Percentage of positive cases in Scotland

[NotAnActualSheep]

Apologies if this has been covered before, but it's something I've just noticed today, and I can't see it being discussed elsewhere. There may also be a really simple thing that I've missed...which is going to make me look really stupid. But here goes.

I've been worried about the increase in positive cases...from around 1%of all tests in early September, passing the 5% WHO "virus is not under control" level a few weeks ago, and it's now at about 14% of all new tests being positive, apparently. NS has been mentioning this rise in her briefings, and it was included in the justification for the "no household mixing" restrictions a couple of weeks ago. Fair enough...it is concerning, and does suggest levels are increasing and not just an artefact of more testing.

But today, there were 1009 new positive cases out of 19664 new tests. This is being reported as a positive test rate of 14%. I thought that didn't sound right (should be nearer 5%) and indeed the numbers here say the methodology is being revised. The page it links to explains how the numbers are currently calculated.

Basically, in order to avoid double counting of people, every person is only recorded in the data once and only your first test is counted in the "number of new tests" figure (the positive % calculation is "number of new positive tests"/"number of new tests" X100) . So say you test negative on your first test, your negative test will be counted as a negative test, and your result will be included in the "number of new tests" figure. But if you then test positive a few weeks later, your positive test will be included in the "number of new positive tests" figure, but not in the "new tests" figure.

As I say, I may be missing something (please go easy on me...stats are not my forte) but this seems an utterly meaningless figure. It is BOUND to increase, as many people will be tested more than once, so won't be counted in the "new test" figure. It could potentially go above 100% if there are more positive cases than people being tested for the first time (whether positive or negative)! And it sounds as if the Scottish government has realised this and is looking at changing the methodology.

But up to now, have we been being scared by this increasing figure, being used to justify further restrictions, when actually it hasn't been tracking the %of positive cases at all? That sounds quite bad if true.

The statistics have been flawed all the way through this.

However, on an anecdotal basis, I know numerous people who are VERY ill at the moment with Covid and thus happy to reduce risk wherever possible.

I don't doubt that, and I hope they recover soon . But don't think it is justifiable to provide bad statistics to support a political/ public health decision. Cases are increasing...and hospitalisations and deaths. I'm not denying that. But also there is a load more testing than in March so you'd expect identified cases to be increasing (even Trump can work that one out). If we don't have an accurate %, it is hard to work out whether we are testing enough people, (in line with WHO recommendations) to identify an actual increase in the population, surely. As our testing programme is only symptomatic people, it's surely important to know what percentage of people with covid symptoms actually have covid...

I know the ONS are doing random population testing and ARE finding an increase across the UK. But Scotland seem to have been using this %positive figure to go beyond the rest of the UK restrictions, and it hasn't been showing what (I assume) everyone thought it was showing.

No you're quite right- it's appalling.

I think the lack of consistency with statistical reporting is possibly why we're in such a bad way- many people choose to disbelieve everything as there's been such misrepresentation.

I agree...I appreciate it must be a tricky set of data to present, with linked results and so on (and that's probably why England doesn't publish the same figure as far as I can find). But continuing to rely on and publish a figure that even a thicky like me can see is mathematically meaningless, but nevertheless shows a trend, seems a bit cynical.

I'm surprised that this hasn't been raised more widely, really. But as you say, it doesn't affect the point that cases are increasing...and if it is revised downwards, it may well increase again up to the current figure over time anyway! (Like figures for hospital cases in Scotland and deaths across the UK which were both decreased when they realised the calculations were a bit pants).