Quote:
Originally Posted by Elbow Jobertski
The 500k number was explicitly a worst case scenario where no measures were taken.
The 20K number assumes that the present strict measures continue.
That's the thing. There are people that will take this revision to mean that the estimates that caused people to take these measures were excessive scare-mongering. That the revised estimates show that the experts don't know what they are talking about so we don't need things to be shut down.
Those people are a menace.
Yes and no.
People who view this as everything is hunkydorie are definitely looking at it the wrong way, for sure.
I'm referring to the predictions on page 13 of the study:
https://www.imperial.ac.uk/media/imp...16-03-2020.pdf
The study always broke things down into three scenarios... do nothing, perform mitigation, perform suppression. The study deduced that only the last of these, which included the most stringent "lockdown" policies, enacted immediately and left in place until a threshold of new cases fell below a certain value (potentially many months later), could avoid a medical catastrophy by severely surpassing hospital resources. "Do nothing" led to 500K+ deaths, but even the middling "mitigation" strategy only cut that in half (middle paragraph on page 16).
If you look at the predictions for a suppression strategy, the death figures ranged from 8700 to 39000 based on estimated R0 at the time and how soon the stringent measures were enacted as a function of ICU confirmed cases. I don't have access to UK ICU confirmation figures to match to their "trigger" values (these would be useful to be able to know which predicted values to use on page 13), but I have to assume they were already at or above the 400 case trigger since at the time the UK already had over 1500 confirmed cases and the very next day that increased by over 400. So, I think at that time, it would be accurate to say their prediction was closer to the "high" end of the suppression estimates (39000) than 8700... and again, these figures assumed immediate implementation of the more stringent policies.
The report was published on 3/16, at a time when the UK strategy might not even have been considered as a mitigation response. Over the course of the next week the UK moved closer to a mitigation strategy (in other words woefully inadequate according to the study). And it wasn't till
3 days ago that the UK enacted the more stringent restrictions, And in enacting them, they put a proposed time limit of 3 weeks on them, unlike the suggested indeterminate time limit in the study.
So, to summarize a lengthy post:
1) The "best case" scenario at the time of the original study was between 8700 and 39000 deaths, assuming an aggressive suppression strategy were immediately put in place, and remained as such for an indeterminate amount of time till ICU confirmations fell below 25% of the value at that time.
2) Based on what the UK strategy was at the time, and had it remained unchanged, it seems like the study predicted between 85000 and 98000 deaths.
3) The even stronger mitigation strategy which the study thought would be very inadequate was put in place and remained as such for a whole week after 3/16.
4) The new more stringent "suppression" policy was not put in place till three days ago and is only scheduled to last 3 weeks. I'm sure the government will re-evaluate at the end of that period, but this is certainly less than what the study proposed in their best case scenario.
5) The current estimate for how "contagious" covid-19 is is a bit higher than it was on 3/16. So this would be a factor leading to
worse predicted numbers.
6) Perhaps my use of the term "revised" in my original post is too strong since the study certainly left itself a fair bit of wiggle room. But given the 5 points above, I think it would be fair to say that the current prediction of 20K deaths (or significantly less), is rosier than the one put forth in the original study. There's nothing wrong with that, and it doesn't completely invalidate the study at all since it has always been an exercise in presenting various "what-if" scenarios.
Last edited by akashenk; 03-26-2020 at 10:07 PM.