An Optimistic Projection of COVID-19 Deaths [with comment by Paul]

The University of Washington’s Institute for Health Metrics and Evaluation is producing data and projections on the COVID-19 epidemic. I don’t know the basis of this group’s projections, but for what it is worth, they are optimistic. IHME expects COVID-19 deaths to peak in the U.S. on April 13, at 2,341. This is their “deaths per day” chart; click to enlarge. You can see the wide range of uncertainty:

Why is this an optimistic assessment? For two reasons. First, IHME expects U.S. deaths to peak soon and then begin to decline. That might reduce the hysteria that, in this instance, may be worse than the disease. Second, IHME projects total deaths from COVID-19 in the U.S. to be only 81,114, with not many additional deaths after around mid-May. Again, click to enlarge:

Does 81,114 sound like a lot of deaths? In a country of 330 million, it actually isn’t. In 2018, the CDC tells us that there were 2,839,205 resident deaths in the U.S. Eighty-one thousand would be less than three percent of that total. As I have written many times, just two years ago, according to the Centers for Disease Control, the seasonal flu virus killed 61,000 Americans while generating almost no publicity and zero government hysteria. If the University of Washington group is right, the Wuhan virus will be a little worse than that, but not much. If it kills 81,000, it will be around twice an average flu season.

Maybe there will be a silver lining to the Wuhan virus: it might focus attention on the main killers of Americans and others. According to the CDC, these are the principal causes of death in the U.S.:

As you can see, at 81,000 COVID-19 is nowhere near the top of the list. It represents an increase in the “Influenza and Pneumonia” category, I think. But these other diseases take their toll year after year, while COVID-19 will soon fade from the scene. Maybe in a year or two, sheepishness over the current overreaction to COVID-19 will translate into a more appropriate focus on these more important killers. Then again, maybe not: politicians are hard to embarrass.

PAUL ADDS: John’s list provides useful perspective on the projected 81,000 deaths estimate. However, it’s important to remember that the 81,000 number assumes the strong measures that have been taken to limit the spread of this virus in the U.S. — what John calls the overreaction.

Without these measures, the University of Washington would very likely be projecting a great many more than 81,000 deaths. In that scenario, the death total would move well up the list John presents.

We keep the number of annual deaths from the flu well below 81,000 through a vaccine. Unfortunately, we don’t yet have a vaccine for the Wujan coronavirus. Thus, officials are trying to keep the number of deaths due to this virus from rocketing into the hundreds of thousands through stringent social distancing measures.

JOHN responds: Paul makes a good point, but it isn’t so simple. Does the U. of Washington model take into account the various shutdown orders around the country, as well as many other measures that federal and state governments have taken? Does it break out the impacts of these measures, many of which I have endorsed, separately? Do any of the models do this? Does the U. of Washington model have a hypothetical alternative death count, as Paul seems to assume, for the counterfactual situation where there are no shutdown orders? Or where there are shutdown orders in all 50 states? Do they adjust the model whenever a new state orders a shutdown? Does this or any other model tell us what benefit would be gained from quarantining New York City, as the administration briefly considered? Not that I can find.

I intend to write about this later.

Notice: All comments are subject to moderation. Our comments are intended to be a forum for civil discourse bearing on the subject under discussion. Commenters who stray beyond the bounds of civility or employ what we deem gratuitous vulgarity in a comment — including, but not limited to, “s***,” “f***,” “a*******,” or one of their many variants — will be banned without further notice in the sole discretion of the site moderator.