Ruckedtobits wrote:domhnallj wrote:Ruckedtobits wrote:Despite having a life-long passion and involvement with numbers, I can't understand how Prof Nolan can confirm that the 'R- ratio' for infection in Ireland is below 1 - given the minimal level and biased nature of tests conducted to date in Ireland.
If the vast majority of testing has been conducted on people exhibiting symptoms of the Virus, but the medical analysis says that asymptomatic people can transmit the virus also, what is the science or mathematics that indicates we have sufficient information to explain or calculate the current rate of infection of the c. 4.6m people untested to date? [i.e. c. 100,000 of population of 4.7m tested to date]
I would be grateful for any explanation or reference to any Paper explaining the basis of Prof Nolan's assertion. My name isn't Thomas, but my instinct always tends to doubting when I can't understand the fundamental basis of any public utterance.
There is an absolute deluge of models hitting the page on this so it can be hard to sort the wheat from the chaff. I've attached a link to a paper from Neil Ferguson's group at Imperial since he is a well noted epidemiologist. I had a quick skim and they use a Bayesian approach with weak priors to generate their results. The relatively wide credible intervals they are reporting suggest they need a lot more data from individual countries and to probably use a more informative prior (but they need data to do this with any confidence). The model's output will change with updating though and this was published over two weeks ago.
https://www.imperial.ac.uk/media/imperi ... 3-2020.pdf[/quote
Thanks @domhnallj.
Interesting Paper and inevitably numerous caveats and suppositions. Primary source data is the 11 EU countries further along the Covid curve. The two most significant, Italy & Spain excluded the majority of Care Home deaths from their initial data. Both are now integrating this data into their Reports and Italy is investigating the death rate in all State owned Car Homes.
My fundamental concern is that the basis of estimation of 'R' is not based on verified data. Lacking any other reference point (which greater daily testing could provide), there is no logical basis for saying that Prof Nolan's figure of 'R'= or < 1 is anything more than wishful thinking.
I hope he's right but I'd prefer more than a guess as my guideline.
Think it's more likely to be a fairly educated guess rather than wishful thinking.
Nolan has a lot more data available to him than is publicly available.
His curves at the very least will be adjusted to reflect when tests were taken rather than when the results became available. The "when" could be further adjusted to take account of when symptoms were first noticed/reported.
No doubt there may be other tinkering that can done to refine the data.
The R is very likely less than 1 but the actual value would fall within a range with a fair degree of room for error.
Also the pressure on the ICU etc is a supporting indicator.
The caveat is what is happening in the "nursing homes".
For example is it appropriate to start an easing of restrictions using figures that exclude the data associated with nursing homes.
The more specific caveat here, of course, is that resources that are being diverted to the nursing homes at present.
It is likely that, until those resources become available again, any easing will be symbolic rather than significant.
However it is important that some easing takes place.
