Tuesday, 28 July 2020

Coronavirus - daily new cases v daily new deaths

From worldometers.info:

There are lots of ways you can interpret that. I'm quite sure that some countries under-report and others over-report, but let's assume it cancels out and charts are a good guide to actual trends. There's a pessimistic and and optimistic way of interpreting anything, even something as grim as this.

1. Is the number of new cases really rising, as you would expect if R is greater than 1, or is the number of infections fairly stable, and the apparent increase is because they are testing more people?

2. Is this the start of the 'second wave' of deaths, or just the inevitable result of lockdowns being eased?

My slightly rosier view is to observe that the number of deaths in any week, while fairly stable, is falling as a fraction of new cases reported three weeks previously*. This might be because the virus is becoming less virulent, which is what such viruses tend to do (some faster than others); or it might be because hospitals are getting better at treating people; or it might be a bit of both. Whichever way, that's got to be A Good Thing.

* Assuming three weeks to be a typical lag between diagnosis and death. For example:

New cases March 28 - 51,000; deaths April 18 - 7,000; death rate = 13.7%

New cases June 20 - 180,000; deaths July 27 - 5,600; death rate = 3.1%


View from the Solent said...

Notice how, in UK, all the scare stories are about number of 'cases'? Little or nothing about the number of current all-cause deaths are below the 5-year rolling average. Or that the number of so-called (unspecified) tests is also increasing.

Mark Wadsworth said...

VFTS, "number of deaths normal" is hardly a headline, I suppose. The media get bored and go and try and terrify us about something else instead.

Derek said...

Doctors are definitely getting better at treating the virus, now that they've had a few months experience of what works and what doesn't. I hope that the virus is getting less dangerous but who knows.

As for the all-cause death count being below the 5-year rolling average, that's good news too. But it does makes me wonder how much of that is down to fewer traffic accidents and other consequences of the lockdown.

Mark Wadsworth said...

D, that cuts both ways.

The number of 'excess deaths' so far has been 60,000, official number of covid deaths (probably overstated) is 45,000.

So maybe there have been fewer car accidents or fights outside pubs. But as the NHS has largely shut itself down, there have been a net total 15,000 collateral damage deaths in addition to covid deaths.

Bayard said...

VFTS, if I was wearing my tinfoil hat, I'd say that the government want to keep us scared of the virus so that we clamour for the jab when it comes out. Otherwise they would have to make it compulsory to keep their sponsors in Big Pharma happy and that wouldn't go down well.

Mark Wadsworth said...

B, sod jabs. No long term point, would need to be redone every year, we just have to sit tight until the virus mutates to just being a bad flu, and we live with it (or indeed, die with it).

benj said...

@ Derek

Could also be that the average age of those catching it is lower?

Derek said...

Could be. In fact I think that's quite likely.