PreprintPDF Available

COVID-19 Deaths A Look at U.S. Data FEB 2021 WORKING PAPER Genevieve Briand

Authors:
Preprints and early-stage research may not have been peer reviewed yet.

Abstract

This paper is a follow-up on the Nov. 11th 2020 webinar, COVID-19 Deaths: A Look at U.S. Data. The effect of COVID-19 on U.S. deaths is assessed by using CDC data. Total deaths as well deaths per age group, cause and jurisdictions are analyzed.
Page 1 of 112
Genevieve Briand, Ph.D.
Assistant Director
MS in Applied Economics
Advanced Academic Programs
Johns Hopkins University
gbriand@jhu.edu
Working paper
February 2021
COVID-19 Deaths: A Look at U.S. Data
Summary
Readers are invited to write their own summary.
Introduction
Professor Hal R. Varian points out that ideas for research projects come as much from everyday life experiences
than from review of previous work in academic journals (19). Everyone has been living COVID-19, but
everyone’s real life experience of it (as opposed to virtual), is unique. COVID-19 has inflicted an economic and
societal toll that is still being measured. This paper concentrates on the effect of COVID-19 on U.S. deaths. For
a comprehensive assessment of that effect, overall deaths and deaths due to select causes are analyzed. Graphs
of weekly All Cause deaths across 2014-152019-20 seasons are also provided for each U.S. jurisdiction (50
states, District of Columbia, New York City and Puerto Rico territory). Readers are invited to consult them.
I would like to thank JHU Advanced Academic Programs for the opportunity to give the Nov. 11, 2020 webinar
this paper is based on (1). This paper has not been peer-reviewed, nor is it endorsed by the Krieger School of
Arts or Johns Hopkins University. It has been written and is being circulated to further discussion and
comments on the original webinar. The data have been updated some, as more became available for more recent
time periods. The analysis has also been expanded to individual U.S. jurisdictions.
Thank you to Ms. Gu for covering the Nov. 11, 2020 webinar in her Nov. 22, 2020 JHU News-Letter Article, A
Closer Look at U.S. deaths due to COVID-19 (2). Thank you to all who emailed me, for their interest, feedback
and testimonies, as well as questions, I will try to address in this paper.
Page 2 of 112
Table of Contents
Methodology .............................................................................................................................................p.3
Overview .......................................................................................................................................p.3
Line Charts ....................................................................................................................................p.4
Data ...........................................................................................................................................................p.6
CDC data sources ..........................................................................................................................p.6
Can the data be wrong? .................................................................................................................p.6
Can deaths be overcounted?..........................................................................................................p.7
MMWR weeks ..............................................................................................................................p.7
Death causes..................................................................................................................................p.8
U.S. deaths: Short-term historic context .................................................................................................p.10
How about excess deaths estimates?...........................................................................................p.16
Weekly deaths per age group: Constancy ...............................................................................................p.17
Weekly deaths per cause: Seasonality ....................................................................................................p.21
Weekly deaths per jurisdiction: Variability ............................................................................................p.45
Conclusion & Testimonials.....................................................................................................................p.53
References ...............................................................................................................................................p.56
Appendix: Graphs 19-71 .........................................................................................................................p.59
Select List of Tables and Graphs
Visually Accurate Graphical Representation of 1% Death Rate: Graph 1.
MMWR Weeks Comprised Within Each Calendar Year and Season: Tables 1-2
U.S. Percentage death rates, years 1999-2019 and seasons 2014-152019-20: Tables 3-8
U.S. Deaths per Age Group (% and average), week ending 01/01/2011/14/20: Graph 3 & Table 9
U.S. Deaths per Week and Cause, from MMRW week ending 01/11/14 to 11/14/20: Graphs 4-10
Week Ending Date (MM-DD-YY) Correspondence to Seasons’ Week Number: Table 10, Tables 12a-12b
U.S. Weekly Deaths per cause, comparison across 2014-15 to 2019-20 seasons: Graphs 11-15
U.S. Deaths per cause (in %) and year: Table 13
Where is Waldo? Graph 16, 70-71, Tables 14-15
U.S. All Cause Deaths per Week and Jurisdiction, across 2014-15 to 2019-20 seasons: Graphs 17-69
Page 3 of 112
1 Methodology
1.1 Overview
The methodology used for this analysis is very simpleput the data downloaded from publicly available CDC
datasets in graphical or tabular form, and try to make sense of them. This analysis can easily be replicated by
anyone who cares to do so. No statistical software per se, but a simple spreadsheet, was used. Microsoft Excel
was chosen, but any spreadsheet software could be used. To guide the interpretation of the graphs and tables,
the methodology adopted is to use simple statistics and logic. The objective of this analysis is not to produce an
“excess deaths number estimate”, but to assess whether the total deaths number the U.S. experienced in 2020
was unexpected or alarming.
In contrast, Roosen et al. produce excess deaths estimates, associated with COVID-19. Their results are
presented in the October 23, 2020 Morbidity and Mortality Weekly Report (5). As explained on the CDC
webpage that Rossen and her team maintain, the methodology they use is much more elaborate, “Estimates of
excess deaths presented in this webpage were calculated using Farrington surveillance algorithms” (7).
The Center for Disease Control and Prevention (CDC) data are the best available data on U.S. deaths. The CDC
indicates that the most recent deaths data made available are still provisional, and that is why the dates the
datasets were downloaded are provided in table captions and list of references. The data presented in this paper
are not estimatesthey are records of past deaths, maintained and made publicly available by the CDC. They
have not been produced, adjusted, nor tampered with, in any way, by the author.
Page 4 of 112
1.2 Line charts
Many visuals about COVID-19 deaths have been deceptive. Plots, or line charts, in this paper, can be too. The
reason for this, is that plots present visuals of changes that cannot directly be interpreted as relevant proportions.
Graph 1: Visually accurate graphical representation of 1 percent death rate.
To illustrate, consider Graph 1. If the blue rectangle is defined as representing 100% of the U.S. population,
then the pink line, at its top, represents 1% of it. Imagine the blue rectangle representing 100 blue lines, like the
pink line atop, stacks up together. In other words, the surface area of the pink line represents 1% of the blue
