Putting coronavirus in perspective: A numbers comparison with SARS, Ebola and flu

By Justin Fox

(Bloomberg Opinion) – The coronavirus outbreak has been turning a lot of us into amateur epidemiologists. Just listen to Mick Mulvaney, the former real estate developer and member of Congress from South Carolina who is now acting White House chief of staff.

“The flu kills people,” he said last week. “This is not Ebola. It’s not SARS, it’s not MERS. It’s not a death sentence, it’s not the same as the Ebola crisis.”

All those statements are true. The flu does kill people: an estimated 61,099 in the US in the worst recent flu season, that of 2017-2018. People who get Covid-19, the World Health Organisation’s shorthand for Coronavirus Disease 2019, are much less likely to die than those infected with Ebola, the Severe Acute Respiratory Syndrome of 2003 and the Middle East Respiratory Syndrome first reported in 2012. And no, this is not the same as the Ebola crisis.

It’s not the same as the 2014 Ebola crisis in part because it appears to be a much bigger deal for the US and other countries outside of West Africa. As Microsoft co-founder Bill Gates, also technically an amateur epidemiologist but by this point quite a well-informed one, put it Friday in the New England Journal of Medicine: “Covid-19 has started behaving a lot like the once-in-a-century pathogen we’ve been worried about.”

How do we reconcile these differing views of Covid-19? Well, I too am an amateur as an epidemiologist, but I find that charts and (very simple) equations help me think through things. On the theory that others might find this helpful, too, let’s start with the approximate case-fatality rates for the diseases listed by Mulvaney and a few others you may have heard of.

These fatality rates can change a lot depending on time and place and access to treatment. The Covid-19 rate is obviously a moving target, so I’ve included both the 3.4% worldwide mortality rate reported this week by the World Health Organisation and the 1% estimate from a study released on February 10 by the MRC Centre for Global Infectious Disease Analysis at Imperial College London that factored in probable unreported cases. The authors of that study also said that, given the information available at the time, they were 95% confident the correct fatality rate was somewhere between 0.5% and 4%. Gates used the 1% estimate in his article, and when I ran it by Caroline Buckee, an actual professional epidemiologist who is a professor at Harvard’s T.H. Chan School of Public Health, she termed it “reasonable”.

In a context that includes Ebola and MERS, the Covid-19 death rates are much closer to those of the flu, and it’s understandable why people find the comparison reassuring. Compare Covid-19 with just the flu, though, and it becomes clear how different they are.

The 61,099 flu-related deaths in the US during the severe flu season of 2017-2018 amounted to 0.14% of the estimated 44.8 million cases of influenza-like illness. There were also an estimated flu-related 808,129 hospitalisations, for a rate of 1.8%. Assume a Covid-19 outbreak of similar size in the US, multiply the death and hospitalisation estimates by five or 10, and you get some really scary numbers: 300,000 to 600,000 deaths, and 4 million to 8 million hospitalisations in a country that has 924,107 staffed hospital beds. Multiply by 40 and, well, forget about it. Also, death rates would go higher if the hospital system is overwhelmed, as happened in the Chinese province of Hubei where Covid-19’s spread began and seems to be happening in Iran now. That’s one reason that slowing the spread is important even if it turns out the disease can’t be stopped.

Could Covid-19 really spread as widely as the flu? If allowed to, sure. The standard metric for infectiousness is what’s called the reproduction number, or R0. It is usually pronounced “R naught,” and the zero after the R should be rendered in subscript, but it’s a simple enough concept. An R0 of one means each person with the disease can be expected to infect one more person. If the number dips below one, the disease will peter out. If it gets much above one, the disease can spread rapidly.

Here are R0s for the same list of diseases as above. These are rough approximations, in most cases the midpoints of quite-large ranges. But they do give a sense of relative infectiousness.

This helps explain why public health authorities want everybody to get vaccinated against the measles. It’s not all that deadly a disease, but once it gets going in an unvaccinated population, everybody gets it.

The numbers also seem to indicate that Covid-19 is a lot more contagious than the seasonal flu. Average R0 isn’t the whole story, though. Why all the worry about MERS, for example, which with an R0 below one shouldn’t spread at all? Well, it’s extremely deadly, its R0 can rise above one in certain environments, among them hospitals, and … you can catch it from your camel.

Then there’s SARS, which is both deadlier than Covid-19 and has a similar R0. Why was it wiped out in about a year, while some experts warn that Covid-19 may be around forever? Because SARS usually didn’t become contagious until several days after symptoms appeared. This meant that, as four British infectious disease experts wrote in 2004, “actions taken during this period to isolate or quarantine ill patients can effectively interrupt transmission.” They proposed adding another variable to disease-transmission models: the proportion of transmission occurring prior to symptoms. For SARS, this was less than 11%, probably much less. For influenza, it is between 30% and 50%, making it far harder to control the disease’s spread.

With Covid-19, “it seems that it can transmit quite a bit before symptoms occur,” Buckee says. How much is still up in the air. World Health Organisation Director-General Tedros Adhanom Ghebreyesus has been arguing this week that pre-symptomatic transmission appears to be low enough that Covid-19 can be controlled in ways that the flu cannot. “If this was an influenza epidemic, we would have expected to see widespread community transmission across the globe by now,” he said on Monday, “and efforts to slow it down or contain it would not be feasible.”

To understand how the spread of such a disease can be contained, it helps to break R0 down to its constituent parts. A simple formula that I got from Buckee is:

  • the probability of infection given contact with an infectious person (b), multiplied by
  • the contact rate (k), multiplied by
  • the infectious duration (d)

In some cases you can shorten the infectious duration with treatment. Quarantining people once you know they’re infected effectively shortens it, too. Variables b and k, meanwhile, are clearly dependent on behaviour. The probability of infection is reduced by things like frequent hand-washing, replacing handshakes with fist bumps and such. The contact rate is reduced by staying home. By putting much of the country on lockdown, Chinese authorities reduced the contact rate enough that Covid-19’s R0 in the country fell below one. They also incurred huge economic and social costs. Now, as China begins to go back to work, the big question is whether a less-draconian approach can keep the disease in check or whether it will just start spreading again.

That’s the big question in the US, Europe and pretty much everywhere else on earth too. It can’t be answered entirely by professional epidemiologists, either. Weighing whether the costs of a particular intervention are worth the benefits is at heart a political decision. So it’s actually good that politicians are moonlighting as amateur epidemiologists. Some of them may just need to study a little harder.

Visited 4,297 times, 2 visit(s) today