On That Censored Johns Hopkins All-Cause Death Analysis
Taken down for a big picture view and the pedestrian observation that all-cause deaths are up only slightly right now (and that's with lockdowns)
You’d think by now our rulers would stop the anemic way they censor what they consider dangerous information. Their blundering attempts only highlight what they want to hide.
It’s particularly silly with coronadoom censorship. Take the Johns Hopkins blog post “A closer look at U.S. deaths due to COVID-19” detailing some work by a lady named Briand. It was up, it was circulated, then it was deemed dangerous and pulled down. The Wayback Machine still has it, of course.
If it weren’t for the censorship, I wouldn’t have got the dozens and dozens of requests to look at it. Now everybody is sure Johns Hopkins is hiding something. Hilarious.
The reason it was censored it particularly stupid, too: “… it was brought to our attention that our coverage of Genevieve Briand’s presentation ‘COVID-19 Deaths: A Look at U.S. Data’ has been used to support dangerous inaccuracies that minimize the impact of the pandemic.”
Yeah, sure. Ninety percent of the population is racing in every direction like extras in a Toho Godzilla movie, only in masks. Johns Hopkins thinks this level of abject irrational terror is just about right. Besides, everybody knows science means only have one unchangeable opinion on every matter.
Funniest thing: they forgot, at least of this writing, to censor the YouTube video where Briand gives a talk.
Anyway, to Briand’s work itself. I appreciate the spirit, but don’t think there’s as much to it as some are hoping.
For starters, stacked bar charts are a mortal sin. She uses one to show the number of attributed coronadoom deaths by age (CDC data).
Here’s another way to do it, the percent of all weekly deaths by age “involving” (CDC’s word, to paper over all difficulties) COVID.
The legend goes top to bottom, but the data runs bottom to top. There isn’t any kind of coronadoom signal until you get to 25-34 year olds, and even for them it is weak and diminishing. As we knew for months, those at the most risk are the oldest—and sickest.
Anyway, you can see that this chart, which can be made prettier, is still more informative than the stacked bar chart.
Two important words of caution. CDC and everybody else counts deaths “involving” coronadoom as fast as possible, but ordinary deaths can take weeks before they’re fully tallied. That means the percentages shown in later weeks are exaggerations: the denominators are too small. The shape will in the ballpark, though. Finally, even doom deaths come in slower than desired, so the drop off at the end of all these pictures is purely an artifact of late counting.
We’ve seen this next chart, which I call the Perspective Plot. All cause deaths, those “involving” COVID, and flu + pneumonia. In early summer CDC stopped separate counts of flu + pneumonia because, we suppose, of the difficulty telling these deaths from doom deaths.
Another way to look at all cause deaths is this, where the black line is an estimate of where deaths would be absent the doom.
It is a mistake, as I always say, and Briand agrees, to imagine that bump is all coronadoom deaths. Not so. It includes others, as we’ll see.
Here’s an estimate of the total cumulative “excess” deaths, updated as of last Monday. The very last point is the estimate for the year so far.
Ignore the late-counting drop off. We’re maybe at 220,000 “excess” deaths this year, a number that may well go down. Again, do not think all these are all doom deaths. They are not.
Here’s two ways to look at that claim (CDC data).
First is number of weekly deaths of the top killers. Note the log scale!
Top two are heart disease and cancer. COPD and stroke battle it out for third and fourth place. Without the doom, that is. The peaking yellow line are mystery diseases (“What should I put down on the form?” “How do I know: don’t bother me”).
Problem with this plot is it exaggerates the doom. These are only the top killers, not all of them. Put all of them together and you get the Perspective Plot.
Now Briand says certain deaths are lower this year in order to boost coronadoom totals. That’s sort of true in some cases, but with some twists. Here’s another way to look at it, breaking each of the top diseases out and plotting one line for 2019 and one for 2020.
The “Natural Cause” is what CDC calls all-cause in this data, so it’s a repeat. Note the changing y-axes!
Late counting explains the red line drop offs. But we also expect, all things equal, the red line to be about 4% or so higher because of population increase.
See septicemia? It peaked in April. Any late counting (not likely for deaths in April) would not remove the bump. I put it to a physician that the explanation for this bump was iatrogenic, i.e. aggressive intubations in the early phase of the panic when the cure-all was thought to be ventilators, ventilators for everybody! He changed the subject.
Well, 2 to 4 thousand people killed by sepsis is still small next to the total 220 thousand—but they are part of it.
All this is very tricky. Can cancer deaths really be down this year? Well, yes: if the deaths are instead tallied, sometimes, as coronadoom deaths. But then look at diabetes, a nice sharp peak same time as the doom peaked. Same with Alzheimers. Both kill a lot of old people, traditionally. The combination of the doom and the chaotic hospital situation (missed treatments etc.) for many of these sick people was one IV tube too many.
Like in this way:
More than 8 months into the pandemic, the very isolation meant to protect nursing home residents is also contributing to their deaths.
Confined to their rooms and largely cut off from visitors, many residents are experiencing mental and physical decline. https://t.co/Z0Eh7MKQUH
— NBC News (@NBCNews) November 28, 2020
Some deaths are down, like flu and pneumonia now. And COPD starting in summer.
Sorting all this out, as I’ve said over and again, will take years. Who died of what is not so easy. Medical data is a mess. This is why I have been insisting for months and months to look at the total all cause deaths to gauge the true severity of the situation.
The needle of that gauge is firmly on “DO NOT PANIC.”
Source: William M. Briggs