rectangle surface area. One could borrow a child’s graph paper and colored pencils to reproduce a similar graph.
Graph 2: U.S. All Cause Weekly Deaths per Season, 2014-15 to 2019-20
100% of population
1%
Page 5 of 112
In contrast, consider Graph 2. This is the plot of U.S. weekly All Cause deaths from season 2014-15 to season
2019-20 (why seasons are considered, and which weeks are included in each season and calendar year, are
questions addressed in the upcoming data section). Graph 2 itself will be considered in further details in the
upcoming “Weekly deaths per cause” section.
When you look at Graph 2, what is your interpretation of the extent of the change in All Cause deaths
experienced by the U.S. during the 2019-20 season, compared to prior seasons? Is your read from Graph 2 that
the total deaths number the U.S. experienced in 2019-20 was unexpected or alarming? Do you come up with a
percentage death rate?
Please, take a second to answer these questions.
Can you read from Graph 2 that the All Cause deaths number in 2019-20 was 9.2% higher than in 2018-19?
With a little bit of time, you could. Now, most importantly, can you read from Graph 2 that the death rate for
season 2019-20 was 0.9%? No, you cannot. Because a death rate is a deaths number, in proportion of its
population level, and you are missing a key piece of informationthe 2019-20 population level. Note, a death
rate of 0.9% is a usual death rate, one that has been experienced in recent yearsmore on that in the upcoming
“US deaths: Short-term historic context” section.
Compare your answers to the previous questions, to the answers given above, and draw your own conclusion
about whether Graph 2 was deceptive.
Despite the limitations of line charts, this tool will be used to analyze changes of death patterns across years and
causes, in sections “Weekly deaths per cause: Seasonality” and “Weekly deaths per jurisdiction: Variability”.
Page 6 of 112
2 Data
2.1 CDC data sources
The data used and their sources are as follows:
Data from page 2 of 2 of “Provisional Death Counts for Coronavirus Disease 2019 (COVID-19) By
Week of Death” table, https://www.cdc.gov/nchs/nvss/vsrr/covid_weekly/index.htm. Downloaded 12-
26-20. (9)
“Weekly Counts of Deaths by State and Select Causes, 2014-2018” dataset found via
https://data.cdc.gov/browse, https://data.cdc.gov/NCHS/Weekly-Counts-of-Deaths-by-State-and-Select-
Causes/3yf8-kanr. Last updated by CDC Nov. 10, 2020. Downloaded 12-26-20. (10)
“Weekly Counts of Deaths by State and Select Causes, 2019-2020” dataset found via
https://data.cdc.gov/browse, https://data.cdc.gov/NCHS/Weekly-Counts-of-Deaths-by-State-and-Select-
Causes/muzy-jte6. Last updated by CDC Dec. 23, 2020. Downloaded 12-26-20. (11)
2.2 Can the data be wrong?
The most recent CDC deaths data are still provisional. This is not an issue. It has always been the case. It is
normal and understandable: Collecting and making U.S. deaths data available to the public, in a timely fashion,
is quite an undertaking! The Center for Disease Control and Prevention (CDC) data are the best available data
on U.S. deaths. Because the most recent deaths data made available are still provisional, the dates the datasets
were downloaded are provided in table captions and list of references.
The reader, who cares to replicate this analysis, will download the data at a later date than the one they were
downloaded at, for this paper. This means, because the most recent deaths data are still provisional, such reader
might get slightly different numbers.
The care with which the CDC collects, keep records of and make publicly available deaths data, is not
questioned here. Reporting errors, if any, are minimized by procedures adopted by the CDC.
The data presented in this paper are not estimatesthey are records of past deaths, maintained and made
publicly available, by the CDC. They have not been produced, adjusted, nor tampered with, in any way, by the
author.
Page 7 of 112
2.3 Can deaths be overcounted?
All deaths are assigned one and only one underlying cause of death (4).
“The International Classification of Diseases (ICD) is designed to promote international comparability in the
collection, processing, classification, and presentation of mortality statistics”, (13). The tenth revision (ICD-10)
covers years from 1999 to the present (14). An online version of all ICD-10 codes is found on the World Health
Organization’s website (20); the latest version is the 2019 one (21).
Old age is not an underlying cause of death.
The new ICD code for COVID-19 deaths, ICD code U071, issued by the World Health Organization (WHO),
was introduced and implemented by the Division of Vital Statistics of the U.S. National Center for Health
Statistics (NCHS), on March 24, 2020 (15).
All Cause deaths are the actual number of dead bodies accounted for at one point in time and no dead body is
double counted. When those numbers are not yet final, All Cause numbers may become higher, at a later date.
2.4 MMWR weeks
Table 1: MMWR weeks comprised within each calendar year, 2014-2020.
Year
1st MMWR Week ending date
Last MMWR Week ending date
Total MMWR weeks
2014
01/04/2014
01/03/2015
53
2015
01/10/2015
01/02/2016
52
2016
01/09/2016
12/31/2016
52
2017
01/07/2017
12/30/2017
52
2018
01/06/2018
12/29/2018
52
2019
01/05/2019
12/28/2019
52
2020
01/04/2020
01/02/2021
53
Deaths are recorded per Morbidity and Mortality Weekly Report (MMWR) week. “MMWR weeks” refers to
the sequential numbering of weeks (Sunday through Saturday) during a calendar year (12). For MMWR year
2014, MMWR Week 1’s ending date is 01/04/2014 and MMWR year 2014’s last MMWR week is Week 53,
ending 01/03/2015. Calendar year 2020 also has 53 MMWR weeks, while years 2015-2019, each have 52: See
Table 1.
Page 8 of 112
Table 2: MMWR weeks comprised within each season defined for this paper, 2014-15 to 2019-20.
Season
1st Week ending date
Last Week ending data
Total MMWR weeks
2014-15
08/23/2014
08/15/2015
52
2015-16
08/22/2015
08/13/2016
52
2016-17
08/20/2016
08/12/2017
52
2017-18
08/19/2017
08/11/2018
52
2018-19
08/18/2018
08/10/2019
52
2019-20
08/17/2019
08/08/2020
52
Because deaths are cyclical (something further analyzed in the upcoming section), another standard way to
report deaths is over a season rather than a calendar year. For example, the CDC reporting period for influenza
seasons begins during MMWR week 40 and ends MMWR week 39 of the following year. In this paper, seasons
are centered around the 2014-19 peaks of deaths, to facilitate their comparison to the 2020 one, and address
concerns that the same weeks were not compared across years during the Nov. 11, 2020 webinar (1)see
upcoming Seasonality of Deaths section. In upcoming Tables 5, 8 and Graphs 11-15, 17-69 and 71, deaths will
be reported or plotted per season, instead of calendar year. Each season has 52 MMWR weeks: See Table 2.
2.5 Death causes
For each cause, deaths are recorded per MMWR week. All the causes/categories of death are not included in
the datasets used here. The list of causes of death included in the datasets used here is:
All Cause
Natural Cause
Septicemia (A40-A41)
Malignant neoplasms (C00-C97)
Diabetes mellitus (E10-E14)
Alzheimer disease (G30)
Influenza and pneumonia (J09-J18)
Chronic lower respiratory diseases (J40-J47)
Other diseases of respiratory system (J00-J06,J30-J39,J67,J70-J98)
Nephritis, nephrotic syndrome and nephrosis (N00-N07,N17-N19,N25-N27)
Symptoms, signs and abnormal clinical and laboratory findings, not elsewhere classified (R00-R99)
Diseases of heart (I00-I09,I11,I13,I20-I51)
Cerebrovascular diseases (I60-I69)
COVID-19 (U071, Multiple Cause of Death)
COVID-19 (U071, Underlying Cause of Death)
Influenza and pneumonia, chronic lower respiratory diseases, and other diseases of the respiratory system, are
three old categories of respiratory diseases. COVID-19 is a new one.
All Cause deaths are total deaths.
Page 9 of 112
All Cause are the sum of Natural Cause and Non-Natural Cause, or “external causes” deaths (21):
   
