Discussion of data from CDC and an article featured in a comment by robc.

It’s important to note that “COVID” deaths below are deaths “of” and “with” COVID-19.  I’m not 100% certain if that is simply evidence of infection, or with symptoms.  It might be logical to presume that since they use the terminology “COVID,” which means “Coronavirus Disease,” that they mean deaths of people who show signs of disease, not simply infection, but I wouldn’t bank on it.  In any case, they are including deaths both caused by, and simply adjacent to, SARS-CoV-2 infection as measured by whatever methods they used (probably the PCR test, as implied/asserted by the original author).  Which is fine, as long as we know what rules we’re playing by.  Just know that these data don’t do a good job of determining whether COVID was causative in the death.

First, let’s look at the data from the article, formatted for this post:

Here is the first data table.  I am copying the data directly from the article, and have not verified that it matches the CDC data.  The note that the source is CDC is the original author’s, not mine.

US Mortality 2020 (Source: CDC)
Total Deaths COVID-19 Deaths
Number % of Total Number % of Total
Total 3,358,814 100.00% 377,883 100.00%
Age Group, Years
<1 19,146 0.57% 43 0.01%
1-4 3,469 0.10% 24 0.01%
5-14 5,556 0.17% 67 0.02%
15-24 35,470 1.06% 587 0.16%
25-34 72,678 2.16% 2,527 0.67%
35-44 103,389 3.08% 6,617 1.75%
45-54 189,397 5.64% 17,905 4.74%
55-64 436,886 13.01% 44,631 11.81%
65-74 669,316 19.93% 80,617 21.33%
75-84 816,307 24.30% 104,212 27.58%
>=85 1,007,114 29.98% 120,648 31.93%
Unknown 86 0.00% 5 0.00%

 

The original author notes (and it’s one of the main points of his article) that COVID related deaths and total deaths, within age cohorts, are closely correlated once you get to the 45-54 year age bracket and above. This relationship does not hold for age brackets younger than that cohort.  The author argues that this is evidence that “The positive Corona-test is a random variable in relation to the observed result ‘death’.”

He then goes on to say:

On average, however, the “COVID-19 deaths” would have left this world at the same time, with Corona or from/with another virus or another disease. (In fact, a high number of them did probably die of something else than COVID-19.)  We are not immortal. On average, we die at our average age of death.

As a group, the COVID-19 deaths are part of normal and, in the last resort, unavoidable population mortality.

Why, then, has the US, why have some (but not all!) countries observed significant excess mortality in 2020?

I do not pretend I have a definite answer to this question; it still needs to be analyzed in much further depth, and we can only hope that this will actually happen.

 

I don’t have an answer to this question, either, but the implication is that perhaps it is something with the response to the COVID-19 pandemic that has caused these excess deaths.  I wanted to take a look at the data and what they say about these excess deaths, compare them across age cohorts, and see if we can in some way quantify the effects.  It should by now be evident to people paying attention that as age increases, the likelihood of dying of/with COVID-19 increases.  I’m not going to get into a discussion of comorbidities, because the data I have don’t include any information on them.

Since the table above includes total deaths and COVID deaths, it is trivial to calculate the Non-COVID Deaths in 2020.  Note that the percentages by age cohort are match closely with the previous numbers, which they should be.  Also, here I’ve calculated the deaths of/with COVID as a percentage of the total deaths within age cohort.  For instance, of all the people in the 45-54 year age bracket who died last year, nearly 10% of them died with/of COVID.

US Mortality 2020 (Source: CDC) My Calculations
Total Deaths COVID-19 Deaths Non-COVID Deaths 2020 COVID Deaths as % of All 2020 Deaths
Number % of Total Number % of Total Number % of Total
Total 3,358,814 100.00% 377,883 100.00% 2,980,931 100.00%
Age Group, Years
<1 19,146 0.57% 43 0.01% 19,103 0.64% 0.22%
1-4 3,469 0.10% 24 0.01% 3,445 0.12% 0.69%
5-14 5,556 0.17% 67 0.02% 5,489 0.18% 1.21%
15-24 35,470 1.06% 587 0.16% 34,883 1.17% 1.65%
25-34 72,678 2.16% 2,527 0.67% 70,151 2.35% 3.48%
35-44 103,389 3.08% 6,617 1.75% 96,772 3.25% 6.40%
45-54 189,397 5.64% 17,905 4.74% 171,492 5.75% 9.45%
55-64 436,886 13.01% 44,631 11.81% 392,255 13.16% 10.22%
65-74 669,316 19.93% 80,617 21.33% 588,699 19.75% 12.04%
75-84 816,307 24.30% 104,212 27.58% 712,095 23.89% 12.77%
>=85 1,007,114 29.98% 120,648 31.93% 886,466 29.74% 11.98%
Unknown 86 0.00% 5 0.00% 81 0.00% 5.81%

