Welcome to our community

Be a part of something great, join today!

Why We Need To End Lockdowns (at least in most places)

Klaas van Aarsen

MHB Seeker
Staff member
Mar 5, 2012
8,679
Incidentally, I have published a blog post (which by no means represents MHB) here on the MHB blog with some data analysis of the causal effect of lockdowns on COVID deaths.
I suggest to take a look at the graphs of other countries.
South Korea is in particular exemplary.
They did their lock down very early - and without all that much force - not like China.
Additionally, they immediately instated drive through tests everywhere, and an app to track where Corona was active.
The result is that they have very few deaths, and the daily new cases are down to zero.
To be fair, they had previous experience with SARS, so they were more ready than western countries.
The USA is pretty much at the opposite end of the spectrum: late to respond and not ready for such a disease.

Graphs and data are for instance here:
Oh, and here is a nice video that explains the effect:


It doesn't make sense to debate on higher levels when the disagreements are more fundamental. I am a Christian, I believe that the Bible is the inspired, inerrant, infallible word of God.
(snip)
Wait, wait. How does religion factor in here?
I thought we were talking about scientific reasons if, when, why, and how long lock down should be applied.
And what its impact is on lives lost, and on the economy.
The bible does not say anything about such things, does it?
 
Last edited:

Jameson

Administrator
Staff member
Jan 26, 2012
4,029
Ok I derailed this last week. Anything I post publicly here should be taken as just a user of the site. If I am going to use my admin or owner status to do something I’ll go another route. Maybe we can work on giving staff the ability to post without their badges when they want.

Anyway, what if we scope this into just the science of “is COVID as infectious as is claimed?” and “does social distancing help at all or less than is claimed?”. These type of questions. Political and religious stuff is way touchier but we can likely stick to the first part.

Who likes this?
 

Joppy

Well-known member
MHB Math Helper
Mar 17, 2016
256
Ok I derailed this last week. Anything I post publicly here should be taken as just a user of the site. If I am going to use my admin or owner status to do something I’ll go another route. Maybe we can work on giving staff the ability to post without their badges when they want.

Anyway, what if we scope this into just the science of “is COVID as infectious as is claimed?” and “does social distancing help at all or less than is claimed?”. These type of questions. Political and religious stuff is way touchier but we can likely stick to the first part.

Who likes this?
I don't think it is a particular bad thing to have discussion on touchy topics. However, the community is quite small here and it wouldn't help to have everyone getting out their pitchforks at every opportunity. 🔪(Evilgrin)
 

Jameson

Administrator
Staff member
Jan 26, 2012
4,029
Incidentally, I have published a blog post (which by no means represents MHB) here on the MHB blog with some data analysis of the causal effect of lockdowns on COVID deaths.
Hi @Ackbach.
  • Could you share the full Python code for the latest data cut? A site like Github would make it easier to digest.
  • From the blog post you are looking at daily state level data. State level is fine to start with but I have seen an extremely extensive project to model the COVID virus counts all the way down to the FIPS code level. State level loses a lot of possible nuances that can separate behavior within a state like my own - Georgia. Atlanta is very unique to the rest of the state in its demographics and population density. Do you have any thoughts on this?

