COVID Problems: CDC's Method of Estimating 'Expected Deaths' and 'Excess Deaths' Is Misleading

One of the main COVID-19 statistics used by the CDC is referred to as the "percent of expected deaths" in 2020 when compared to previous years.  The CDC estimated a number for total deaths it expects in 2020.  The CDC then compares the actual reported number of deaths in 2020 to the number of deaths that the CDC estimated to be expected in 2020.  Deaths above the "expected deaths" number are then suggested to have been caused by COVID-19.

On its main COVID-19 data page, the CDC labels the comparison between the reported number of deaths in 2020 and the expected number of deaths in 2020 as the "percent of expected deaths."  To estimate the expected deaths in 2020, the CDC is using the average total deaths (also known as the mean of total deaths) from 2017, 2018, and 2019.

The CDC's number of expected deaths in 2020 requires close evaluation.  The CDC's data for 1999–2018 are found here (click "I agree").  Data from 2019 are more difficult to find but are obtainable here (click "Downloads," then "Download data"; the total deaths for 2019 was determined by adding column "K" from rows 171 to 222, which are data for weeks 1–52 of 2019).

 

 

 

 

Year

Total Deaths

CDC's Reported

Total U.S. Population

1999

2,391,399

279,040,168

2000

2,403,351

281,421,906

2001

2,416,425

284,968,955

2002

2,443,387

287,625,193

2003

2,448,288

290,107,933

2004

2,397,615

292,805,298

2005

2,448,017

295,516,599

2006

2,426,264

298,379,912

2007

2,423,712

301,231,207

2008

2,471,984

304,093,966

2009

2,437,163

306,771,529

2010

2,468,435

308,745,538

2011

2,515,458

311,591,917

2012

2,543,279

313,914,040

2013

2,596,993

316,128,839

2014

2,626,418

318,857,056

2015

2,712,630

321,418,820

2016

2,744,248

323,127,513

2017

2,813,503

325,719,178

2018

2,839,205

327,167,434

2019

2,845,796

Not Available

 

 

 

If the above data are true, then there are some questions that may be entering one's mind.  One question is only indirectly related to this article, and a complete commentary is not possible here.  But notice that from 1999 to 2010, the number of deaths per year stayed relatively constant from approximately 2,390,000 to 2,470,000 deaths per year while the total U.S. population increased by about 30 million people.  Since 2010, though, deaths per year have reportedly increased by about 400,000 per year compared to 2009, while up to 2018, the CDC estimated the population only increased by 20 million people.  Such a difference in reported deaths per year compared to estimated population change is a big deal; is the huge increase in yearly deaths in America a result an aging population or a huge increase in population due to legal and illegal immigration or "refugee re-settlement"?

Now for the main topic of this article: the CDC misrepresents excess deaths in 2020 by using a misleading and inappropriate method to estimate expected deaths.  Look above at the total deaths in 2017, 2018, and 2019 compared to the previous years.  The CDC reports that the total deaths increased in large amounts from 2009 to 2016 and then suddenly stopped increasing in 2018 and 2019.  While that is possible, it is worth noting as an odd trend in the data.

The most obvious falsification or error, though, is the CDC's method of estimating expected deaths in 2020 by averaging the total deaths from 2017, 2018, and 2019.  Notice how variable the total yearly deaths are from 2009 to 2019.  As one textbook on evaluating medical literature states, when reading medical research literature,

Readers should carefully assess the use of means [averages] in studies.  Means provide little information if the data are extremely variable. (p. 153)

That is, the CDC's method of averaging the total deaths for the years 2017, 2018, and 2019 to calculate expected deaths in 2020 is inappropriate because of the significant variation of the yearly total deaths; the use of averages by the CDC is a potential indicator of data manipulation — using an inappropriate statistical method in attempt to achieve a desired result or make the results ("excess deaths" and COVID-19 deaths in 2020) look more convincing.  The potential data manipulation could call into question some or all of the other COVID-19 data — especially the COVID-19 death count data — provided by the CDC.

One can see the absurdity of the CDC's method by calculating it oneself.  Using the CDC's numbers provided above, the average total deaths number for the years 2017, 2018, and 2019 is 2,832,835.  This means that the CDC expected 2,832,835 deaths in 2020 — and it means that the CDC expected the total deaths to actually decrease in 2020 compared to previous years.  The CDC's method results in a lower number of expected deaths in 2020 than what the CDC reports for both 2018 and 2019.

