[quote StatisticallyChallenged]@NotAnActualSheep I need to look in more detail too, but I think that from what you're saying there are two different tests in their criteria
- positivity rate assessed from sentinel sites
- actual positive cases per 100k
At the moment we're over the 50 cases although not in all areas so the positivity rate is a slightly moot point but there have been times in the past when the positivity rate was used to keep areas in higher tiers.
For the sentinel type testing the way the test population are tested is as important as the sample size. If you are testing only symptomatic and high risk people then your positivity rate will be higher than it is in the actual population, so 5% positive in this group would not mean 5% in the population[/quote]
Yes, that's how I understand it too.
The late lamented Scottish tiers seemed to be based on different numbers... So tier 2 was (from memory) 30-75 cases per 100k, and 3-5% positivity, tier 3 75-150 cases and 5-10% positivity and tier 4 over 150 cases per 100k and over 10% positivity (there were other criteria too, based on hospital capacity etc). But it was the "worst" criteria that mattered even if all the others indicated that an area should be in a lower tier. Which was why the % positive was so annoying, as it could so easily be lowered by changing the testing protocol (in March/ April when they were only testing people in hospital, it was 30-40%, which is obviously ridiculous, but it fell rapidly as soon as they introduced community symptomatic testing, but not low enough, seemingly). Edinburgh was in tier 3 forever even though cases were well below 100 in the autumn, but %positivity was always above 5%... until it fell to below 5%and it still wasn't moved!
My guess is that if they reintroduce tiers it will use the (harsher) WHO criteria rather than the previous ones.
And you're right - 5% of tests being positive isn't the same as 5% of the population having it. At the moment around 5% of tests are positive Scotland-wide, but only 0.55% of the population have it according to the ONS. But if we are to get to the magic 50 per 100k (even if we're only counting out of the tests done, so still likely to miss many infected people) we need to halve the current number of infections, even if we get the %positive figure really low! Obviously what we really need is for loads of people to test negative to get both figures down. But that's hardly a surprising observation 