Review and meta-analysis of the effects of lockdowns on Covid-19 mortality [pdf]
sites.krieger.jhu.edu
sites.krieger.jhu.edu
"Overall, our meta-analysis fails to confirm that lockdowns have had a large, significant effect on mortality rates. Studies examining the relationship between lockdown strictness (based on the OxCGRT stringency index) find that the average lockdown in Europe and the United States only reduced COVID-19 mortality by 0.2% compared to a COVID-19 policy based solely on recommendations. Shelter-in-place orders (SIPOs) were also ineffective. They only reduced COVID-19 mortality by 2.9%."
But there is some nuance in there, worth reading.
That said, in April 2020 I assure you that economic activity in restaurants and non-essential businesses would have been severely disrupted regardless of government action. Most were already closed or take-out only.
There was no real lockdown in the USA, if you go beyond just looking at the mandates and take into account people's actual behavior. So I'm not sure how they review effects of a lockdown that never happened.
Not really true.
Hawaii closed parks, trails, and beaches. https://www.travelandleisure.com/travel-news/hawaii-coronavi... ' Honolulu, city parks were closed on Aug. 8 along with beaches and state parks and indoor attractions like bowling alleys and arcades. '
Colorado closed all ski areas. https://coloradosun.com/2020/03/14/jared-polis-executive-ord...
The stringency index should be aggregated as the minimum of all areas that allow for freedom of movement.
Importantly, the key benefit of this approach is that it allows you to open up in a controlled fashion areas that are clear of any outbreaks. So while the measures are most extreme and most costly up front, they ultimately are economically advantageous because they are of shorter duration.
Here in Lincoln, NE, everything was locked down to the point where open air parks had bright orange fencing around the playground equipment.
Grocery stores, Home Depot and Walmart were all that were open.
Walmart had a block long line, socially distanced of people waiting to get in. A Walmart employee was controlling the flow and ensuring masks.
The streets were empty. You couldn't go anywhere even if you wanted to.
And I dunno, Louis Rossman has a lot of videos up, NYC was definitely impacted.
Only the legislation does not support using that wording, so recommendation was chosen.
Now, we have evidence from other sources that the vaccine is definitely highly effective in every particular state (this should not require a citation, but here you go anyways: https://covid.cdc.gov/covid-data-tracker/#rates-by-vaccine-s...). But it seems there are too many confounding variables here (age, population density, etc.) such that attempting a simple region-by-region comparison over any particular time period might incorrectly conclude that the vaccine does not work.
Was JHU able to address that data collection and comparison problem here?
It's also interesting that lockdowns seem to be a wasted effort. Much suck for Ireland that imposed some of the worst lockdowns.
https://m.belfasttelegraph.co.uk/news/health/coronavirus/nis...
Thankfully all restrictions have just been lifted here in Denmark.
That sums it up nicely.
Good that some of Europe is opening.
Also, this study only measures impacts of lockdowns on Covid mortality, and doesn't include the human and financial costs of lockdowns.
https://91-divoc.com/pages/covid-visualization/?chart=countr...
The PDF here actually mentions New Zealand, as one of the studies is about it. But reading that, it doesn't compare New Zealand to other countries, but does within-country analyses (different parts of the same country, like US states, or the same parts at different times).
I am not an expert on this topic, but isn't it fair to say that New Zealand's massively lower death rate is due, in part, to lockdowns? Without a lockdown to get things under control things likely would have progressed as in almost all other countries.
With that said, I think it's possible that for most countries lockdowns might have a minor effect. If say California goes into lockdown but other states don't, for example, then the spread might reach it anyhow eventually. So I would bet that both things are true: lockdowns can be a critical part of a COVID response, even if on average perhaps they are not.
Regardless, all this is moot at this point. The early lockdowns were critical for reasons like US hospitals not having enough masks and other protective equipment, and vaccines were far off. Lockdowns bought us time - they were justified back then. The situation is obviously different now, and perhaps they are not justified today in most places.
(Again, caveats that I am not an expert on any of this.)
From the discussion section of the paper:
>We propose four factors that might explain the difference between our conclusion and the view embraced by some epidemiologists.
>First, people respond to dangers outside their door. When a pandemic rages, people believe in social distancing regardless of what the government mandates [...]
>Second, mandates only regulate a fraction of our potential contagious contacts and can hardly regulate nor enforce handwashing, coughing etiquette, distancing in supermarkets, etc. [...]
>Third, even if lockdowns are successful in initially reducing the spread of COVID-19, the behavioral response may counteract the effect completely, as people respond to the lower risk by changing behavior. [...]
It's plausible that new zealand was already predisposed to contain a pandemic well (because they're well isolated) and/or that their citizens are conscientious.
Point 1 is basically "lockdowns are unnecessary because people will behave in a safe manner regardless".