The total deaths in the U.S. have reportedly not decreased compared to a previous year since 2009 - yet the CDC estimates that total deaths would decrease in 2020?  In fact, according to the CDC's data above, the deaths significantly increased several years up until 2017, 2018, and 2019, when the CDC reports that total deaths suddenly remain almost the same.  One should see how inappropriate the CDC's method is to estimate expected deaths of 2020.

One can look at the data from a different perspective to arrive at the conclusion that the CDC is using an inappropriate method to estimate expected deaths in 2020.  From 2010 to 2016, the number of deaths increased by 275,813.  Using an arbitrary method as the CDC does, one could estimate deaths in the U.S. to increase by 275,000 every six years or so.  Such a method of estimation would result in nearly 3,000,000 expected deaths in 2020 — which would be 170,000 more than the CDC's current estimate of expected deaths in 2020.  If a more accurate estimate of expected deaths in 2020 is around 3,000,000, then there may not have been a large excess in deaths in 2020.

It is possible that the apparent falsification may be an error by the CDC rather than deliberate deception.  Even so, the method of calculating expected deaths inappropriately leads to lower expected deaths in 2020 and a higher number of deaths that can be (potentially wrongly) attributed to COVID-19.  Thus, the CDC's method of calculating expected deaths appears to be similar to the type of potential research misconduct known as "trimming," in that the expected deaths calculation method appears to manipulate the data to make the results look more convincing.  It appears to manipulate the expected deaths to a low number, thus misleadingly and unreliably making the excess deaths in 2020 appear to be higher; the CDC then apparently presumes that the wrongly calculated excess deaths in 2020 are caused by COVID-19.

There are more examples of the CDC and other public health officials skewing or otherwise misrepresenting COVID-19 data.  There is the CDC's March 2020 guide to death certificate–certifiers stating that "the rules for coding and selection of the underlying cause of death are expected to result in COVID-19 being the underlying cause more often than not."  There is the observation that the CDC is ignoring their own COVID-19 severity criteria, which suggests that COVID-19 is not nearly as severe as they are claiming, using fear appeals to artificially heighten risk perceptions.  There is the observation that the CDC is apparently ignoring basic medical facts on pneumonia and respiratory viruses in favor of potentially wrongfully increasing COVID-19 death counts.  Such questionable interpretations, misrepresentations, and methods could collectively lead one to question the reliability of some or all of the CDC's COVID-19 data — including the CDC's weekly COVID-19 anecdotes.

One of the main COVID-19 statistics used by the CDC is referred to as the "percent of expected deaths" in 2020 when compared to previous years.  The CDC estimated a number for total deaths it expects in 2020.  The CDC then compares the actual reported number of deaths in 2020 to the number of deaths that the CDC estimated to be expected in 2020.  Deaths above the "expected deaths" number are then suggested to have been caused by COVID-19.

On its main COVID-19 data page, the CDC labels the comparison between the reported number of deaths in 2020 and the expected number of deaths in 2020 as the "percent of expected deaths."  To estimate the expected deaths in 2020, the CDC is using the average total deaths (also known as the mean of total deaths) from 2017, 2018, and 2019.

The CDC's number of expected deaths in 2020 requires close evaluation.  The CDC's data for 1999–2018 are found here (click "I agree").  Data from 2019 are more difficult to find but are obtainable here (click "Downloads," then "Download data"; the total deaths for 2019 was determined by adding column "K" from rows 171 to 222, which are data for weeks 1–52 of 2019).

 

 

 

 

Year

Total Deaths

CDC's Reported

Total U.S. Population

1999

2,391,399

279,040,168

2000

2,403,351

281,421,906

2001

2,416,425

284,968,955

2002

2,443,387

287,625,193

2003

2,448,288

290,107,933

2004

2,397,615

292,805,298

2005

2,448,017

295,516,599

2006

2,426,264

298,379,912

2007

2,423,712

301,231,207

2008

2,471,984

304,093,966

2009

2,437,163

306,771,529

2010

2,468,435

308,745,538

2011

2,515,458

311,591,917

2012

2,543,279

313,914,040

2013

2,596,993

316,128,839

2014

2,626,418

318,857,056

2015

2,712,630

321,418,820

2016

2,744,248

323,127,513

2017

2,813,503

325,719,178

2018

2,839,205

327,167,434

2019

2,845,796

Not Available

 

 

 

If the above data are true, then there are some questions that may be entering one's mind.  One question is only indirectly related to this article, and a complete commentary is not possible here.  But notice that from 1999 to 2010, the number of deaths per year stayed relatively constant from approximately 2,390,000 to 2,470,000 deaths per year while the total U.S. population increased by about 30 million people.  Since 2010, though, deaths per year have reportedly increased by about 400,000 per year compared to 2009, while up to 2018, the CDC estimated the population only increased by 20 million people.  Such a difference in reported deaths per year compared to estimated population change is a big deal; is the huge increase in yearly deaths in America a result an aging population or a huge increase in population due to legal and illegal immigration or "refugee re-settlement"?

