I'll do two cross checks of different data sources
I trust the ONS data to some extent, but I'll take their data, and cross check it with the ZOE app data which I also trust a bit, but it is a totally different data source. If all is consistent then both sources are giving good data.
The ONS do random sampling for covid19, and that they do their own PCR tests that are (as I understand it ) considerably more reliable than the Lighthouse Labs, I've also also heard that the ONS tries to estimate and correct for false positive results.The ONS now say that a million Britons are currently infected,
https://www.bbc.co.uk/news/uk-55550906
Hm… so to cross check this, I'll use data from the ZOE symptom checker app .
https://covid.joinzoe.com/data#levels-over-time
The app is presently showing 760k symptomatics. Tim Spector estimates 20% more asymptomatic cases exist on top , making 912k, so it's nearly a million. It is consistent to an impressive degree, +/- 100k is good enough for government work.
So with two data sources giving much the same answer, I'm sorry to say I think there are a million British people who have covid-19 now.
Using an ifr of 0.23% (from the US CDC, which is much lower than UK govt or SAGE admit), it means, in two weeks, we will be seeing deaths of 2,300 per day. We have to wait 2 weeks before we know, it's 8 to 9 hundred a day at present, if govt. can be believed. But I'll cross check that anyway.
Deaths now pertain to prevalence on 21st Dec (time from symptom onset to death is ~ two weeks) According to ZOE cases on 21 Dec were 375k*1.2 = ,450k total (symptomatic+ asymptomatic.)
At ifr 0.23%, we should be seeing 1035 deaths per day, against actual 800, consistent enough for government work.
Conclusion: close enough for government work, ZOE and ONS give the same data.There are a million Brotons with it, deaths will increase to 2000+ a day unless prevalence falls swiftly, hence flap in Number 10 about vaccine.
Although it seems OK its worth noting that even ONS themselves expressed concerns over their data quality due to a massive drop in people responding.
hi fon,
Theres a flaw here.
I think the ZOE app shows those people who have symptoms on each day recorded. So if yesterday there are 706,000 who have symptoms it may be that *if* symptoms last for 10 days around 10% of those will still have symptoms having recorded as such 10 days earlier and might not have today on 6th ( having recovered or died) while maybe 10% of those have just started and will still have symptoms on 15th. Everyone else is moving through that day in accordance with the distribution of symptomatic cases.
You can have symptoms for many days but if you are going to die you can die on only one day . So if it is that symptoms last on average for 10 days at a IFR of 0.23% there will be 1/10 of the number of deaths or 200/day from that cohort infected 10 days earlier. Of course the time from symtoms to recovery or death is a more complex re;lationship and deaths take longer (20-22 days or much longer) Of the million with symptoms I'd expect to see in a rising symptomatic case load ( ZOE 300k-706k in 20 days average of last 10 days ) less than 500-700 deaths which is what is being recorded now. Of course it seems that around 50-70% are other respiratory infections reported as instructed by authrities as covid so real IFR is quite a lot less ( WHO admitted 0.14%)
I'm not worried there will be 2300 deaths per day.
if I ask somebody at the office, how are you? and they say "I've got a slight cold." If I ask them again later I do not think "That's two people with slight colds", one this morning and another now! I assume he has the same cold that he had before, and it would surprise me if the zoe app does not do that too!
real IFR is quite a lot less ( WHO admitted 0.14%)
That is old fake news.
https://www.thejournal.ie/factcheck-admit-covid-19-false-5276375-Nov2020/
I'm not worried there will be 2300 deaths per day.
We will know in just over a week and a bit.
Fon
I'm sorry if my explanation was rather long winded. It was not clear to me what the average or predicted symptomatic period was. You can be ill for 2 or 3 weeks but if you are going to die, you can only be dead on the day you die !
Rather than use the ZOE predicted number of cases in the UK which their site states as 806,696 today, look at COVID Symptom Study from the ZOE site. It reports 69,958 Total number of new daily cases across the UK Using CDC 0.23% IFR thats 161 deaths from todays number in what might be 2 weeks or a little longer.
It is not really important whether the CDC or the WHO IFR is used - they are both an estimate based on a standard all-agegroup profile - which SAGE have been keen to present the so-called new strain as targeting the younger groups. If it really was 0.23% several months ago, then by SAGE's logic the new strain being presented as more transmissible amongst the young but conversely no more deadly it must have a lower IFR.
I wouldn't believe what 'fact-checkers' try to tell you either.
As I said, I'm not worried but as you say lets see.
