The COVID undercount, May 16, United States: 32.8%.

Every morning, if you like, you can go to the Johns Hopkins dashboard and get the current count of COVID-19 cases and deaths in the United States. Neither is particularly reliable; we don't test everyone, so we don't know the real number of cases. And COVID-19 deaths are not reliable either. Here's why:

  1. At the beginning of the epidemic, in good faith we were failing to code deaths as COVID-19, because the disease was still mysterious at the time
  2. Very old people often don't show the canonical symptoms of COVID-19, and simply die unexpectedly. Unless you make an effort to test, you don't know COVID-19 was the cause
  3. And finally, some states' governors, all or mostly Republicans, are for political reasons miscoding COVID deaths as pneumonia
So these day-by-day numbers, while useful and fairly immediate, don't catch all the deaths.

Fortunately there's a more reliable but slower alternative. All deaths in the US are recorded by the CDC as death registrations. These have cause of deaths as reported by the attending physician. And for any given week of the year, total US deaths are remarkably constant over periods of about half-a-decade. Tell me how many people died this corresponding week last year, or better yet over the last 5 years, and I can usually tell you to high accuracy (within 1%) deaths this year. Death rates decay slowly throughout the spring, but in a predictable way.

So in a severe epidemic, we can measure the death rate simply by subtracting actual deaths from expected deaths. It doesn't depend on diagnosis. It simply tells you how many more people died than expected.

These are deaths for six age groups, averaged for each week over the last five years, compared with deaths this year.

As you can see, the expected deaths (orange) and deaths in 2020 (blue) follow each other closely for the first 10 weeks in the year.

(Parenthetically, I think the very slight mismatch in thr 65-74 age group is due to Vietnam, which killed 60,000 Americans, mostly male. In 2015-2016, the 65-74 group corresponded to those born in 1941-1951. The tail end of that group were draft age during the height of the war. Enough young men from that group were killed that there is a significantly higher percentage of women in that clade, with a correspondingly lower death rate. But, whatever...)

. But after week 10, there's a huge increase in 2020, primarily in the older age-groups. That is COVID. Under 25s don't show it; it's relatively huge for over-85s. I was interested that the risk for 25-45s was a bit higher than I thought it would be. Older than 25, you aren't safe.

Anyway, for the moment, none of that matters. We just add them all up, giving this:

...and we just integrate the difference between the two curves between Week 10 and week 19 (March 8 - May 16, inclusive), getting 124770 deaths. Problem is, when we look at the Johns Hopkins numbers, we get 82543. So Johns Hopkins is only recording 66.2% (almost exactly ⅔) of all excess deaths. Right now, where we have 117,717 deaths according to Johns Hopkins, the true number of excess deaths is better estimated at 178,000 deaths.

Now, you could argue that there might be other contributions to excess deaths -- untreated cancer, heart disease, etc.. But there's no evidence for any of that; it's mere speculation. In fact, the lockdown may well have reduced accidental and automobile deaths. In fact, excess deaths are the best estimate we have of COVID deaths, and the data quality looks excellent.

Tomorrow we'll dig down into the excess deaths, and try to see what else the data tell us.

Comments

Post a Comment

Popular posts from this blog

Benny Johnson and Turning Point USA at UNL: stealing people's stuff to own the libs

Image
The newly 'recognized' Turning Point USA UNL chapter have immediately put paid to speculation they might be any different than the other chapters of the alt-right campus freak show, by recruiting the unsavory Benny Johnson as their first speaker. Benjamin Arthur Johnson graduated from the University of Iowa in 2009. Although it's been variously reported he also has a degree in Organic Chemistry, U. Iowa seems to think he only graduated in Psychology . Since he was an honors student, I hoped to find the thesis he is supposed to have written; running that through a plagiarism checker might be a laugh riot. Unfortunately, it seems not to be in their digital repository. Johnson commenced his career with a listicle-laced gala performance of promiscuous plagiarism at Buzzfeed, which led to his firing. He stole from no fewer than 41 articles. Then, after a brief break to let the scandal die down, he got a job at IJR Review . The man has no shame; a hate click, he claims,...
Read more

Dan Whitney, aka Larry the Cable Guy, a short biography

Image
I have known of Dan Whitney, Larry the Cable Guy , longer than most people have. In the car on the way into work at the University of Nebraska at Lincoln, we used to listen to one of the usual mostly-talk rock radio shows. Larry would quite regularly call in. His accent fascinated me. Probably because I’ve lived in so many places, I like to try to place accents. Larry the Cable Git's accent was vaguely southern, or was meant to be southern, but I couldn’t place it anywhere in particular in the South. In any case, he’d distorted it so much you could make out only half of what he was saying.  What he was saying was mumbled, consciously transgressive hate.  Later on, I discovered he was was born and grew up in Pawnee City Nebraska, until his preacher daddy hauled him down to Florida. It’s a long story… (1) Daddy issues Daniel Lawrence Whitney, better known by his stage name and character Larry the Cable Guy, was born on February 17, 1963 in Pawnee City, Nebraska, to the Reverend ...
Read more

The Fortenberry slaves: a prelude

Image
Toiling in the mid-morning sun, in a hot, muddy cotton field in Copiah County, in the south of Mississippi, three young African American men, brothers, aged 18, 20 and 22, hoed the rows. It was late June, 1860, and so the men were slaves. The ground was wet from yesterday’s rain, the air heavy, and the men were anxious to finish their work before the clouds gathering on the western horizon flooded the fields with another summer cloudburst. Inside the plantation house, unbeknownst to them, they were being enumerated . James R Fortenberry, the slave-holder, a man of about 35, was listing their important personal details to Elijah Peyton, Assistant Marshall. These details were three; age, gender, and race. The race was officially ‘mulatto’; at that time, and in that place, mulattos were usually the sons or daughters of a black woman raped by a white slave-holder. They had no names, as far as the United States Government was concerned. Instead they would be listed, in Peyton’s elegant,...
Read more