Quote:
Originally Posted by OLD BOY
No, I was commenting on the number of reported cases - that figure that appears on our TV sets each night. I am not arguing that the number of actual cases will fall.
The thing is, nobody knows the actual number of cases there are in the country because we cannot know what isn’t reported. One thing for certain is that there are far more positive cases than are reported.
The data indicating the proportion of tests that are positive proves very little. After all, there are specific reasons why people get PCR tests. What about the rest of positive cases where no PCR tests are performed?
|
Yes, the absolute number of cases is a crude estimate of the severity of any outbreak and certainly shouldn't be used to establish a health policy. However, it is a measure easily digested by the public which is critical in a public health emergency.
As always, the absolute number numbers reported should never be taken as gospel for the reasons you state - many people might get missed due to not testing. More important is the trending - are cases going up or down and therefore are things likely to get better or worse? Of course, trending is only possible if the drivers for testing stay the same,
i.e. people do a test for the same reasons every time.
The other option is to randomly sample a subset of the population and extrapolate from there. That is what the ONS reports do where around 180,000 people are randomly selected per fortnight for PCR testing. This probably gives a better guide of what is out there in the wild than people who self select (or not) through booking and taking tests.
Here is the current report. This is where you see the 1 in 15 in England testing positive. I know some on here don't like statisticians but these surverys are pretty up front about the limitations in the data (
article here and
here)
Unfortunately, the ONS surveys only come out weekly and probably aren't timely enough considering the doubling time of COVID when it runs riot