A New Analysis Predicts 1.1 Million Coronavirus Deaths in a Medium Case Scenario

After weeks of downplaying the severity of the coronavirus pandemic, why did President Trump suddenly sound so serious about it on Monday? Everyone’s best guess is that he was briefed on a new study from Imperial College in London, which estimated the death rate from the pandemic if no control measures are taken:

In such scenarios, given an estimated R₀ of 2.4, we predict 81% of the GB and US populations would be infected over the course of the epidemic….In total, in an unmitigated epidemic, we would predict approximately 510,000 deaths in GB and 2.2 million in the US, not accounting for the potential negative effects of health systems being overwhelmed on mortality.

That’s enough to put the fear of God into anyone. If no control measures are put in place, the researchers estimate that we’ll exceed critical care capacity about 25 days from now and see a peak of about 50,000 deaths per day by early June.

So what should we do? The Imperial College team modeled two basic strategies: suppression and mitigation. In both cases, the goal is to reduce the number R₀ (pronounced “R nought”), which tells you how many people are likely to catch the virus from someone who already has it. The team assumes that the virus, if uncontrolled, has an R₀ of about 2.4, which means that everyone who’s infected passes the virus along to an average of 2.4 additional people. This is what causes the exponential growth that you see in the chart.

In the case of suppression, the goal is to get R₀ below one. Obviously, if each infected person passes the virus along to, say, an average of half a person, you get an exponential decline: 1,000 cases quickly becomes 500 cases, then 250 cases and so forth. The problem here is threefold. First, it would take massive—and probably unrealistic—control measures to get R₀ so low. Second, the economic cost of full suppression is likely to be very large in the long term. Third, the control measures would have to be in place for the entire population for well over a year. Until then, if the control measures fail at any point, the epidemic will almost immediately come roaring back because no one has developed any immunity to it.

Mitigation is more realistic, though it still requires plenty of social change. The goal is not to stop the spread of the virus completely but to slow it down. The downside of mitigation is that since it does less to reduce the spread of the virus, it’s a certainty that at some point the health care system will become overwhelmed. Also, a lot more people will die.

The Imperial College team concludes that suppression is the only “viable strategy,” but the “social and economic effects of the measures which are needed to achieve this policy goal will be profound.” Mitigation, conversely, is feasible only if our critical care bed capacity is increased by at least eight fold. And even at that, about 1.1 million people would die in the US. Here’s the inevitable chart:

If there’s any good news in the report it’s this: if we did immediately adopt a maximum suppression strategy, using the most intense versions of each social intervention and keeping them in place for about 18 months, we could reduce deaths by 95-99 percent—though even that depends on having an effective vaccine in place by then. Personally, I consider this so unlikely that it’s not really worth pretending we can do it.

Beyond that, I’m no expert and can’t pretend to have an informed opinion about all this. However, I can say this much: if the Imperial College team is right and the choice is between (a) immense social changes starting instantly and being kept in place for 18 months, and (b) building a huge number of critical care beds fast, I’d say the construction project is actually more feasible—though only if it starts now and is given about the same urgency as the World War II Manhattan Project.

Most likely, the best we can do is to combine a strong but reasonably practical mitigation strategy—in which people are informed that this is the new reality for at least several months—with the immediate, emergency construction of something like 200,000 temporary critical care beds—or the closest we can come to that. And we need to be prepared for a large number of deaths even if we succeed.

Alternatively, we can hope the Imperial College team is wrong.

POSTSCRIPT: One other thing the team says is that it takes a while for mitigation strategies to have an effect. The stuff we’re doing now will start to reduce infection rates in about 2-3 weeks, but before that the growth rate of the virus is pretty much set in stone.

POSTSCRIPT 2: Also of interest is that the team’s model suggests that stopping mass gatherings has little impact “because the contact-time at such events is relatively small compared to the time spent at home, in schools or workplace and in other community locations such as bars and restaurants.” It’s much better to focus on case isolation (widespread testing combined with isolation of detected cases), social distancing (either of the entire population or only of those age 70+), home quarantine (i.e., quarantining the entire family of anyone infected), and school closures.