Point 2 is basically "lockdowns are ineffective because they don't stop all forms of dangerous behavior".
If people are willing to continue some dangerous behavior as part of point 2, why do we think they would stop other dangerous behavior without the lockdown enforcing that decision?
You have one group of Careful Carlas who always behave safely, lockdown or not. You also have a group of Reckless Rogers, standing too close to people in supermarkets and not washing their hands. If either of these two groups is too big, you won't see a positive effect from lockdowns.
Either one can ignore the lockdown and continue their usual behavior or one can't ignore the lockdown and their behavior changes.
The fact that they discarded 83 of 117 possible studies, including large notable studies that disagree with their results leaves a lot of room for selection bias. Plus they include non-peer-reviewed studies, despite being otherwise extremely prejudiced with their selections.
I think this is the kind of thing where you should wait for some peer review at the very least before drawing any conclusions, or sending it to the front page of hacker news.
I'm old enough to remember the financial crisis, when it was always assumed that things couldn't be as bad as they were because there's no way thousands of high paid bankers could all be so reckless and have it wrong at once. After all, these are the smartest people in the world, right? Surely they saw this coming. Surely they're not that foolish...
Well, it turns out they didn't see it coming and they were that foolish. It turns out they were all consumed by arrogance and, collectively, were far more incompetent than anyone was willing to give them credit for.
The inescapable fact is we don't know the long term effects of lockdowns.
Additionally, it would be very surprising to learn that lockdowns are ineffective given that covid spreads via human-to-human interaction, and lockdowns enforce a reduction in those interactions.
So, while I believe it's not impossible you are right, I would be very surprised.
"Lockdowns are defined as the imposition of at least one compulsory, non-pharmaceutical intervention (NPI). NPIs are any government mandate that directly restrict peoples’ possibilities, such as policies that limit internal movement, close schools and businesses, and ban international travel."
renders the exercise kind of useless because it conflates unsuccessful half-measures implemented in most countries with countries that followed lockdown and elimination strategies. For China, New Zealand and a few others which largely eliminated the disease the almost complete prevention of deaths proposed by the models discussed in the articles seems accurate.
And while I suppose you can write NZ off as a small island nation, China is the most populous country on earth. Yet it appears in the paper once in the introduction.
Calling masks a form of lockdown is utterly bogus.
And even if they had proven their point all they would be showing is that half measures don't work.
Simple thought experiment: If the world had basically been put on pause for April 2020--everyone stocks up in March, then spends the entirety of April at home where would we be now? (Use the stocks of PPE to protect those who have to be at their posts or people will die and they do not go home.)
(Now, I'm not sure if there was enough food and the like around to actually permit this.)
The places that eliminated CoVID-19 in 2020 did far better than those that didn't, both in terms of preventing death (no cases means no CoVID deaths) and keeping society functioning more or less normally. Once CoVID is gone, you can reopen bars, restaurants, etc. without having to worry about the virus spreading.
The main challenge after local elimination is preventing the virus from getting back in and reestablishing itself. The first line of defense is border quarantine. Behind that, you need ongoing surveillance (e.g., test everyone who shows up at a clinic with a fever or sore throat) to detect new outbreaks early. Once you detect an outbreak, you need to use tools like contact tracing (with mandatory quarantine and isolation), mass testing and, when absolutely necessary, local lockdowns to end it.
The question is why more countries didn't follow this strategy. For any country able to pull it off, it was obviously worth it. China reopened in April 2020. Most people haven't experienced a lockdown since, life has been mostly normal, and there hasn't been any large outbreak since (the largest one was about 2,000 people in total). The main cost is that international travel is difficult, because you have to quarantine for 3 weeks when you return, but that doesn't affect most people.
You are correct that it is not useful for evaluating the outcome for what didn't happen.
In an epidemic, there is a sharp transition between exponential growth and decline. There is also a fundamental difference between temporarily reducing case numbers and then letting them grow again, and reducing case numbers to zero.
Some countries, like China, New Zealand and Australia, competely eliminated the virus using lockdowns and other measures (like contact tracing).
Other countries locked down, but not hard enough to get rid of the virus. Then they reopened and let the virus spread like wildfire.
Treating these two strategies as the same is nonsensical.
The zero-CoVID strategy (China, NZ, etc.) is fundamentally different from the mitigation strategy (most countries). They both use lockdowns, but in very different ways and with competely different outcomes.
https://www.cbsnews.com/news/covid-in-china-cases-lockdown-o...
China is still locking down whole cities. Denmark is back to normal. Once Omicron gets into a country, that's it. Eventually, they will have to bite the bullet.
At the height of the "wave" this winter (which means about 200 infections/day), about 20 million people in China were under lockdown, out of a population of 1.4 billion (that's less than 2% of the population). That "wave" was successfully contained, and the major city that was under lockdown, Xi'an, reopened (the lockdown lasted 3 weeks).