“Technically the natural causes are codes A-R, and external causes are V through Y. U-codes are ‘codes for
special purposes’ and COVID-19 is U071, which we group in with the natural causes (4)”.
Let Old Select Cause be the sum of deaths from the following old natural causes:
Septicemia (A40-A41)
Malignant neoplasms (C00-C97)
Diabetes mellitus (E10-E14)
Alzheimer disease (G30)
Influenza and pneumonia (J09-J18)
Chronic lower respiratory diseases (J40-J47)
Other diseases of respiratory system (J00-J06,J30-J39,J67,J70-J98)
Nephritis, nephrotic syndrome and nephrosis (N00-N07,N17-N19,N25-N27)
Symptoms, signs and abnormal clinical and laboratory findings, not elsewhere classified (R00-R99)
Diseases of heart (I00-I09,I11,I13,I20-I51)
Cerebrovascular diseases (I60-I69)
Hereafter, in the text of this paper, malignant neoplasms will be referred to as cancers.
COVID-19 (U071, Underlying Cause of Death) deaths are a subset of COVID-19 (U071, Multiple Cause of
Death) deaths: 5 to 13 % of deaths from weekly counts of COVID-19 (U071, Multiple Cause of Death) deaths
have something other than COVID-19 specified as the underlying cause of death (4). Consequently, for this
analysis, COVID-19 (U071, Underlying Cause of Death) deaths numbers are used.
Natural Cause deaths are the sum of deaths from Old Select Cause, Other Old Natural Cause deaths and
COVID-19 (U071, Underlying Cause of Death):
      
Page 10 of 112
3 US deaths: Short-term historic context
The standard way the Centers for Disease Control and Prevention (CDC) and the National Center for Health
Statistics (NCHS) National Vital Statistics System (NVSS) report deaths is a rate by 100,000 population. These
rates can be found reported in the NVSS mortality tables (16, 17) and CDC WONDER tables (18), among other
places.
The way the CDC and NCHS report deaths, as a rate, or proportion of a population, instead of, a number of
deaths, makes a lot of sense. A number of deaths by itself is meaningless. For example, is 500,000 deaths an
alarmingly high number of deaths? Well, it depends. If it is 500,000 deaths over a year, for a population that, at
the beginning of the year, had 1 million individuals, then it would mean half of the population died. But if it is
500,000 deaths over a year, for a population, that, at the beginning of the year, was 100 times that (50 million
individuals) then, it would mean 1% (1 per 100) of the population passed away that year.
Crude death rate per 100,000 population and percentage death rate (rate per 100 population) are defined as
follows:
   
   (Eq. 1)
   
   (Eq. 2)
Table 3: U.S. death rates per 100,000 population and per 100 population (or percentage death rates), years 1999-2019
Year
Deaths
Population
Crude Rate per 100,000
Percentage Death Rate
1999
2,391,399
279,040,168
857.0
0.9
2000
2,403,351
281,421,906
854.0
0.9
2001
2,416,425
284,968,955
848.0
0.8
2002
2,443,387
287,625,193
849.5
0.8
2003
2,448,288
290,107,933
843.9
0.8
2004
2,397,615
292,805,298
818.8
0.8
2005
2,448,017
295,516,599
828.4
0.8
2006
2,426,264
298,379,912
813.1
0.8
2007
2,423,712
301,231,207
804.6
0.8
2008
2,471,984
304,093,966
812.9
0.8
2009
2,437,163
306,771,529
794.5
0.8
2010
2,468,435
308,745,538
799.5
0.8
2011
2,515,458
311,591,917
807.3
0.8
2012
2,543,279
313,914,040
810.2
0.8
2013
2,596,993
316,128,839
821.5
0.8
2014
2,626,418
318,857,056
823.7
0.8
2015
2,712,630
321,418,820
844.0
0.8
2016
2,744,248
323,127,513
849.3
0.8
2017
2,813,503
325,719,178
863.8
0.9
2018
2,839,205
327,167,434
867.8
0.9
2019
2,854,838
328,239,523
869.7
0.9
Source: https://wonder.cdc.gov/ucd-icd10.html; tables organized by year, generated February 1st 2021.
Page 11 of 112
The first four columns from Table 3 were generated using CDC WONDER by choosing to organize the table
layout by year (18). The last column was computed using (Eq. 2). According to Table 3, U.S. death rates have
been from 0.8% to 0.9%, each year, from 1999 to 2019. A difference of 0.1 percent point in the death rate of
one year over another has been observed in the recent past and is thus to be expected.
A difference of 0.1 percent point in the death rate for a population of 330 million is 330,000 deaths (C 1). For a
population of 330 million with a death rate of 0.9%, the expected number of deaths is 3.0 million per year or
8,137 deaths per day (C 2-3). An additional 0.1 percent point in the death rate would bring the expected number
of deaths for a population of 330 million to 3.3 million per year or 9,041 deaths per day (C 4-5).
 
 deaths: 0.1 % of a population of 330 million (C 1)
 
  deaths per year (C 2)
 
  deaths per day (C 3)
 
  deaths per year (C 4)
 
