Quarantine is terrible, but it works

By Mark Buchanan

IN MARCH, many governments around the world imposed strict controls on human meetings and movement — “lockdowns” — to slow the spread of coronavirus. The pandemic’s curves were soon “flattened,” sparing health services from being overrun. Now, in China, Europe, the US, and elsewhere, authorities are cautiously relaxing the rules. Lockdowns have been costly, but they helped us avoid a far worse catastrophe.

Yet our very success — tenuous as it is — has suddenly stirred contrarian voices, with some pondering the possibility that the lockdowns may not have had much effect. Statistician William Briggs insists that “there’s no evidence the lockdown made a difference,” noting that the virus waned even in some nations — Taiwan, for example — that did not go into lockdown. An article in a right-wing British newspaper insists that there’s not even a “shred of real proof” the lockdown did any good.

None of this should be taken seriously. The position flies in the face of overwhelming evidence, and accepting it requires the commission of a litany of basic errors in reasoning.

To begin with, Briggs relies on a terrible statistical argument, citing the lack of a strong correlation between the number of coronavirus deaths in a nation and its imposition of a lockdown policy. The UK imposed a lockdown yet has still had the second-highest death rate per capita of any nation. No-lockdown Taiwan, in contrast, had very few deaths. On a graph of lockdown or no lockdown versus epidemic outcome, there’s no clear pattern, so Briggs thinks: “What can we conclude? Only one thing: We cannot conclude that lockdowns worked.”

Yes — if we only consider this kind of misguided statistical comparison. We shouldn’t expect any clear correlation, because we know that lots of other factors matter besides lockdown or no lockdown, including the timing of the lockdown and the presence (or absence) of an effective public-health service to carry out testing, contact tracing, and quarantining of the infected.

Take timing. Of countries taking similar measures, those that acted earlier were more successful in containing the epidemic. New Zealand, instituting lockdown when it had only 102 cases, has had 22 deaths. On March 16, the UK had 3,000 confirmed cases yet delayed lockdown for another week. This one-week delay, a recent study in pre-print estimates, caused roughly 28,000 further, avoidable deaths. Not really surprising, as the number of infected was doubling every three to four days at the time.

If lockdowns were effective, why did some nations without lockdowns have very few deaths? Obviously lockdown isn’t the only way to fight the spread of an epidemic. Some nations didn’t lock down because they didn’t need to. South Korea, Hong Kong, and Canada each had very-well-developed infrastructure for aggressive testing, tracing, and quarantining. They applied more precise anti-epidemic tools than others did.

If lockdown didn’t make any difference, then why, in all the nations imposing lockdowns, did the growth rate of the epidemic soon show marked reductions? Take China, the US, the UK, Spain, France, Italy — each shows the same pattern of sharp change in growth rate soon after lockdown. Is this just a weird coincidence?

There is one conceivable alternative explanation: that the epidemic slowed everywhere as enough people became infected and populations reached herd immunity. Exponential growth of a new epidemic does normally slow when so many get infected that the virus has a harder time finding new uninfected targets. But the evidence weighs heavily against this idea. Estimates in Spain and France find about 5% of their populations may have been infected by early May. The number is similar for the UKoverall, rising to perhaps 15% in London. This isn’t close to herd immunity, with almost everyone in the population remaining susceptible.

Aside from all this evidence, it’s also just very hard to imagine how a lockdown, which greatly reduces the frequency of human contact, could possibly fail to have an impact on the spread of an infectious agent. This is epidemiology in another universe. Indeed, scientists have clearly seen the effects of lockdown on this year’s flu in Northern Hemisphere countries. In January, we had expected this year to be among the worst flu seasons in several decades. But the number of flu infections dropped sharply in April, not long after COVID-19 lockdowns were imposed. Another weird coincidence?

It is hard to understand the motivation for “lockdowns had no effect” thinking. Of course lockdown is unpleasant. It has been socially and economically costly. But it never had any effect? That’s about as plausible as COVID-19 being caused by 5G microwave signals.

BLOOMBERG OPINION