The vast majority of people in China have not experienced a lockdown since April 2020.
> Once Omicron gets into a country, that's it.
China has already successfully controlled a few Omicron outbreaks. Omicron is a bit harder to control, because the serial interval (the time from one infection to the next in a transmission chain) is a bit shorter, but it's not fundamentally different. Your contact tracing has to be faster to control Omicron, and Chinese contact tracing is very fast. They identify all close contacts of an infected person within hours.
> Denmark is back to normal.
The vast majority of China has been back to normal since April 2020. The advantage of zero-CoVID is that it allows life for the vast majority of people to go back much closer to normal for the vast majority of the time. If your neighborhood has an outbreak or if you're a close contact of an infected person, you may have to quarantine, but outbreaks are few and far between in China, so these measures only ever affect relatively few people.
Danish people have been living with on-and-off restrictions for two years. The average Dane's daily life has been more restricted than that of the average Chinese person.
> Eventually, they will have to bite the bullet.
Not in the way that most countries have "bitten the bullet." Chinese public health experts have said that they will reevaluate the policy when drugs and vaccines render SARS-CoV-2 less deadly than the flu. We're not at that point yet. When/if China allows the virus to spread, it will be much less deadly, and China will not suffer the same sort of enormous death toll that the US and Europe have suffered.
People have been saying this for a year, with every new variant, yet China has been perfectly capable of keeping CoVID-19 under control and 98% or more of the country open. Contact tracing, mass testing and the other measures China uses work.
> And more importantly, why?
Because the risks of letting CoVID-19 spread in China are poorly known but potentially massive.
Looking outside China, the US is currently going through its 2nd most lethal CoVID-19 wave. There are more deaths per day now than there were at the height of the Delta wave.
It's prudent to wait and continue to assess the situation, rather than gambling with public health, as most countries are doing. China is able to do so.
> If very few people get seriously ill, why are economically destructive policies justified. I've felt that way from the beginning of the pandemic.
Well, first of all, China has had better economic performance than any other major country during the pandemic. One of the reasons is that it has spent far less time under lockdown or with restrictive social distancing measures than most countries. That's because it quickly eliminated the virus. International travel is affected, but factories can run normally, people can go to work in person, public venues are open, etc.
Second of all, if China had responded the same way as Europe and the US, it is likely that around 4 million people would have died in China. That's not "very few people."
This seems like a really weird take. The entire concept of a lockdown is solely about timing, since the whole purpose of the exercise is to reduce the R value in order to buy time for healthcare services.
By that part alone I'd call them a total failure: Healthcare didn't take advantage of it to prepare, instead the lockdowns just kept going on.
If anything the opposite has happened, with hospitals firing nurses and ending up with reduced capacity.
I've been living in Germany, where lockdowns seem to happen at the time you expect them to happen (when cases start to spike), and the deaths-per-capita are about half that of the UK, where they happened at the last minute (after cases had peaked).
In countries like Korea, Japan, and China, they had very early lockdowns, and have had basically negligible Covid impact.
I would really like it if everybody sat down and started consciously copying the states with approaches that worked, rather than trying to solve the problem from first principles. Public health is complicated, but what's not complicated is that some states are obviously way better at it, and these successes should be emulated.
I would want to see longer data periods that would tell me that early lockdowns slowed down the spread enough to identify better treatment options as well as allow more people to survive till vaccines were available in 2021. In other words, whether people who caught Covid in the fall 2020/winter 2021 had better survival odds than if they caught it in the Spring/Summer of 2020.
https://www.foreignaffairs.com/articles/china/2022-01-26/chi...
Also, while it worked well against the earlier strains they were already having some trouble with Delta and I think they're going to fail spectacularly against Omicron. It seems to be outrunning the ring fencing.
That's really huge.
Now, even if Omicron rips through the population, you will have a small fraction of the deaths that you would have otherwise.
I doubt that you would be very successful at attempting to bolt Americans inside their homes even in the bluest of blue cities and states.
The arithmetic average across the seven studies is -7.3% and the median is -2.4%, however the authors instead calculate a "precision-weighted average" where one study is weighted significantly higher than the rest (7390 vs 11-256 for the others, see Table 3 for details) and the number from that one study (0.1%) pretty much overwhelms all the others.
I don't think there is a reasonable explanation for doing the calculation this way unless you are intentionally trying to massage the numbers to support a pre-determined outcome. Very suspicious.
What about contact tracing? We're figuring out that masks, lockdowns, hygiene, etc. don't have that much of an effect. But where do people get COVID, shouldn't we have pretty good data on that now? It seems like a few superspreader events have a lot to do with it but that's the only anecdote that stuck with me.
Clearly you haven't looked at the death rates in Japan or Australia.