  deaths per day (C 5)
Table 4: U.S. death rates per 100,000 population and per 100 population (or percentage death rates), years 2014-2020
Year
Deaths
Population
Crude Rate per 100,000
Percentage Death Rate
2014
2,646,843
317,344,917
834.1
0.8
2015
2,698,943
319,680,287
844.3
0.8
2016
2,731,848
321,969,165
848.5
0.8
2017
2,810,935
324,157,064
867.2
0.9
2018
2,839,076
325,991,287
870.9
0.9
2019
2,852,609
327,598,547
870.8
0.9
2020
3,382,503
329,197,954
1027.5
1.0
Source: https://www.census.gov/popclock/, consulted February 1st 2021, January 20th population numbers.
https://data.cdc.gov/browse , Weekly Counts of Deaths by State and Select Causes, for 2014-19. Page 2 of 2 of
“Provisional Death Counts for Coronavirus Disease 2019 (COVID-19) By Week of Death” table,
https://www.cdc.gov/nchs/nvss/vsrr/covid_weekly/index.htm, for 2020.
Because one needs a population number to compute a death rate against and because Table 3 does not show a
death rate for calendar year 2020, Table 4 uses different sources of information, to include a death rate for year
2020. Population data for Table 4 were retrieved from the U.S. Census Bureau Population Clock (6), for
January 20th of each year. Death data were computed as the sum of weekly All Cause deaths numbers for each
calendar year, from the CDC Weekly Counts of Deaths by State and Select Causes datasets (8, 10, 11).
Page 12 of 112
Table 5: U.S. death rates per 100,000 population and percentage death rates, season 2014-15 to season 2019-20
Season
Deaths
Population
Crude Rate per 100,000
Percentage Death Rate
2014-15
2,698,735
319,680,287
844.2
0.8
2015-16
2,695,241
321,969,165
837.1
0.8
2016-17
2,789,845
324,157,064
860.6
0.9
2017-18
2,843,558
325,991,287
872.3
0.9
2018-19
2,832,269
327,598,547
864.6
0.9
2019-20
3,093,240
329,197,954
939.6
0.9
Source: See Table 4.
Because deaths are cyclical (something further analyzed in the upcoming section), another standard way to
report deaths is over a season rather than a calendar year. For example, the CDC reporting period for influenza
seasons begins during Morbidity and Mortality Weekly Report (MMWR) week 40 and ends week 39 of the
following year, where “MMWR weeks” refers to the sequential numbering of weeks (Sunday through Saturday)
during a calendar year (12). In this paper, seasons are centered around the 2014-19 peaks of deaths, to facilitate
their comparison to the 2020 one, and address concerns that the same weeks were not compared across years
during the Nov. 11, 2020 webinar (1)see upcoming section about the seasonality of deaths. Each season has
a total of 52 weeks of reported deaths. A season starts MMWR week 33 and ends MMWR week 32except for
the 2014-15 season which starts MMWR week 34 and ends MMWR week 32. Table 5 uses the same data
source as Table 4, but does group U.S. deaths per season rather than calendar year. Season 2014-15 spans from
week ending 8/23/2014 to week ending 8/15/2015. Season 2015-16 spans from week ending 8/22/2015 to week
ending 8/13/2016. Season 2016-17 spans from week ending 8/20/2016 to week ending 8/12/2017. Season 2017-
18 spans from week ending 8/19/2017 to week ending 8/11/2018. Season 2018-19 spans from week ending
8/18/2018 to week ending 8/10/2019. Season 2019-20 spans from week ending 8/17/2019 to week ending
8/08/2020.
Population estimates from Table 3 (CDC WONDER table) are U.S. Census Bureau estimates of July 1 resident
populations (see Dataset Documentation, (18))mid-year: This makes sense. For example, over the calendar
year 1999, the US experienced 2,391,399 deaths. A total of 2,391,399 deaths experienced by a population of
279,040,168 individuals, means the death rate that year was 0.9%.
Now, because the objective of this analysis is to assess whether the total deaths number the U.S. experienced
during the course of the 2020 calendar year or 2019-20 season was unexpected or alarming, July 1, for the U.S.
population, cannot be the reference. Indeed, by July 1, 2020 the US population level would potentially have
been altered by COVID-19. So, instead, the date chosen for population levels in Tables 4-5, is January 20th
pre-COVID-19. The interpretation of the data does not change though. Over the calendar year 2020, the U.S.
experienced 3,382,503 deaths. Given that the U.S. started the year with a population of 329,197,54 individuals,
this means the U.S. death rate was 1.0%.
Page 13 of 112
The percentage death rates shown in Tables 3-5 range from 0.9% to 1.0%. Percentage death rates shown in
Table 3, which covers years from 1999 to 2019, went from 0.9% to 0.8%, and back to 0.9%, in the past. A
difference of 0.1 percent point in the death rate of one year over another has been observed in the recent past
and is thus to be expected. Tables 3-5’s population data and death rates show that total deaths experienced in the
U.S. during season 2019-20 and year 2020 were normal, within the U.S. historical context reviewed here.
Table 6: U.S. percentage death rates, percentage change in deaths, percentage change in population, years 1999-2019
Year
Deaths
change
% change
Population
change
% change
% Death Rate
1999
2,391,399
279,040,168
0.9
2000
2,403,351
11,952
0.5
281,421,906
2,381,738
0.9
0.9
2001
2,416,425
13,074
0.5
284,968,955
3,547,049
1.3
0.8
2002
2,443,387
26,962
1.1
287,625,193
2,656,238
0.9
0.8
2003
2,448,288
4,901
0.2
290,107,933
2,482,740
0.9
0.8
2004
2,397,615
-50,673
-2.1
292,805,298
2,697,365
0.9
0.8
2005
2,448,017
50,402
2.1
295,516,599
2,711,301
0.9
0.8
2006
2,426,264
-21,753
-0.9
298,379,912
2,863,313
1.0
0.8
2007
2,423,712
-2,552
-0.1
301,231,207
2,851,295
1.0
0.8
2008
2,471,984
48,272
2.0
304,093,966
2,862,759
1.0
0.8
2009
2,437,163
-34,821
-1.4
306,771,529
2,677,563
0.9
0.8
2010
2,468,435
31,272
1.3
308,745,538
1,974,009
0.6
0.8
2011
2,515,458
47,023
1.9
311,591,917
2,846,379
0.9
0.8
2012
2,543,279
27,821
1.1
313,914,040
2,322,123
0.7
0.8
2013
2,596,993
53,714
2.1
316,128,839
2,214,799
0.7
0.8
2014
2,626,418
29,425
1.1
318,857,056
2,728,217
0.9
0.8
2015
2,712,630
86,212
3.3
321,418,820
2,561,764
0.8
0.8
2016
2,744,248
31,618
1.2
323,127,513
1,708,693
0.5
0.8
2017
2,813,503
69,255
2.5
325,719,178
2,591,665
0.8
0.9
2018
2,839,205
25,702
0.9
327,167,434
1,448,256
0.4
0.9
2019
2,854,838
15,633
0.6
328,239,523
1,072,089
0.3
0.9
Source: See Table 3.
Table 7: U.S. percentage death rates, percentage change in deaths, percentage change in population, years 2014-2020
Year
Deaths
change
% change
Population
change
% change
% Death Rate
2014
2,646,843
317,344,917
0.8
2015
2,698,943
52,100
2.0
319,680,287
2,335,370
0.7
0.8
2016
2,731,848
32,905
1.2
321,969,165
2,288,878
0.7
0.8
2017
2,810,935
79,087
2.9
324,157,064
2,187,899
0.7
0.9
2018
2,839,076
28,141
1.0
325,991,287
1,834,223
0.6
0.9
2019
2,852,609
13,533
0.5
327,598,547
1,607,260
0.5
0.9
2020
3,382,503
529,894
18.6
329,197,954
1,599,407
0.5
1.0
2021
330,835,933
1,637,979
0.5
Source: See Table 4.
Page 14 of 112
Table 8: U.S. percentage death rates, percentage change in deaths and change in population, season 2014-15 to 2019-20
Season
Deaths
change
% change
Population
change
% change
% Death Rate
2014-15
2,698,735
319,680,287
0.8
2015-16
2,695,241
-3,494
-0.1
321,969,165
2,288,878
0.7
0.8
2016-17
2,789,845
94,604
3.5
324,157,064
2,187,899
0.7
0.9
2017-18
2,843,558
53,713
1.9
325,991,287
1,834,223
0.6
0.9
2018-19
2,832,269
-11,289
-0.4
327,598,547
1,607,260
0.5
0.9
2019-20
3,093,240
260,971
9.2
329,197,954
1,599,407
0.5
0.9
Source: See Table 5.
Further consider Tables 6-8. Tables 6-8 contain the same deaths and population data as Tables 3-5, and
consequently the same percentage death rates. Nothing has changedall the deaths, population and death rate
numbers from Table 3 and, now Table 6, are from the CDC’s WONDER table (18); deaths and population
numbers from Tables 4-5, and now Tables 7-8, are from the CDC (8, 10, 11) and the U.S. Census Bureau (6)
while death rates were computed based on those, the exact same way they are on the CDC WONDER tables
(18). Tables 6-8 include only the percentage death rates, to leave room for the change in deaths and population
numbers from one calendar or season, over another, and the expression of those changes in percentage terms.
Note, the January 20th 2021 population level is added to Table 7 (it was not included in Table 5).
Changes in deaths and population, as well as percentage changes in deaths and population were computed as
follows:
     (Eq. 3)
  