  • Regardless of granularity, I believe you are making the claim that the daily new infection counts and daily deaths do not show signs of benefitting from "social distancing". There are other comments as to why but the data science part interests me more for now. I won't try to box in what you should say, but I think that you are claiming the data does not demonstrate that social distancing "flattens the curve" as we have been told it does. For the sake of avoiding ambiguity, "flattening the curve" is meant to convey the idea of changing the infection rate distribution from very skewed right and having a very front loaded shape to something closer to a uniform distribution. This may or may not change the total deaths but this cumulative death count is no the main idea, it is about the timing of the deaths. The argument is that hospitals have finite resources and if overrun, they would not be able to handle all infected patients which would lead to even more deaths. Keeping hospitals below capacity should allow for an overall better quality of care and monitoring. This is the argument as I have heard.
  • I find your approach with the data you found interesting, but unless I'm missing something big I do essentially disagree with the main premise. You are estimating the 1st and 2nd derivatives for a cohort's death count over time. This isn't a widely used technique for time series modeling but it's interesting and has merit I think, so no issues from me with method of approximation. Using your results I think this is your major conclusion that is the crux of your argument -
    • The derivatives before and after social distancing requirements took effect show that the death rate did not slow down. It might slow down the higher order changes but this is not immediately provable.
    • Furthermore, a true "test" of the effectiveness of social distancing would be to somehow have the same location try both paths and compare. This is not possible unfortunately so we have to do our best to find two areas that are good proxies for each other. When you do this you do not see evidence of social distancing leading to significant differences in the death rate.
Ok, please correct me above if I'm not understanding you correctly. Without confirmation that nothing is hugely wrong I don't want t can't reliably know that continuing is fair. So I'll pause a while and wait for your comments. Until then I can say independent of this topic that it's so easy to talk about different aspects of COVID and actually be debating two different topics but think that both sides mean the same thing. This is why setting the ground assumptions and terms helps I think. As an example, on the effectiveness of social distancing we could look at:
  • How diseases spread typically and the validity of modeling with SIR type methods. For these the parameters are crucial, one being the infection rate. When this rate is >1, the spread becomes exponential. When between 0 and 1, it will naturally dampen. This can be bad or not that bad depending on the medical nature of infection.
  • Many propose that social distancing lowers this rate to something still larger than 1, but lower than without intervention. If true, this would be an argument for the concept working.
  • There are medical arguments that can be made I'm sure, but I have no expertise to be able to really speak on this. I hope we all agree that COVID is either infectious or its not, and if it is then there are properties about it that are very helpful in understanding its ability to spread.
Last sentence. What do you or anyone else here reading my long post think is the ideal test for social distancing's effectiveness? With the constraints of reality and not being able to run the same scenario with just this one change and watch, what method(s) are best and remove the most human biases?

@Ackbach - If helpful, I can summarize this into 3-4 sentences and only keep the truly essential parts. I don't want it to seem lost in details.
 

Ackbach

Indicium Physicus
Staff member
Jan 26, 2012
4,183
Reply to Jameson at #29:

Very thoughtful reply. Here are my thoughts:

1. I really was interested in COVID deaths, and not the least bit interested in COVID cases. Other people might have been interested in cases, or flattening the curve, but I can't get excited about healthy people getting the virus for a week or two and recovering. That really is no different from the flu or a cold. Moreover, the vast majority of hospitals in the US haven't even been close to being overrun, so that the need to flatten the curve seems non-existent. I do get excited about people dying from COVID, though. If there are reasonable prevention methods possible, then the people of the United States should volunteer do those, and they should not be mandated by the government whose job it ... isn't.

2. My interest in the whole thing tanked when I saw this post. Those graphs tell me that the lockdowns didn't change anyone's actual behavior, and therefore could have had no causal effect on anything at all. Except the economic downturn, of course. Effectively, it makes my entire blog post moot! I just haven't bothered to post an update or to take the blog post down.

3. With everything loosening up, that's just another reason to lose interest. With the whole George Floyd thing, and the resulting catastrophes in Minneapolis - somewhat close to home for me - my focus has been elsewhere.

Obviously, there's still lots of work to be done. My institution, Mayo Clinic, is ramping back up appointments in a major way, and we aim to be back to 100% by July. Hopefully we can make up for some lost ground, and get people in for healing.
 

Ackbach

Indicium Physicus
Staff member
Jan 26, 2012
4,183
I took down the blog post because, as I mentioned, I think it's all moot. The lockdowns accomplished nothing at all beyond destroying a lot of businesses and a lot of jobs. I find the whole thing rather amusing in one way: the American people were ahead of the government in fighting the virus every single step of the way, and the government has done nothing but get in the way. So much for the government solving problems.