Sam Lowry said:
ATL Bear said:
Sam Lowry said:
ATL Bear said:
Sam Lowry said:
ATL Bear said:
Sam Lowry said:
ATL Bear said:
Sam Lowry said:
ATL Bear said:
Sam Lowry said:
ATL Bear said:
Sam Lowry said:
ATL Bear said:
Sam Lowry said:
ATL Bear said:
Sam Lowry said:
But again, the triggers only apply to social distancing and school closures. "Other policies are assumed to start in late March and remain in place."
Those are the only other variables/recommendations outside of the infected and immediate family stay at home recommendations. %A0Those were definitely in place and only until after mass vaccination were those eased, which was also the study recommendation. %A0And again, the model accounted for a 75 (infected) and 50 (family) percent compliance. %A0
We're talking about government mandates, not just recommendations. Where did you see these in place?
https://www.instituteforgovernment.org.uk/charts/uk-government-coronavirus-lockdowns
I'm sorry, but that's not very informative. You're showing me about 30 times when different restrictions were imposed or lifted. Which ones were in place continuously? To review the specific policies in question:
-Mandatory isolation of suspected cases
-Mandatory quarantine of families
-Mandatory distancing of entire population 2/3 of the time
You're arguing something I'm not, and you're applying a standard that isn't part of the model. %A0These were recommendations not required government mandating. %A0But the government did mandate many of the recommendations, and the overreaching guidance on isolation was always there. %A0
That's a timeline of institution of specific policies starts, easing of strictness and restarting of many of them. %A0Some form of social distancing was in place all the way until mid 2022. %A0During the first 18 months of the pandemic most indoor gatherings were banned and then limited to small sizes. %A0Schools were face to face for only 3 months from April 2020 through April 2021. %A0So as I've said, the recommendations were in place and many directly government mandated. %A0The model didn't say zero contact, it forecasted the impact of a reduction outside of work and school. %A0
I'm applying the standard that is part of the best-case scenario. It's not zero contact, but it is continuous. Banning mass gatherings wasn't projected to make much difference. As it turns out, neither did school closures. The model assumes that children transmit to each other at the same rate as adults. We now know that isn't true, but there was no way to know at the time. I think lockdowns made a difference--we've disagreed on that before, so I'm glad you changed your mind--but they were immediately politicized and sporadically applied.
And I'm utilizing their worst case scenario with the recommendations they suggested, and they missed by a ton. %A0If I used your best case model they would miss by a factor of over 10! %A0Hence back to the original issue of a bad model used as justification for policies that didn't work (like school closings). %A0
And I didn't say lockdowns don't work. %A0I said lockdowns in shared living spaces with common air systems aren't going to work with a respiratory virus of this type. %A0If everyone could hunker down in their own isolated home, sure that's going to work. %A0But of course there are other considerations around something like that. %A0
The UK didn't mandate case isolation until late September 2020. The reason for the mandate was a study that showed a compliance rate of 18% (as opposed to the 70% assumed by the IC model). This is particularly important because the success of the suppression model depends on early implementation. So I just can't agree that all measures were in place as intended. As for school closings, you can't run a model without assumptions. Some of them will always be wrong. That doesn't mean the model is bad.
Self isolation (infected) and close contact quarantine was part of the guidelines from the beginning, so not sure what you're referencing, but after codifying it deaths and cases spiked the most over the next 12 months, so improved compliance didn't really matter. %A0 %A0
Models are judged by their proximity to forecasted outcome. %A0No one expects exactitude, only a reasonable delta range. %A0Even if it was off by 50% or even 100%, you might say it was a pretty fair estimate, and the recommendations were appropriate. %A0But this was 500 - 1000%+ off, and the predictions were aided by other measures not considered like mask mandates, business closures, and vaccines. %A0It missed badly. %A0
And that's just not how any of this works. The whole suppression strategy was based on quickly getting the reproduction rate down and maintaining it there. You can't assume that tweaking the policy six months later is going to magically reboot the whole pandemic and put you on the same track as before. That does not mean improved compliance doesn't matter. It just doesn't work retroactively.
The policies were in place from the start. %A0If they worked to slow infection and death it wouldn't get worse AFTER the measures were tightened. %A0And I don't mean a little worse, it got monumentally worse. %A0There were 340,000 COVID cases in the UK when the government mandated self isolation and quarantine. %A0Six months later there were over 4 million cases, and that's pre-Delta, which got worse. %A0Simple logic would tell you an equal or moderate increase case trend would follow improved suppression efforts, %A0The opposite happened. %A0
You're not thinking geometrically. If it takes six months to add 340K, that doesn't mean it will take another six months to add another 340K. The effect of interventions is on transmission rates. If you're on a steep rise to begin with, you can curb it significantly and still see a large increase. And governments typically give these orders when hospitalizations and deaths are already rising, as in this case.
Yes, they measure transmission rates through daily new cases. %A0You might see a spike rise, then a plateau and perhaps a decline. %A0Prior to the government mandate on isolation, they were concerned about the increase from 8-900 cases a day jumping to 2000 that moved to 3000 shortly after implementation. %A0Two months after mandates they were running over 20,000 per day, and 4 months in they were over 40,000 per day. %A0Not only was there no impact to transmission, rates got exponentially (geometry) worse.
A spike rise, plateau, and decline is a good description of what happened. It was followed by another spike in December due the Alpha variant, which was 40-80% more transmissible. Also, the relevant metric in terms of the model's predictions is deaths, not cases. Despite the dramatic rise in cases during the second wave, peak deaths per day were only a few hundred above the first peak in April.
C'mon Sam. Deaths increased exponentially (compared to predictions) and I used that, but you said we needed to look at transmission rate, not deaths. So I showed where transmission rate increased exponentially, and now you're back to deaths with the argument, well cases increased, but the daily death rate wasn't as high and the virus was more transmissible. I'm sorry, but the model was poor, the recommendations were mostly ineffective, and the data shows it any way you want to evaluate it. A weaker and slightly more transmissible virus profile doesn't explain away the huge difference in model performance.
BTW. The R0 variable already took into account the Alpha variant transmission range, and why I didn't include Delta or Omicron period comparisons which have significantly higher R0.
I didn't say not to look at deaths. I said you can't look at them as a simple function of how long interventions are in place. Where you are on the curve and what happened before the intervention affects what happens after.
I did exactly that. Deaths were low and trending flat when stricter interventions started. Cases were trending up relatively
modest prior to stricter intervention measures, including a 10,000 pound fine for violating self isolation if infected. Below are the visuals. Sept 22, 2020 is noted as that's the start date of the stricter measures (others were already in place).
Quote:
Whether Alpha transmission was within the expected range is irrelevant. The point is that it accounts for the relative difference in case numbers before and after the spike
.
It's relevant because the model incorporated the Alpha variant spread factor. Alpha R0 has been assessed between 2-3. The model gave forecasts with R0 values of 2.2, 2.4, and 2.6. So saying the mitigation/suppression forecasts were extraordinarily wrong because the COVID variant was more infectious than anticipated isn't correct because its transmissibility was accounted for. And its spread factor would not account for such an exponential increase in actual outcome, especially when it's occurring during increased mitigation/suppression measures. That would be a reflection of the effectiveness of the measures themselves in slowing/reducing spread.
With COVID, the virus is going to virus, regardless of what we do. I'm just glad it wasn't as deadly as others, like the flu when it was novel, because this one has been highly infectious. It's as infectious as its cousin the common cold but not as deadly as its other cousins SARS-1 and MERS. Let's hope that doesn't change.