    (Eq. 4)
     (Eq. 5)
  
    (Eq. 6)
The first thing to notice is that the U.S. population has not decreased during calendar year 2020 (Table 7). In
other words, COVID-19 deaths have not decimated the U.S. population. Instead, in 2020, the U.S. population
increased by 0.5 percent over the previous yearjust like it did for the two years prior to that. It is the fact that
the US population has increased over the years, that explains the increase of deaths numbers over the years. The
more people, the more of them pass away. The change in population over the years is the reason why the
relevant statistics, to assess whether deaths experienced in 2020 were unexpected or alarming, are death rates.
A death rate is a deaths number in proportion to its population level. A death rate does account for the increase
in population over the years.
Page 15 of 112
Tables 7-8 show the percentage changes in population ranged from +0.7% for earlier years or seasons to + 0.5
this past three years. Those percentage changes are slightly different on Table 6, since the population numbers
they are based on, are slightly different. The percentage changes in deaths vary a lot more, ranging from -2.1%
from 2003 to 2004 (Table 6) to +2.9% from 2016 to 2017 (Table 7) or +3.5% from season 2015-16 to 2016-17
(Table 8). With + 9.2% (2018-19 to 2019-20, Table 8), or + 18.6% (2019 to 2020, Table 7), this past season and
year; these numbers would seem oddif it was not for the relevant percentage death rates telling us otherwise.
The emphasis is onif it was not for the relevant percentage death rates telling us otherwise.
In other words, while Table 8 shows a change of +260,971 deaths from season 2018-19 to season 2019-20 and
table 7 shows a change of +529,894 deaths from calendar year 2019 to calendar year 2020, given the level and
change in population over the same periods, those deaths numbers are normalas the respective death rates of
0.9% and 1.0% show.
The death rates show that the total deaths increase experienced in 2020 was not unexpected, nor alarming, but
rather, is explained by the population increase. Note, this result, in conjunction with the facts that (1) every
recorded death is allocated one exclusive cause, (2) a new cause was especially created for COVID-19 and, (3)
means to test and record COVID-19 deaths were considerably expanded, points to suspicions that some
reclassification between other causes of death and COVID-19 deaths might have occurred.
Also, note, it is not assumed, or even suggested, that such re-classification, if they occurred, were malicious or
intended to inflate COVID-19 numbers. Nonetheless, individuals within organizations do respond to financial
incentives and groupthink, to some extent. Similarly, it is not assumed, or even suggested, that mitigation
measures adopted were designed to harm targeted groups of the U.S. population. Nonetheless, the unequal toll
that those measures have had on individuals across our society cannot be ignored.
Page 16 of 112
3.1 How about excess deaths estimates?
Back in October 20, 2020, the CDC published estimates by the National Center for Health Statistics (NCHS) of
299,028 excess deaths from late January through October 3th 2020, with two thirds attributed to COVID-19”
(5). This begs the question: How can those results be reconciled with the above conclusions?
The objective of this analysis is to assess whether the total deaths number the U.S. experienced in 2020 was
unexpected or alarming. In contrast, Rossen et al. take as given that COVID-19 had effects of pandemic
proportion on U.S. deaths and question whether COVID-19 deaths numbers were underestimated. They then,
set themselves to produce excess deaths estimates, (5):
As of October 15, 216,025 deaths from coronavirus disease 2019 (COVID-19) have been reported in the
United States; however, this number might underestimate the total impact of the pandemic on mortality.
Measures of excess deaths have been used to estimate the impact of public health pandemics or
disasters, particularly when there are questions about underascertainment of deaths directly attributable
to a given event or cause.
The methodology used here is simple and transparent: Put publicly available CDC data of past deaths into
tabular and graphical forms and use simple logic and statistics to make sense of them. In contrast, Rossen et al.
produce estimates of excess deaths, using Farrington surveillance algorithms.
In the discussion part of their paper, Rossen and co-authors allude to one of the reasons why their results seem
to contradict the conclusions reached here that total deaths experienced in the U.S. in 2020 were not unexpected
nor alarming. Rossen et al. point out that, because of population growth over the years, using the average
number of deaths from past years as a benchmark, like they dooverestimates excess deaths, (5). This is the
fourth limitation of their findings:
The findings in this report are subject to at least five limitations. Fourth, using the average numbers of
deaths from past years might underestimate the total expected numbers because of population growth or
aging, or because of increasing trends in certain causes such as drug overdose mortality.
Because they defined excess deaths as “the number of persons who have died from all causes, in excess of the
expected number of deaths for a given place and time” (5), when they state in the discussion part of their paper
that they “might underestimate the expected number of deaths”, what they are saying is, they might have
overestimated excess deaths. In other words, the model used to produce their excess deaths estimates did not
account for the increase in population over the years. In contrast, by using death rates, which are simple
statistics anyone who cares to do so can compute, this analysis does account for the increase in population over
the years.
Page 17 of 112
4 Weekly deaths per age group: Constancy
Looking at the distribution of deaths across age groups is important. If a new disease is more fatal to one age
group compared to another (despite mitigation efforts adopted), this will be reflected in a higher percentage of
total deaths from that age group. If a new disease is equally fatal to all age groups, then the distribution of
deaths across age groups will remain constant.
Graph 3: Percentage of U.S. Deaths per Age Group, from MMWR week ending 2/1/2020 to week ending 11/14/2020
Source: Page 2 of 2 of “Provisional Death Counts for Coronavirus Disease 2019 (COVID-19) By Week of Death” table,
https://www.cdc.gov/nchs/nvss/vsrr/covid_weekly/index.htm. Downloaded 12-26-20.
Graph 3 is a 100% stacked column graph. It is used to compare the percentages across age groups that each
contributes towards the total weekly deaths. Each column represents 100% of all deaths experienced that week.
Each colored chunk of each column represents the portion or percentage of those deaths experienced by its