 

Some have made the argument that old people are going to die anyway, and, looking at the rates of death above, we see that the death rates of/with COVID are equivalent to the overall death rate for those cohorts, but others note that the COVID deaths are in addition to the total deaths we would normally expect.

Well, is that true?  In order to figure this out, we need to look at historical death rates and make some projections for 2020, since we’re getting into hypothetical territory here.  Let’s look at some more data.  Thankfully, the original author pulled some more historical data from CDC and presented it.

Here is the original author’s second data table.

US Death Rates per 100,000 Population
Total 2019 Total 2020 Change COVID19 2020
Age Group, Years
<1
1-4 23 22 -1 0
5-14 13 14 0 0
15-24 70 83 14 1
25-34 129 158 29 6
35-44 199 246 47 16
45-54 392 468 75 44
55-64 883 1,029 145 105
65-74 1,765 2,069 304 249
75-84 4,308 4,980 672 636
>=85 13,229 15,007 1,779 1,798

 

Here, the author compares the death rates per 100,000 population in the years 2019 and 2020.  He also shows them alongside the rates of death with/of COVID-19.  Following some links in the article, we come to a CDC data brief on Mortality in the United States, 2019.  Linked within that data brief are some raw data for death rates per 100,000 population in 2018 and 2019.  Using these historical data we can back-calculate the population in each age cohort, as well as a projected population for 2020, since I can’t find 2020 population from the Census Bureau split along the appropriate age cohorts.  We can see below, a net projected increase in population from 2019 to 2020 of about 1.37 million, which compares reasonably to the 2018-2019 increase of 1.3 million.  It’s almost a 5% increase in the increase, but it’s not out of the ballpark, so we’ll use it.

US Death Rates per 100,000 Population Calculation of 2020 Population (Not Available From Census Yet)
2018 2019
Age Group, Years Number of Deaths Death Rate per 100k Population Back-Calculated Population Number of Deaths Death Rate per 100k Population Back-Calculated Population Population Delta % Change in Population Projected 2020 Population
<1 N/A N/A N/A N/A N/A N/A N/A N/A N/A
1-4 3830 24 15,958,333 3,676 23 15,776,824 -181,509 -1.14% 15,597,379
5-14 5450 13 40,977,444 5,497 13 41,022,388 44,944 0.11% 41,067,382
15-24 30154 70 42,954,416 29,771 70 42,713,056 -241,360 -0.56% 42,473,052
25-34 58844 129 45,686,335 59,178 129 45,945,652 259,317 0.57% 46,206,441
35-44 80380 195 41,284,027 82,986 199 41,659,639 375,612 0.91% 42,038,668
45-54 164837 396 41,636,019 160,393 392 40,874,873 -761,147 -1.83% 40,127,640
55-64 374836 887 42,273,148 374,937 883 42,447,300 174,152 0.41% 42,622,170
65-74 543778 1,783 30,492,794 555,559 1,765 31,483,566 990,771 3.25% 32,506,529
75-84 675205 4,386 15,394,200 688,027 4,308 15,969,802 575,603 3.74% 16,566,927
>=85 880280 13,451 6,544,492 873,746 13,229 6,604,977 60,485 0.92% 6,666,021
Unknown N/A N/A N/A N/A N/A N/A N/A N/A N/A
2,817,594.00 323,201,208 2,833,770 324,498,077 1,296,868 325,872,210

 

Please note that the CDC did not include death data for the age groups “<1” and “Unknown” in 2018 and 2019, so they are “N/A” from here on.

So, now what we need to do is to take the number of deaths in 2018 and 2019 and use them to calculate an expected death count in 2020, absent the effects of COVID-19.  Yes, Yes, I know, we shouldn’t extrapolate a line from two points, but right now, all I have are the 2018 and 2019 death data, and I don’t have time at the moment to get prior years’ data.   Consider this table to be a rightward extension of the previous tables.

