In this post I would like to discuss a little more of the data around lockdown that shows why we are able to start easing this now. Post 3 discussed how we could be easing lockdown when the number of confirmed cases from testing in the UK looks like it is remaining constant:
Following this, someone made the very good point that I said:
“This data therefore probably does represent a flattened curve, it just doesn’t show it well, which is why restrictions are being eased.”
They pointed out that “probably” isn’t good enough when it comes to their health, which is absolutely right! I completely agree!
I used the word “probably” because as a scientist I try to use language accurately in order to avoid communicating misinformation (I will try to post more about how scientists talk later), so I wouldn’t say the data actually showed cases had dropped off without having seen this analysis – and as it’s not my actual job to do this analysis, I didn’t do it… The important thing here is that I trusted that following the catastrophic mistakes our government has made (that I hope there will be an inquiry into – I can’t imagine many virologists in SAGE pushing for herd immunity), scientists who are far more intelligent than me, with expertise in this area and whose job it is to do this analysis would have done it, and they would not have made that critical decision to recommend easing lockdown without good evidence.
But to test the validity of that assumption, I downloaded the testing numbers and confirmed cases data to calculate the % positive cases. The government actually publish the daily testing numbers and confirmed cases etc., but the site is a little clunky to navigate and/or process from. Luckily there are people who are logging this data daily and putting it into processable formats such as tables and excel files (the one I used was here). This gave me a list of the total cases and total tests by date, so I just had to subtract each days total to get the daily change in tests and cases and could then calculate the % of positive tests. Just to show I am analysing the same case numbers as in the first graph, this is the daily cases I calculated which I hope you would agree look the same:
So, what does the actual % of COVID tests coming back positive look like? Here it is:
This is very much back-of-a-beermat data analysis and there are many limitations/caveats that could be explored – and I’m sure there are far more intelligent people than me exploring them – but I hope that you can see the data does actually show an infection that spread and has now mostly receded. Another comment someone made was that as we didn’t have the widespread testing data at the start of the outbreak, we will never know what happened. Yes, it is true we don’t have all the data from that period, but the effect of the virus is still there in the data that we do have, we just have to know how to look for it… So, I hope that explains a little around how we could be easing lockdown with the current daily case numbers.
Interestingly to me, the peak of the positive tests starts ~4th April (graph C). Compare this to people reporting symptoms (graph D), where the peak was 1st April:
The progression of COVID symptoms is undoubtedly complex, but there appears to be ~4 day gap between patients who recover and patients who develop more severe symptoms and need hospitalisation. The point I am trying to make is that simply reporting symptoms is actually matching incredibly well with the biochemical testing data, and can be vastly more widespread (3.6 million people currently) – so to make a point that I will probably keep coming back to – why not contribute to this as it only takes 10 seconds per day?