corresponding age group. The data for this graph were downloaded on December 26, 2020. Because “data for
the most recent 5 weeks are typically less than 90% complete, with lower levels of completeness in more recent
weeks” (9), those most recent 5 weeks of data were omitted from this graph. Graph 3 includes 42 weeks of data,
from MMWR Week 1 ending 02/01/2020 to MMWR Week 42 ending 11/14/2020. This time period covers the
weeks since the U.S. started recording COVID-19 deaths, March 24th 2020, and also pre-COVID weeks. For a
corresponding of week number (from 1 to 42) to the date the week ended, please see Table 9.
Page 18 of 112
Table 9: Week ending date (MM-DD-YY) corresponding to week number, Graph 3
season
2019-2020
Week #
1
2
3
4
5
6
7
8
9
10
11
12
13
ending date
02-01-20
02-08-20
02-15-20
02-22-20
02-29-20
03-07-20
03-14-20
03-21-20
03-28-20
04-04-20
04-11-20
04-18-20
04-25-20
Week #
14
15
16
17
18
19
20
21
22
23
24
25
26
ending date
05-02-20
05-09-20
05-16-20
05-23-20
05-30-20
06-06-20
06-13-20
06-20-20
06-27-20
07-04-20
07-11-20
07-18-20
07-25-20
Week #
27
28
29
30
31
32
33
34
35
36
37
38
39
ending date
08-01-20
08-08-20
08-15-20
08-22-20
08-29-10
09-05-20
09-12-20
09-19-20
09-26-20
10-03-20
10-10-20
10-17-20
10-24-20
Week #
40
41
42
ending date
10-31-20
11-07-20
11-14-20
Page 19 of 112
Notice Graph 3’s pattern of total deaths across age groups: The percentages of total deaths from each age group
are fairly constant throughout all those weeks. The chunks of blue columns have approximately the same height
throughout all those weeks. This means that every week, 30% of all deaths are from individuals who were 85
years old or older. That is true for the other age groups as well. Throughout the weeks, every week, 25% of all
deaths are from individuals who were from 75 to 84 years old. And so on.
Table 10: Average Weekly U.S. Deaths from MMWR week ending 2/1/2020 to week ending 11/14/20, per Age Group
Age Group
Average Weekly Deaths
Percentage
All ages
62,084
100%
85 and over
18,473
30%
75-84 years
14,985
24%
65-74 years
12,344
20%
55-64 years
8,152
13%
45-54 years
3,565
6%
35-44 years
1,962
3%
25-34 years
1,392
2%
15-24 years
681
1%
14 and under
531
1%
Source: See Graph 3.
Every week, all age groups experience deaths. No one age group escapes this reality. Table 10 indicates that
531 individuals, 14 years and younger, pass away every week in the United States. An average of 62,084
individuals died per week in the United States, between week ending 02/01/2020 and week ending 11/14/2020.
The good news is, every week, a much lower number of younger individuals pass away than the number of
older individuals who pass away. That is good news.
Data from Graph 3 and Table 9 point to the reason a higher number of COVID-19 deaths has been reported
among older individuals than younger individuals, is simply because, every day in the U.S., COVID-19 or not,
those aged 85 or older are at a higher risk of dying than those aged 84 or younger. This should come as no
surprise to anyone.
Family and friends mourn the passing of their loved ones, no matter their age at the time of their passing. The
passing of a loved one can be, devastating. Nonetheless, as a society, the passing of older individuals should not
be something to dread or fear, but instead, to celebratelong lives lived are something to be grateful for and
worth celebrating.
Page 20 of 112
While the general pattern of Graph 3 has not changed over the weeks shown, those of you with eagle eyes might
have noticed the percentage of weekly deaths from the 85 and over age group increased to reach a high of
32.1%, week 12, which corresponds to MMWR week ending 04/18/2020the week with the highest number of
total deaths of all ages in spring 2020. Could this change in the age distribution of deaths, at the U.S. level,
reflect the abrupt increase in deaths from New York City and the State of New Jersey experienced during those
weeks (see US deaths per jurisdiction section)? More work is needed to determine what effect COVID-19 and
mitigation efforts have had on deaths, on how it might have differed, across age groups, as well as U.S.
jurisdictions.
In the next section, the seasonality of weekly deaths across causes is addressed.
Page 21 of 112
5 Weekly deaths per cause: Seasonality
5.1 Seasonality of Deaths across Causes
Graph 4: U.S. Deaths per Week and Cause, from MMWR week ending 1/11/2014 to MMWR week ending 11/14/2020
01/10/15
61,883
03/12/16
56,232
01/14/17
61,124
01/13/18
67,664
03/09/19
58,683
04/11/20
78,999
47,888
08/08/15
49,382
07/23/16
49,493
07/29/17
50,432
07/28/18
51,139
08/17/19
57,634
10/17/20
46,479
08/30/14
Page 22 of 112
Graph 4 is the plot of weekly U.S. deaths per cause, from MMWR week ending 1/11/2014 to MMWR week
ending 11/14/2020, a total of 358 weeks. The blue line is the plot of weekly deaths from all causes. The first
two lines below that, are the plots of weekly deaths due to heart diseases and cancers, the two main causes of
death in the United States. Note, the third leading cause of death is respiratory diseases. The bottom cluster of
lines are plots of weekly deaths due to other select causes on record in the datasets used here. Those datasets,
labeled “Weekly Counts of Deaths by State and Select Causes”, are publicly available CDC datasets and the
same ones used to get Tables 4-5 and 7-8’s deaths numbers (10, 11).
Weekly U.S. deaths numbers, highs and lows, seem to be seasonal, occurring at similar time intervals, over the
period shown on Graph 4. The All Cause weekly U.S. deaths blue line shows an upward sloping trend, which is
consistent with the increasing deaths numbers over the years shown in Tables 3-8. This trend can most clearly
be identified by the troughs consistently going up, despite the peaks going up and down, over the years.
From Graph 4, the seasonality of deaths is more striking with the All Cause line. It can also be discerned for
deaths due to heart diseases, chronic lower respiratory diseases as well as influenza and pneumonia. The scale
of the graph, chosen to accommodate All Cause deaths numbers, as well as its clutter, might obscure seasonality
of deaths occurring for other select causes. Next, in Graphs 5-10, this possibility is investigated. To get a clearer
picture of this seasonality, the data series corresponding to the newly introduced category of death “COVID-19”
is taken off those graphs.
Page 23 of 112
Graph 5: Heart diseases and cancers, U.S. Weekly Deaths from MMWR week ending 1/11/2014 to MMWR week ending 11/14/2020
change in scale
Page 24 of 112
Graph 6: Less than 4,500 deaths/week causes, U.S. Weekly Deaths from MMWR week ending 1/11/2014 to MMWR week ending 11/14/2020
Page 25 of 112
Graph 7: Chronic lower respiratory diseases and influenza & pneumonia, U.S. Weekly Deaths from week ending 1/11/2014 to 11/14/2020
Page 26 of 112
