(10-20-2020 07:15 PM)salukiblue Wrote: But it is pretty damn close to being predictable.
The simple point is that the fluctuation in death totals from year to year is normally +/- 0.2% and in more "extreme" years it's been about 1%.
This year it's looking to be a 10% increase.
Hmmm. What is different about this year?
Occam's Razor.
There are lots of ways to look at numbers, so it's easy to draw different conclusions from the same data. I don't think we're really saying different things.
Here is how I look at it, and consistent with the 'expected deaths calculation' (if you look into the CDC methodology).
Annual variances of 50k+ are not that unusual... and yet there are also 5 yr periods where the 'expected deaths' would be declining and then followed by a 50k+ rise... so 'expected over actual' of approaching 100k, without a pandemic.
Now let's look at flu deaths.
Our lowest estimate (these are estimates because unlike COVID, flu can be suspected but not confirmed thus not hit the death cert) in the past 10 years was 12k deaths during the 2011-2012 season (winter 2011). The HIGHEST estimate during that period was 61,000 in winter 2017... so a variance of 50k, just in flu... which is a small number of the total deaths in the US... far behind Cancer and Heart disease.
But before we go with that, we should also look at the 95% confidence rates since these are estimates. the 95% confidence rate for 2017 is a range from 46k-95k... so the number could have been more like 85k, not 50k, again, just from Flu.
Consistent with that, Flu in winter 2019 is estimated to be VERY light at 22k... the second lowest in the past 10 years and 40k off from the high just 2 seasons earlier. The 2013-2018 5yr average would ave been very high, so 2019-2020 would have been a SIGNIFICANT under-performer vis a vis the 'expected' number... and that is how the cdc does it... by the underlying components and not simply the aggregate.
And now consider that flu season is basically contracting the disease in late October until around February, meaning most flu deaths for 'season 2019-20' likely come in 2020... weeks to months after contracting the disease. I don't trust this estimate because this was precisely the time frame when we were unaware of COVID/just learning about it...
My ultimate point being that if we're 200k over 'expected deaths' up to 100k of that could be annual variance in flu and other events unrelated to COVID. I also think a meaningful number of those deaths could be from 'other causes', often related to job losses or other results of the shut down.... and not Covid... which I THINK is your point.
Without a Pandemic, without a shut down... the 'expected' deaths should have been +25k from last year... but the RANGE on that number could have been up to +75k from last year. Throw in the shut down stresses plus loss of access to care for many and that number is potentially double that... so perhaps 150k. There's likely been a reduction in traffic deaths, but a rise in 'riot' deaths or other criminal activity... and then throw the Pandemic in on top of that.
Too many people simply see numbers that somewhat align... actual - expected of +/- 200k and 200k reported Covid deaths and they assume that everything else was constant. That idea is demonstrably not true. EVERY component has changed... and if we saw that death from all sorts of URI diseases that could look like COVID but aren't like FLU are 'down' (like we're seeing for FLU) then it very clearly implies that SOME of those COVID deaths are actually likely misassigned flu.