Now for the main topic of this article: the CDC misrepresents excess deaths in 2020 by using a misleading and inappropriate method to estimate expected deaths.  Look above at the total deaths in 2017, 2018, and 2019 compared to the previous years.  The CDC reports that the total deaths increased in large amounts from 2009 to 2016 and then suddenly stopped increasing in 2018 and 2019.  While that is possible, it is worth noting as an odd trend in the data.

The most obvious falsification or error, though, is the CDC's method of estimating expected deaths in 2020 by averaging the total deaths from 2017, 2018, and 2019.  Notice how variable the total yearly deaths are from 2009 to 2019.  As one textbook on evaluating medical literature states, when reading medical research literature,

Readers should carefully assess the use of means [averages] in studies.  Means provide little information if the data are extremely variable. (p. 153)

That is, the CDC's method of averaging the total deaths for the years 2017, 2018, and 2019 to calculate expected deaths in 2020 is inappropriate because of the significant variation of the yearly total deaths; the use of averages by the CDC is a potential indicator of data manipulation — using an inappropriate statistical method in attempt to achieve a desired result or make the results ("excess deaths" and COVID-19 deaths in 2020) look more convincing.  The potential data manipulation could call into question some or all of the other COVID-19 data — especially the COVID-19 death count data — provided by the CDC.

One can see the absurdity of the CDC's method by calculating it oneself.  Using the CDC's numbers provided above, the average total deaths number for the years 2017, 2018, and 2019 is 2,832,835.  This means that the CDC expected 2,832,835 deaths in 2020 — and it means that the CDC expected the total deaths to actually decrease in 2020 compared to previous years.  The CDC's method results in a lower number of expected deaths in 2020 than what the CDC reports for both 2018 and 2019.

The total deaths in the U.S. have reportedly not decreased compared to a previous year since 2009 - yet the CDC estimates that total deaths would decrease in 2020?  In fact, according to the CDC's data above, the deaths significantly increased several years up until 2017, 2018, and 2019, when the CDC reports that total deaths suddenly remain almost the same.  One should see how inappropriate the CDC's method is to estimate expected deaths of 2020.

One can look at the data from a different perspective to arrive at the conclusion that the CDC is using an inappropriate method to estimate expected deaths in 2020.  From 2010 to 2016, the number of deaths increased by 275,813.  Using an arbitrary method as the CDC does, one could estimate deaths in the U.S. to increase by 275,000 every six years or so.  Such a method of estimation would result in nearly 3,000,000 expected deaths in 2020 — which would be 170,000 more than the CDC's current estimate of expected deaths in 2020.  If a more accurate estimate of expected deaths in 2020 is around 3,000,000, then there may not have been a large excess in deaths in 2020.

It is possible that the apparent falsification may be an error by the CDC rather than deliberate deception.  Even so, the method of calculating expected deaths inappropriately leads to lower expected deaths in 2020 and a higher number of deaths that can be (potentially wrongly) attributed to COVID-19.  Thus, the CDC's method of calculating expected deaths appears to be similar to the type of potential research misconduct known as "trimming," in that the expected deaths calculation method appears to manipulate the data to make the results look more convincing.  It appears to manipulate the expected deaths to a low number, thus misleadingly and unreliably making the excess deaths in 2020 appear to be higher; the CDC then apparently presumes that the wrongly calculated excess deaths in 2020 are caused by COVID-19.

There are more examples of the CDC and other public health officials skewing or otherwise misrepresenting COVID-19 data.  There is the CDC's March 2020 guide to death certificate–certifiers stating that "the rules for coding and selection of the underlying cause of death are expected to result in COVID-19 being the underlying cause more often than not."  There is the observation that the CDC is ignoring their own COVID-19 severity criteria, which suggests that COVID-19 is not nearly as severe as they are claiming, using fear appeals to artificially heighten risk perceptions.  There is the observation that the CDC is apparently ignoring basic medical facts on pneumonia and respiratory viruses in favor of potentially wrongfully increasing COVID-19 death counts.  Such questionable interpretations, misrepresentations, and methods could collectively lead one to question the reliability of some or all of the CDC's COVID-19 data — including the CDC's weekly COVID-19 anecdotes.