Graph 8: Cerebrovascular diseases, Alzheimer’s, and diabetes U.S. Weekly Deaths from MMWR week ending 1/11/2014 to MMWR week ending 11/14/2020
break in scale
Page 27 of 112
Graph 9: Nephritis, nephrotic syndrome & nephrosis, septicemia, and other diseases of the respiratory system, U.S. Weekly Deaths, 1/11/14-11/14/20
change in scale
Page 28 of 112
Graph 10: Symptoms, signs & abnormal clinical & laboratory findings, not elsewhere classified; U.S. Weekly Deaths, weeks ending 1/11/14-11/14/20
change in scale
Page 29 of 112
Graph 5 takes a closer look at deaths due to heart diseases and cancers. Graph 6, which has the plots of data
series for causes with less than 4,500 deaths per week, is further declutter and/or rescaled on Graphs 7-10.
Graphs 5-10 reveal that all old select causes of deaths are characterized by the seasonality first observed on
Graph 4, with peaks of deaths occurring during the same weeks across years and causes of death. This is quite
amazing.
Deaths have followed a pattern of highs and lows, year after year, and the magnitude of those highs and lows,
one year relatively to another, has been consistent for each and every select cause of death. In other words, the
peaks of deaths, from all the different causes, have coincide (happened over the same weeks)year after year,
and the magnitude of those peaks for one year relative to another, for each cause, has also been consistent
across causes. But not for 2020. Why?
Does the historically observed seasonality point to general conditions, usually reoccurring in winters, that make
individuals more vulnerable and at a higher risk of dying, no matter what their underlying health conditions may
be? Could some of those general conditions have prevailed “off-season”, in 2020? Could have mitigation efforts
artificially created some of those general conditions, which are harmful to individuals’ health and increase their
risk of dying?
Table 11: Copy of Table from Nov. 11, 2020 COVID-19 Deaths: A Look at U.S. Data, webinar
Page 30 of 112
For each cause, the magnitude of those peaks for one year relative to another, being consistent across causes,
means that, when the 2020 peak of weekly deaths caused by heart diseases is smaller than its 2018 one, while
the 2020 peak for All Cause is higher than its 2018 one, it is not consistent with observed historical patterns
and it points to an anomaly. This prompted the question, where have all the heart attacks gone? (1)
During the webinar this paper is based on, Table 11 was presented. Table 11 looks at the changes in deaths
numbers from select causes from one week over another, over the three weeks the U.S. experienced the highest
COVID-19 deaths numbers. The sum of the decreases in deaths numbers across causes was then compared to
the number by which COVID-19 deaths exceeded the deaths due to heart diseases. And lo and behold, those
two nearly matchedwhich gave suspicions that some re-classification between various select causes of death
and COVID-19 might have occurred. Many found this result intriguing while, for some, this measure did not
make sense. Dr. Shrestha argued it was just a coincidence (3). A few asked whether such results held true for
later weeks too.
Now, these results cannot possibly hold true for weeks-on, because the number of COVID-19 deaths did not
continue to increase week after week, and the number of weekly deaths due to other select causes never goes to
zero (or below of course). There is a seasonality to the number of deaths the U.S. experiences from all the
different select causes reviewed herebut no matter the cause, the deaths numbers never go to zero. That is
worth emphasizing. The lowest number of weekly deaths due to respiratory diseases (influenza and pneumonia
+ chronic lower respiratory diseases + other diseases of the respiratory system) has been 3,699 deaths per week
over the 358 weeks period from MMWR week ending 1/11/2014 to MMWR week ending 11/14/2020. Why
should we expect weekly deaths due to COVID-19 to go to zero?
An alternative to the measure adopted in Table 11 must be found. As stated before, because there is,
historically, synchronicity of the relative increases of death peaks across year, for all causes, one death peak (for
a specific select cause) not increasing as much as expected could be an indication that some re-classification,
between this select cause and the newly introduced COVID-19 cause, occurred. On the other hand, could, one
death trough (for a specific select cause) not decreasing as much as expected, indicate that mitigation efforts
(isolation, lockdowns, …) have taken a toll on individuals and put them at a higher risk of dying from such
specific select causes, such as diabetes or suicides?
To get a clearer picture of those potential changes, that might have occurred, we turn to the comparison of
weekly deaths per cause, across seasons. Doing so also facilitates the comparison of the past peaks of deaths to
the 2020 one, and addresses concerns that the same weeks were not compared across years, during the Nov. 11,
2020 seminar (1).
Page 31 of 112
5.2 Causes of Deaths across Seasons
Although weekly deaths exhibit seasonal patterns (see sub-section on Seasonality of Deaths across Causes),
every peak of weekly deaths and every trough does not occur at the exact same week year after year.
Nonetheless, the peaks have occurred in the winter and the troughs have occurred in the summer: From 2014 to
2019, peaks occurred from MMWR week ending January 10th to MMWR week ending March 12th, and troughs
occurred from MMWR week ending July 23rd to MMWR week ending August 30th. But in 2020 (see Graph 4),
the peak of deaths occurred MMWR week ending April 11th and the trough occurred MMWR week ending
October 17th. This in itself is puzzling. This could also mean comparing the 2020 peak to previous ones might
be misleading, as it occurred “off-season”. To address this, instead of comparing peaks to peaks, seasons are
compared to seasons. The seasons are delineated by defining their centers around the 2014-19 peaks of deaths
since the objective here is to assess whether the 2020 peak was unexpected or alarming.
All seasons have a total of 52 weeks of reported deaths. A season starts MMWR week 33 and ends MMWR
week 32except for the 2014-15 season which starts MMWR week 34 and ends MMWR week 32. Season
2014-15 spans from week ending 8/23/2014 to week ending 8/15/2015. Season 2015-16 spans from week
ending 8/22/2015 to week ending 8/13/2016. Season 2016-17 spans from week ending 8/20/2016 to week
ending 8/12/2017. Season 2017-18 spans from week ending 8/19/2017 to week ending 8/11/2018. Season 2018-
19 spans from week ending 8/18/2018 to week ending 8/10/2019. Season 2019-20 spans from week ending
8/17/2019 to week ending 8/08/2020.
Page 32 of 112
Table 12a: Week ending date (MM-DD-YY) corresponding to week number on Graphs 2, 11-15, 19-71 and 73, for each season
season
2014-2015
Week #
1
2
3
4
5
6
7
8
9
10
11
12
13