 

Calculation of Expected Deaths in 2020 and Comparison to Actual
Age Group, Years 2018 to 2019 Change in Death Count Death Count Rate of Change Expected Death Count Change for 2020 Expected Total Deaths in 2020, Excl. COVID-19 Expected 2020 Death Rate per 100k Population, Absent COVID-19 Actual 2020 Total Deaths Excess Deaths in 2020 (Actual 2020 Deaths minus Projected 2020 Deaths) 2020 COVID Deaths Excess Deaths Not Explained By COVID
<1 N/A N/A N/A N/A N/A 19,146 N/A 43 N/A
1-4 -154 -4.02% -148 3,528 23 3,469 -59 24 -83
5-14 47 0.86% 47 5,544 13 5,556 12 67 -55
15-24 -383 -1.27% -378 29,393 70 35,470 6,077 587 5,490
25-34 334 0.57% 336 59,514 129 72,678 13,164 2,527 10,637
35-44 2,606 3.24% 2,690 85,676 199 103,389 17,713 6,617 11,096
45-54 -4,444 -2.70% -4,324 156,069 392 189,397 33,328 17,905 15,423
55-64 101 0.03% 101 375,038 883 436,886 61,848 44,631 17,217
65-74 11,781 2.17% 12,036 567,595 1,766 669,316 101,721 80,617 21,104
75-84 12,822 1.90% 13,065 701,092 4,311 816,307 115,215 104,212 11,003
>=85 -6,534 -0.74% -6,486 867,260 13,225 1,007,114 139,854 120,648 19,206
Unknown N/A N/A N/A N/A N/A 86 N/A 5 N/A
16,176 16,941 2,850,711 3,358,814 488,871 377,883 111,036

 

Now we are getting to some of what I think are really interesting numbers.  Look at the “excess deaths” column.  Here we see that there were over 480,000 deaths in 2020 beyond those expected from a simple extrapolation from 2018 and 2019.   That is a lot of people–and they account for 17% more than the total deaths in 2019!  Also, look at the “excess deaths” that cannot be explained by COVID, the disease.  That number is an astonishing additional 111,000 lives lost due possibly to the effects of the response to the COVID pandemic.  To put it another way, they represent 22% of the excess deaths in 2020, and nearly 30% of the total number of COVID deaths.  Could it be that our response to the COVID pandemic has increased the total lives lost by thirty percent?  Proponents of the COVID response will, of course, counter that we do not know how many would have died of/with COVID if the response had not bee undertaken, but neither I nor they have, or ever will have, data to argue either side.  It’s a counterfactual.  All we can say is that for some reason, there has been an additional loss of life during this pandemic that is unassociated with the actual illness.

Now we come to the crux of what I want to present.  Look at this table below, where I calculate the “Person-Years” of life lost.  My method for this calculation was to assume an 80-year average lifespan for all age cohorts.  Then I calculated the average lifespan remaining for each cohort.  Then, I took the average remaining lifespan for each cohort and multiplied it by the deaths in each cohort.  This gives a total number of person-years of life lost.  I have a feeling that many of us would have expected the numbers to be slightly different–for instance, those who argue that “old people have less life ahead of them, and therefore will not contribute greatly to the total amount of life lost” would be dead wrong.  In fact, for all excess deaths, the 55-84 year cohort accounts for a nearly equal amount of lost person-years of life as the 1-54 year cohort.  Arguments can be made about quality of life, but that is a very subjective measure, and I believe that most individuals value their remaining life quite highly–probably more so than an outsider would.

 

Calculation of Person Years of Life Lost (Assumption of Average Human Life Span in the U.S. of 80 years)
Age Group, Years Average life Span Remaining Person-Years Lost of Life (Expected) Person-Years Lost of Life (Actual) Person-Years Lost of Life (COVID) Person-Years Lost of Life (Unexplained by COVID)
<1
1-4 77.5 273,435 268,848 1,860 -6,447.4
5-14 70.5 390,881 391,698 4,724 -3,906.1
15-24 60.5 1,778,268 2,145,935 35,514 332,153.2
25-34 50.5 3,005,452 3,670,239 127,614 537,173.8
35-44 40.5 3,469,898 4,187,255 267,989 449,368.2
45-54 30.5 4,760,099 5,776,609 546,103 470,407.3
55-64 20.5 7,688,280 8,956,163 914,936 352,947.9
65-74 10.5 5,959,750 7,027,818 846,479 221,589.5
75-84 0.5 350,546 408,154 52,106 5,501.3
>=85 -5 -4,336,302 -5,035,570 -603,240 -96,027.5
Unknown
23,340,305 27,797,147 2,194,082 2,262,760

