My previous post about Essex covid data generated a bit of discussion. In particular, the key question seems to be: why do the hospital admissions seem to go up 7-14 days after the case numbers go up? It struck me yesterday that there’s a simple and straightforward answer: people who have symptoms get tested. In fact, I understand that tests tend to track people who have symptoms, I’m not sure what the exact ratio is.
So if people who have symptoms are getting tested, then when covid is spreading there will be more tests, which means more cases. That would be true even if every single positive case is a false positive (which I am in no way suggesting). But even if the testing regime is flawed, it doesn’t really matter because it’s actually tracking symptomatic people.
I still find it strange that cases across different areas can seem to track each other so exactly. For example, the latest data shows cases across Colchester and Tendring:
We’ll probably never know the answer.
The key question to my mind is, does testing add anything which we couldn’t have if we just asked people who had symptoms to report them (maybe via the ZOE App), and then asked people who had symptoms to self-isolate? Testing could then simply focus in on those who are symptomatic, maybe if less tests were being run then it might help with some of the problems I mentioned.
In all honestly there’s not much else to say without more data. The coronavirus dashboard doesn’t seem to provide local area data for the number of tests being run. It’s difficult to get a number of excess deaths for this time of year because that information isn’t broken down by region (as far as I can tell).
This might be a fruitful area for someone to look into more – but not me. I just don’t have the time at the moment!
Update to the update: the Probability and Risk blog have been looking into similar issues over the last couple of months. Check out this blog and the links at the bottom for more information.