay attention to science and Republicans spend their days punching hippies and making liberals cry. This characterization of Democrats vs. Republicans is pretty much unassailable, but the effect on COVID-19 right now is a bit more complicated. Population size and density and other factors probably matter more, and it is possible that all the different virus related factors come together in the New York City Metropolitan area (which, to the surprise of Federal health authorities, includes about 10-12 counties, not four as they have been saying) to make that a hot zone no matter what.

Otherwise, the data, as shown in the following graphic which has a smudge on each state with the most rapid recent increase in COVID-19 infection, speak for themselves.

Sometimes, when data speaks for itself, it mumbles.

Not shown on that graphic because things are happening too quickly is the sudden and dramatic increase in cases, and deaths, in Louisiana. It thought that COVID-19 was active in that state during Mardi Gras. Contagious carnivalians were literally parading around on floats throwing the virus (on beads and such) to innocent revelers. There might have been some other forms of exposure. Right now, this morning, it appears that Louisiana has the second highest infection rate in the US, second only to New York and Washington, but possibly rising at a meteoric rate soon so surpass everyone.

And, of course, all those people who went home after Mardi Gras took it with them. I want to see travel to and from NOLA and other carnival sites mapped against infection outbreaks. Globally; Mardi Gras is only one carnival of many.

Flattening the curve is a nice idea, but there are two problems with it. First, it is probably very difficult to do unless cases are truly isolated prior to multiple infections. The idea of flattening the curve is to reduce R-naught, the number of people, on average, that are infected per infected person. Social distancing can help, but the only way to make a huge difference is to identify ill individuals very early in the course of their infection, and take them totally away from society. Social distancing does not do that enough. One can somewhat attenuate the curve, but mere social distancing is not going to do what happened in South Korea, Singapore, and China.

Moving the curve is somewhat difference. This involves recognizing that a spike will happen (though maybe a lower one than otherwise), but one moves the spike about two to three weeks ahead in time. Why? Because a given region probably has about 1/10th (or maybe in better scenarios, 1/5th) of the ICU beds needed to save most lives of the critically ill.

Flattening the curve is, explicitly, making the maximum infection rate low enough to duck under the bare of ICU bed number. Like this:

If the curve flattens a bit but fails at this objective, it was not flattening the curve, but rather, failing and losing to the virus. Like this:

Pushing the curve to later, which would probably reduce the amplitude of the peak but mostly result in the same huge increase in number of cases, allows the build up of ICU bed number. Like this:

That is what we are doing in Minnesota. We expect thousands of people to require ICU beds, no matter what happens. We are partly locked down, and increasing the lockdown on Friday. We are building new ICU facilities, apparently, at a sufficient rate to handle the eventual need.

Communities that are just flattening the curve, and that expect it to work, may run into trouble if they are not building out infrastructure now. We can argue about what the best approach is, but trying different methods in different states or regions is a hell of a way to test a hypothesis.

I learned yesterday that about 15% of the ventilators used in the US are made in Minnesota. We have about 2% of the population. So we’re good.