Europe’s Lockdown Experiment

Elaine He’s piece at Bloomberg on the effectiveness of European countries’ lockdowns strikes me as quite important. I’ll try to summarize it for you. In terms of cubic inches much of the piece consists of three infographics: a heat chart indicating the stringency and onset of lockdown measures for 17 European countries, graphs illustrating excess mortality over time for 12 of them, and, finally, a scattergram illustrating the change in industrial production and, presumably, “economic pain” from the lockdowns for 13 of them.

Here are some telling passages from the piece:

While not a gauge of whether the decisions taken were the right ones, nor of how strictly they were followed, the analysis gives a clear sense of each government’s strategy for containing the virus. Some — above all Italy and Spain — enforced prolonged and strict lockdowns after infections took off. Others — especially Sweden — preferred a much more relaxed approach. Portugal and Greece chose to close down while cases were relatively low. France and the U.K. took longer before deciding to impose the most restrictive measures.

But, as our next chart shows, there’s little correlation between the severity of a nation’s restrictions and whether it managed to curb excess fatalities — a measure that looks at the overall number of deaths compared with normal trends.

On excess mortality:

In Europe, roughly three groups of countries emerge in terms of fatalities. One group, including the U.K., the Netherlands and Spain, experienced extremely high excess mortality. Another, encompassing Sweden and Switzerland, suffered many more deaths than usual, but significantly less than the first group. Finally, there were countries where deaths remained within a normal range such as Greece and Germany.

Yet the data show that the relative strictness of a country’s containment measures had little bearing on its membership in any of the three groups above. While Germany had milder restrictions than Italy, it has been much more successful in containing the virus.

The overall impression is that while restrictions on movement were seen as a necessary tool to halt the spread of the virus, when and how they were wielded was more important than their severity. Early preparation, and plentiful health-care resources, were enough for several countries to avoid draconian lockdowns. Germany, with better testing and contact tracing and more intensive care units than its neighbors, could afford to keep the economy a bit more open. Greece, by acting quickly and surely, appears to have avoided the worst, so far.

and here’s her conclusion:

The Covid-19 experience has taught us that it’s far better to respond quickly and smartly, with the right technology and mass testing and tracing, rather than only relying on the crudest of shutdowns. If there are second waves of the virus, we shouldn’t repeat the mistakes of the first.

I’m not sure I’d draw those conclusions from the data she’s presented. I think I’d be more likely to conclude:

  1. Lockdowns as such are not an effective method of reducing mortality but they are an effective method of reducing economic activity. It doesn’t matter much how early they were imposed or how strict they were. Lockdown measures as such do not reduce excess mortality. Despite imposing lockdown measures as early as February, Italy still suffered considerable excess mortality.
  2. Confounding factors are probably more significant in reducing excess mortality than testing or contact tracing. Confounding factors include demographics, the level of health of the populace, the robustness of the health care system, who knows how many other factors, and, probably dumb luck.

I continue to believe that we should be looking more closely at Portugal’s example. They imposed their lockdown measures later than most and, initially, rather lightly, they’re opening up earlier than many, and their excess mortality has been relatively low. Maybe Portugal’s success is just a matter of being small, comparatively isolated, and having high social cohesion.

5 comments… add one
  • jan Link

    The evidence from this article syncs with the one posted regarding Ioannidis’s studies. I also recall you, Dave, mentioning Farr’s Law much earlier, as being something to consider when generally reviewing the course of past pandemics.

    Finally, as reality rolls out, vanquishing unreliable models, and questioning panicked speculation these models caused, I tend to think it will be difficult to forgive those who churned the pot vigorously, calling for the implementation of lengthy, over-the-top, unnecessary measures, including business and school closures.

  • steve Link

    ” as early as February, Italy still suffered considerable excess mortality.”

    Excess compared with? Also, you have to account for compliance. There may have been some lockdowns somewhere in Italy, but they were only early if you are looking at eh calendar. They waited until they had thousands of documented cases and hit 100 deaths/day. Remembering the time lag that meant they already many more hundred, maybe thousands, who were going to die. Italy’s problem was that they waited too long and it was not taken seriously at first.

    “Lockdowns as such are not an effective method of reducing mortality”

    Not what we are seeing. I expect the same elsewhere.



  • CuriousOnlooker Link

    Or maybe Portugal did the small things turn out to be important.

    Here is an interview of Gov DeSantis from the national review on why he thinks Florida controlled Covid-19 much better then New York (contrary to many predictions).

    “Many states simply did not have the data infrastructure that Florida has,” says Mary Mayhew, secretary of Florida’s Agency for Healthcare Administration. “We have an emergency status system that gets stood up, as I mentioned, in the case of a hurricane. Hospitals and nursing homes and other long-term-care providers are required to submit data on a daily basis, twice-daily basis, regarding their bed availability.”

    “We advised, before there was even mitigation,” DeSantis points out, “if you’re 65 and older, stay home as much as possible and avoid crowds. And that was just something that made sense.” The state talked to senior communities like The Villages about what they were doing to mitigate risk…

    And it required staff to wear PPE. “We put our money where our mouth is,” he continues. “We recognized that a lot of these facilities were just not prepared to deal with something like this. So we ended up sending a total of 10 million masks just to our long-term-care facilities, a million gloves, half a million face shields.”

    “Every day on these calls,” she says, “I would hear the same comments and questions around, we need to get these individuals returned back to the nursing home. We drew a hard line early on. I said repeatedly to the hospital, to the CEOs, to the discharge planners, to the chief medical officers, ‘I understand that for 20 years it’s been ingrained, especially through Medicare reimbursement policy, to get individuals in and out. That is not our focus today. I’m not going to send anyone back to a nursing home who has the slightest risk of being positive.’”

    At the other end of the equation at the nursing homes, the state made it clear, according to Mayhew, “if you are unable to adhere to these infection-control standards, if you are unable to safely isolate and dedicate staff to an isolation wing or unit, you need to transfer that individual to a hospital.”

    When the state was seeing infections at nursing homes presumably caused by staff, DeSantis deployed what he calls “an expeditionary testing force,” 50 National Guard teams of four guardsmen together with Department of Health personnel that tested staff and residents.

    Finally, it has established several COVID-19-only nursing homes, with a couple more in the pipeline. The idea, again, is to get COVID-19-positive residents out of the regular nursing homes to the maximum extent possible.

    The National Review is conservative-leaning; so you have to decide how that colors the article.

    But I hope other (at least Republican) Governors take note. The focus on nursing homes is smart. Focusing on real time data instead of models is smart. Separating new nursing home cases from the general population as soon as possible and keeping them away until they are non-contagious is really smart (it is what every East Asian country would do).

  • bob sykes Link

    William Briggs has assembled some lockdown vs. per caput death data here,

    There is no obvious pattern to the data, and no evidence that lockdowns work.

    Of course, the rationale for lockdowns was that they would spread out the total number of deaths so as to reduce the work load on our hospitals. They were never intended to reduce the total number of deaths or cases.

Leave a Comment