 

Sadly for anyone who values life, in 2020 we lost nearly thirty million person-years of life in the United States alone.  Imagine the potential!  It’s astounding.  It’s even scarier that, had 2020 been a “normal” year for deaths, we would have expected well over twenty million person-years of life to be eliminated.  Think of what a human being is capable of in a year of life.  You can write a novel; design a manufacturing plant; fall in love and have your heart broken; produce a hit film; and a million other things.  If every one of those people had been living life to the fullest, imagine what they have lost.  Of course there are negatives to life as well, and it’s not all sunshine, puppies, and productivity, but the sheer potential for things to not only happen, but be made to happen, is nearly unfathomable.

But beyond the simply incomprehensible loss of human potential, consider some other points:  For instance, the excess person-years lost by COVID are actually fewer in number than those lost, not explained by COVID.  How can that be?  To understand that, we have to look closely at the numbers and see where the deaths are, and where the remaining lifespan is.  Go back up and compare the excess deaths, COVID deaths, and COVID-unexplained deaths by age cohort.

Now we look at the fraction of the life lost as it varies among the age cohorts.

 

Calculation of Person Years of Life Lost (Assumption of Average Human Life Span in the U.S. of 80 years)
Age Group, Years Average life Span Remaining Person-Years Lost of Life (Expected) Person-Years Lost of Life (Actual) Excess Person-Years Lost of Life (COVID) Excess Person-Years Lost of Life (Unexplained by COVID) COVID-Explained Fraction of Excess Person-Years Lost COVID-Unexplained Fraction of Excess Person-Years Lost Projected Potential Person-Years of Life based on 2020 Population
<1
1-4 77.5 273,435 268,848 1,860 -6,447.4 0.01% -0.02% 1,208,796,889
5-14 70.5 390,881 391,698 4,724 -3,906.1 0.02% -0.01% 2,895,250,417
15-24 60.5 1,778,268 2,145,935 35,514 332,153.2 0.13% 1.19% 2,569,619,655
25-34 50.5 3,005,452 3,670,239 127,614 537,173.8 0.46% 1.93% 2,333,425,262
35-44 40.5 3,469,898 4,187,255 267,989 449,368.2 0.96% 1.62% 1,702,566,046
45-54 30.5 4,760,099 5,776,609 546,103 470,407.3 1.96% 1.69% 1,223,893,034
55-64 20.5 7,688,280 8,956,163 914,936 352,947.9 3.29% 1.27% 873,754,477
65-74 10.5 5,959,750 7,027,818 846,479 221,589.5 3.05% 0.80% 341,318,557
75-84 0.5 350,546 408,154 52,106 5,501.3 0.19% 0.02% 8,283,464
>=85 -5 -4,336,302 -5,035,570 -603,240 -96,027.5 -2.17% -0.35% -33,330,105
Unknown
23,340,305 27,797,147 2,194,082 2,262,760 13,123,577,696

 

The excess deaths really start to be noticeable in the 15-24 cohort, where they jump to over 6,000.  But note that these are overwhelmingly in the “COVID-unexplained” realm.  These are also the people among the excess deaths who have the most life, on average, ahead of them.  So they represent an incredible number of person-years lost above the expected amount–332,000 or so.

If we look at the total excess COVID-unexplained deaths for ages 25-64, we see that there were over 54,000 excess deaths not explained by COVID, covering 1.8 million lost person-years.  The 45-64 year bracket accounts for somewhat less than half of those.

Interestingly, the fraction of excess person-years lost to COVID peaks among the older population (45-80) while the COVID-unexplained lost life is focused around the younger population (15-54).  What is particularly tragic is that the total life lost to COVID is actually secondary to that lost to excess causes unexplained by COVID, even though the number of COVID excess deaths exceeds the the number of COVID-unexplained deaths by a factor of three.

One of the many tragedies of this pandemic, which will hopefully be recognized sooner rather than later, is that the total person-years of life lost due to causes not explained by the particular disease in question appears to actually exceed the loss to the disease itself.  As this becomes more clear to the general populace, hopefully we will see more care taken in the future when dealing with major public health emergencies, better evaluation of the danger of the emergencies themselves, and more rational and scientific approaches to treating and containing pandemic diseases.