I know it was rr but it's happened and is still happening that the reality. Figures and charts and so on can all paint a picture but the real human economic and social cost are felt and evident daily. The front cover seems to have taken everything I wrote into an article on the front page.. 😀
Its sad and depressing but a reality..hered is hoping the tide will be changing this year..
😀

Modellers should always bear in mind the first rule of modelling: all models are wrong, only occasionally are some less wrong than others. If this simple maxim were recognised a bit more by their protagonists perhaps their advice to our politicians would be rather more measured.

The actual quote is slightly different but more accurate.

All models are wrong but some models are useful..

All models are simplifications but some models give us a better understanding of the process and dynamics involved. The problem with models always start when the accuracy of the output from the model is claimed to be far greater than the actual accuracy of the model that created that output.

My favorite example is AGW models. Where the 100% certainty of the truth of the results by its creators is based on a model where the key magnitude value is calculated using equations that assumes the surface of the earth has no intrinsic temperature. It is 0 Kelvin. Which is quite interesting as its actually between 282K and 285K. Now the key value in AGW model is 290K (the median surface air temp) and the difference between this number and a theoretician 257K (black body temp). No accounting for the actual physical surface temp anywhere in the models. Just surface radiation. Turns out the AGW modelers never opened a geophysics text book. If they had they would have created a much more accurate full system model rather than the not very accurate partial system model used so far. But then they would have been far less likely to get the results they wanted. And needed.

Most model failures are due to using inaccurate partial system models rather than much more accurate and representative full system models. Which has been the track record for the various epidemiological models. Simplistic ones fail, more subtle and sophisticated ones are often quite useful.

So a general rule is that the more modest the claims the creator of a model, the less forthright they are in stating the result as being "the truth", the more likely the model is to have real value in helping to understand the question at hand.

Models are just tools, they are never the truth.

All models are wrong but some models are useful..
...
Models are just tools, they are never the truth.

"For whatever, and in whatever way, they think, the facts are ever other"
(The Buddha)

Oh good, so I was right then? lockdowns do alter the course of the disease. I said that in my original post: lockdowns work to delay disease progress, lifting the lockdown allowed disease to resume from the same place. Lockdowns have some effect, sure, that's what the plots show.

Yes, I would say that lockdowns have some effect on the dynamics. How significant that effect is and how it is manifest in the shape of the curve is, I think, not an easy task to figure out. I apologize if it seemed I was disagreeing with your general conclusion (although I wouldn't use the term "worked" here - it's too strong). I don't think the reasoning you have used (based on looking at peaks) is all that sound though.

Clearly some kind of brutal lockdown with everyone locked into their houses for a couple of months would drastically alter the course of the disease. Crudely, we might want to describe a lockdown severity parameter and set the brutal lockdown to have a value of 1 for this parameter. Doing nothing would have a value of 0. But what about a more partial lockdown that we might judge to have a severity parameter of 0.5 say?

I don't know - and I don't think this kind of thing is something that can be determined without some serious analysis. I might be wrong on this, but I would also expect the efficacy of a lockdown to be dependent on the susceptibility. In the first outbreak, where we had around 70% of the population susceptible according to some studies, a lockdown of severity 0.5 would probably have a bigger effect than the same lockdown now where we probably have something like 30-40% of the population still susceptible.

But either way, I think the 'efficacy' of a lockdown is going to be a non-linear function of this severity parameter (yes, I know it would be a crude thing to use a single parameter like this to parametrize lockdown severity).

I agree with your overall conclusions fon, just not totally in agreement with how you've arrived at them 🙂

Oh good, so I was right then? lockdowns do alter the course of the disease. I said that in my original post: lockdowns work to delay disease progress, lifting the lockdown allowed disease to resume from the same place. Lockdowns have some effect, sure, that's what the plots show.

Yes, I would say that lockdowns have some effect on the dynamics. How significant that effect is and how it is manifest in the shape of the curve is, I think, not an easy task to figure out. I apologize if it seemed I was disagreeing with your general conclusion (although I wouldn't use the term "worked" here - it's too strong). I don't think the reasoning you have used (based on looking at peaks) is all that sound though.

The first thing I did at the end of January after the Hong Kong Med School press conference about what was really happening in Wuhan was find some good text books on the subject.

This was a very good overview

https://www.amazon.com/Introduction-Mathematical-Epidemiology-Applied-Mathematics/dp/1489976116

This was a great book on the nuts and bolt of field models.

https://www.amazon.co.uk/Mathematical-Models-Epidemiology-Applied-Mathematics-ebook/dp/B07YZ3HC1F

It turns out that with air transmitted infectious respiratory diseases there are only two barriers to a standard random diffusion spread though the susceptible fraction of the population. Therapeutic treatments or a safe effective vaccine available at the start of the outbreak. Absolutely no other measures will have any effect on the eventual area under the curve. The total number of people infected. Measure like lock-downs etc just create pinch points, temporary changes in the slope of the curve, but once those pinch points are removed there is always an above the curve rebound then a return its its normal slope.

Very quickly what became obvious is that A) there are no effective therapeutic treatments for viral pneumonia, only palliative ones. The only treatment regime that has any measurable impact on severe viral pneumonia survival rates were antibiotic to reduce Hospital Acquired Infections while in the hospital. And B) all human corona-virus vaccines had failed in the past for very good virological and immunological reasons. There was not and could not be in the 1 to 3 years span of the pandemic any safe and effective vaccine for SARs CoV 2.

This would have been know by all the senior public health people familiar with the subject back in January. This pandemic was going to run its course no matter what draconian public health measures were taken. They would have zero effect on its long term course.

What would have also been know at the time, January, by those most familiar with the literature regarding the SARs CoV 1 outbreak in 2003 was that the quality of key epidemiological values for SARs CoV 1 were severely compromised. In fact the most important value of all when it came to the potential novel virus health risk made no sense at all. It was a purely arbitrary number. No basis in any verifiable data.

You should try playing around with even simple epidemiological models in R and you will see how introducing and removed impermanent physical barriers to diffusion like lockdowns has no effect on the final total of infections. But introduce permanent physical barriers like effective safe vaccines and effective therapeutic treatments and just watch the infection curve reduce quite quickly. This is very basic physics of any random diffusion process..

When the final tally is made in 3 to 5 years time I expect that it will show no measurable lives saved in the high risk group due to lockdowns etc. That the final toll of excess deaths due to severe viral pneumonia will be of the same order as H1N1-09 Swing Flu in 2009/2011 and much lower that with H3N2 in 1968/69. And that the number of excess deaths due to the severe dislocation of the medical system and the wider society due to government health policies will be at least three to five times greater (and probably a lot more) than the total of people who died directly from SARs CoV 2 infections.

So the SARs CoV 2 pandemic of 2020 (which was over by June) will turn out to be the greatest single peace-time public health disaster since 1918. But most of the excess deaths will be due to incompetent government polices in response to the initial outbreak rather than the actual primary pandemic. Totally unlike 1918. Or any other pandemic in the last 100 years for that matter.