In that set you have 3 different "cultures" already. Plenty of controls.
There was never really an analogous situation in the USA since we maintained open borders to the virus (US citizens were allowed to come back from Europe even after the travel ban in March 2020, seeding the virus across the country.)
And in fact they would have been better off because society would not have incurred the monumental monetary costs of lockdowns and would have saved a very large amount of money that is now not going to be spent on the elderly.
All and all, pretty obvious lockdowns work if they’re actually enforced.
The high level of noncompliance we saw made it pretty much useless. Half measures cause economic pain without providing much benefit.
The careful people wear masks anyway, the reckless won't obey the protective measures.
Lots of people don't comply with speed limits but that doesn't prove speed limits ineffective.
Speed limits are a good counter-example though, where they can provide a baseline expectation (eg mask wearing is good in crowded places) and then begins the negotiations from there for what is expected social behavior.
For those who remember Chicken Pox, almost everyone got chicken pox. To the extend any NPIs reduced transmission (e.g. keeping kids out of school), the primary result was delaying chicken pox or avoiding major outbreaks, and delaying chicken pox may have been a bad thing.
As a different analogy, if you have a Windows XP system in the Internet, it will get pwned. If you update it to Vista, you get a worse UI and it still gets pwned. Sometimes half measures are much less than half effective.
I’m not claiming that NPIs are useless. I’m saying that half an NPI may be vastly less than half as effective. I think the OP can speak for itself.
Assuming all other things are equal in both cases, sure.
In reality though, by flattening the curve we have avoided greater collapse of health systems, we got more time to understand better what works and what doesn't when fighting the disease, we got more vaccinated people, etc. These things combined might have saved a lot of lives.
I'm not sure why delaying chicken pox may have been a bad thing, what do you mean by that?
Regarding chickenpox, it becomes more and more dangerous the later in life you catch it. Getting it as a 5yo is annoying. Getting it at 15 can be quite a bit more dangerous. getting it for the first time at 40 can be lethal.
How does that work out in reality though? There are always people in all those age groups in society and if you let the disease run amok, that just means the 40yo of now are more likely to catch it before there's a vaccine or better treatments, doesn't it? OTOH if you slow down the kids getting it, they will be less likely to spread it to the teens and middle-aged kids :D until those are available. i.e. pretty much the same, is it not?
Disclaimer: I'm pulling those ideas out of my arse, I have no real data or knowlege to back my thoughts.
Now we have a vaccine so I assume this doesn't happen anymore.
> I guess our parents had all had it, so there was no risk of getting it again.
This is not the case with COVID though. Everyone is getting infected for the first time, and we're still unsure how long and how strong immunity-through-exposure is. In fact, there are multiple reports of people catching it multiple times (I know of a person who had it 5 times :-o ).
Sorry, but not too long ago I ran into a figure that 2/3 of Omicron cases had previously been infected with another variant.
Covid is mutating fast enough that prior infection provides only limited protection. If you're lucky your body locks onto the spike protein and the protection works pretty well against all pre-Omicron variants. But you're not guaranteed that, your body might go for a much less stable target and do little against a variant.
We are seeing this with the Chinese vaccines. They're killed-virus vaccines, they worked well at first. Not anymore.
Reinfections are meaningless. What matters is severity of symptoms. Infection by earlier variants gives most recovered patients durable cellular immunity against Omicron.
I didn’t listen to the podcast, but I did read the notes, and I thought “wow, Omicron is no big deal! The data is great! Let me click those links!” Then I clicked the links and they didn’t seem amazing. Apparently previous infection reduces severity, and Omicron is less severe, but neither of these effects is anywhere near as strong as the show notes made them out to be.
This is false because COVID also has animal reservoirs. It will never get eradicated like polio or smallpox, which only had human reservoirs.
It appears that in 2022 it is now finally understood that COVID-19 will not get eradicated, the vaccines do not prevent transmission or infection (in fact, the original RCT never even showed it -- it was just assumed!), and that cloth masks have much more limited effectiveness than initially assumed. Omicron is highly contagious (but much milder!) so it appears that almost everyone will get it eventually.
I recall multiple studies early on stating cloth masks were not effective and should be avoided if you have access to surgical or n95 masks.
My impression is that the (US) government didn't want to nuisance this discussion as it'd only open further criticism of mask mandates.
Quick aside, polio hasn't been eradicated. Only smallpox and a cat virus I forget the name of.
An easy fallacy to handwave away the possibility that the lockdown strategy just doesn’t work so well. Which is kind of obvious if you think it through; either the result is you eradicate the virus completely or the lockdown just delays the inevitable. And it should be clear by now that eradication is simply impossible and from very early on in the pandemic was just never going to happen.
However, that strategy fails when you don't have access to the vaccine.
With booster shots it should never be an option to shut down society.