ending date
08-23-14
08-30-14
09-06-14
09-13-14
09-20-14
09-27-14
10-04-14
10-11-14
10-18-14
10-25-14
11-01-14
11-08-14
11-15-14
Week #
14
15
16
17
18
19
20
21
22
23
24
25
26
ending date
11-22-14
11-29-14
12-06-14
12-13-14
12-20-14
12-27-14
01-03-15
01-10-15
01-17-15
01-24-15
01-31-15
02-07-15
02-14-15
Week #
27
28
29
30
31
32
33
34
35
36
37
38
39
ending date
02-21-15
02-28-15
03-07-15
03-14-15
03-21-15
03-28-15
04-04-15
04-11-15
04-18-15
04-25-15
05-02-15
05-09-15
05-16-15
Week #
40
41
42
43
44
45
46
47
48
49
50
51
52
ending date
05-23-15
05-30-15
06-06-15
06-13-15
06-20-15
06-27-15
07-04-15
07-11-15
07-18-15
07-25-15
08-01-15
08-08-15
08-15-15
season
2015-2016
Week #
1
2
3
4
5
6
7
8
9
10
11
12
13
ending date
08-22-15
08-29-15
09-05-15
09-12-15
09-19-15
09-26-15
10-03-15
10-10-15
10-17-15
10-24-15
10-31-15
11-07-15
11-14-15
Week #
14
15
16
17
18
19
20
21
22
23
24
25
26
ending date
11-21-15
11-28-15
12-05-15
12-12-15
12-19-15
12-26-15
01-02-16
01-09-16
01-16-16
01-23-16
01-30-16
02-06-16
02-13-16
Week #
27
28
29
30
31
32
33
34
35
36
37
38
39
ending date
02-20-16
02-27-16
03-05-16
03-12-16
03-19-16
03-26-16
04-02-16
04-09-16
04-16-16
04-23-16
04-30-16
05-07-16
05-14-16
Week #
40
41
42
43
44
45
46
47
48
49
50
51
52
ending date
05-21-16
05-28-16
06-04-16
06-11-16
06-18-16
06-25-16
07-02-16
07-09-16
07-16-16
07-23-16
07-30-16
08-06-16
08-13-16
season
2016-2017
Week #
1
2
3
4
5
6
7
8
9
10
11
12
13
ending date
08-20-16
08-27-16
09-03-16
09-10-16
09-17-16
09-24-16
10-01-16
10-08-16
10-15-16
10-22-16
10-29-16
11-05-16
11-12-16
Week #
14
15
16
17
18
19
20
21
22
23
24
25
26
ending date
11-19-16
11-26-16
12-03-16
12-10-16
12-17-16
12-24-16
12-31-16
01-07-17
01-14-17
01-21-17
01-28-17
02-04-17
02-11-17
Week #
27
28
29
30
31
32
33
34
35
36
37
38
39
ending date
02-18-17
02-25-17
03-04-17
03-11-17
03-18-17
03-25-17
04-01-17
04-08-17
04-15-17
04-22-17
04-29-17
05-06-17
05-13-17
Week #
40
41
42
43
44
45
46
47
48
49
50
51
52
ending date
05-20-17
05-27-17
06-03-17
06-10-17
06-17-17
06-24-17
07-01-17
07-08-17
07-15-17
07-22-17
07-29-17
08-05-17
08-12-17
Page 33 of 112
Table 12b: Week ending date corresponding to week number on Graphs 2, 11-15, 19-71 and 73, for each season
season
2017-2018
Week #
1
2
3
4
5
6
7
8
9
10
11
12
13
ending date
08-19-17
08-26-17
09-02-17
09-09-17
09-16-17
09-23-17
09-30-17
10-07-17
10-14-17
10-21-17
10-28-17
11-04-17
11-11-17
Week #
14
15
16
17
18
19
20
21
22
23
24
25
26
ending date
11-18-17
11-25-17
12-02-17
12-09-17
12-16-17
12-23-17
12-30-17
01-06-18
01-13-18
01-20-18
01-27-18
02-03-18
02-10-18
Week #
27
28
29
30
31
32
33
34
35
36
37
38
39
ending date
02-17-18
02-24-18
03-03-18
03-10-18
03-17-18
03-24-18
03-31-18
04-07-18
04-14-18
04-21-18
04-28-18
05-05-18
05-12-18
Week #
40
41
42
43
44
45
46
47
48
49
50
51
52
ending date
05-19-18
05-26-18
06-02-18
06-09-18
06-16-18
06-23-18
06-30-18
07-07-18
07-14-18
07-21-18
07-28-18
08-04-18
08-11-18
season
2018-2019
Week #
1
2
3
4
5
6
7
8
9
10
11
12
13
ending date
08-18-18
08-25-18
09-01-18
09-08-18
09-15-18
09-22-18
09-29-18
10-06-18
10-13-18
10-20-18
10-27-18
11-03-18
11-10-18
Week #
14
15
16
17
18
19
20
21
22
23
24
25
26
ending date
11-17-18
11-24-18
12-01-18
12-08-18
12-15-18
12-22-18
12-29-18
01-05-19
01-12-19
01-19-19
01-26-19
02-02-19
02-09-19
Week #
27
28
29
30
31
32
33
34
35
36
37
38
39
ending date
02-16-19
02-23-19
03-02-19
03-09-19
03-16-19
03-23-19
03-30-19
04-06-19
04-13-19
04-20-19
04-27-19
05-04-19
05-11-19
Week #
40
41
42
43
44
45
46
47
48
49
50
51
52
ending date
05-18-19
05-25-19
06-01-19
06-08-19
06-15-19
06-22-19
06-29-19
07-06-19
07-13-19
07-20-19
07-27-19
08-03-19
08-10-19
season
2019-2020
Week #
1
2
3
4
5
6
7
8
9
10
11
12
13
ending date
08-17-19
08-24-19
08-31-19
09-07-19
09-14-19
09-21-19
09-28-19
10-05-19
10-12-19
10-19-19
10-26-19
11-02-19
11-09-19
Week #
14
15
16
17
18
19
20
21
22
23
24
25
26
ending date
11-16-19
11-23-19
11-30-19
12-07-19
12-14-19
12-21-19
12-28-19
01-04-20
01-11-20
01-18-20
01-25-20
02-01-20
02-08-20
Week #
27
28
29
30
31
32
33
34
35
36
37
38
39
ending date
02-15-20
02-22-20
02-29-20
03-07-20
03-14-20
03-21-20
03-28-20
04-04-20
04-11-20
04-18-20
04-25-20
05-02-20
05-09-20
Week #
40
41
42
43
44
45
46
47
48
49
50
51
52
ending date
05-16-20
05-23-20
05-30-20
06-06-20
06-13-20
06-20-20
06-27-20
07-04-20
07-11-20
07-18-20
07-25-20
08-01-20
08-08-20
Page 34 of 112
Next, Graphs 11 through 15 are reviewed.
Graphs 11-15 are plots of weekly deaths per cause and per season, from season 2014-15 to season 2019-20.
Because each of the seasons has a total of 52 weeks of reported deaths, Graphs 11-15’s horizontal axes show
weeks numbered 1 to 52. For a corresponding of week number (from 1 to 52) to the date the week ended, for
each season, please see Tables 12a-12b.
Page 35 of 112
Graph 11: U.S. All Cause Weekly Deaths per Season, 2014-15 to 2019-20 and COVID-19’s for 2019-20 season
Week # 35: Week ending 04/11/20
https://www.cdc.gov/nchs/data/nvss/coronavirus/Alert-2-
New-ICD-code-introduced-for-COVID-19-deaths.pdf
March 24, 2020
Week # 22: Week ending 01/13/18
All Cause deaths
Page 36 of 112
Graph 11
Graph 11 is the plot of the U.S. All Cause weekly deaths per season, from season 2014-15 to season 2019-20.
The 2019-20 peak occurred week ending April 11th 2020 (in spring, which is unusual) while the 2018-20
occurred week ending January 13th 2018 (in winter, which is usual). What is most striking though is that the
2019-20 weekly deaths were following the pattern of deaths from previous years, very closelyuntil March 24.
Do you see that?very closely, until March 24.
March 24, 2020 is the date a new code was introduced to record COVID-19 deaths (15). Prior to March 24,
there was no category or cause of death code devoted to keeping track of deaths due to COVID-19. This was a
new virus, so it makes sense there was no code for it, before. Still, other categories or causes of deaths due to
respiratory diseases already existed: Chronic lower respiratory diseases, influenza and pneumonia, and other
diseases of the respiratory system. The CDC performs a very important task, collecting health and death data.
Although new codes and categories need to be introduced, they can make the interpretation of the data collected
more difficultespecially if the use of a new code or category of death is heavily incentivized. The departure
from the historical pattern of deaths, after March 24, 2020 is so abrupt though, there has to be other
explanations!
Doesn’t it?
Page 37 of 112
Graph 12: U.S. Weekly Deaths per group of Causes, 2014-15 to 2019-20 season
All Cause deaths
Other Old Natural Cause deaths
= Natural Cause
Old Select Cause COVID-19
Old Select Cause deaths
Non-Natural Cause deaths (which includes accidents, homicides, suicides)
= All Cause Natural Cause
Page 38 of 112
Graph 12
To Graph 12, a 2019-20 without COVID-19 series, is added to the All Cause deaths graph. Below that, are the
following plots: Old Select Cause deaths, Other Old Natural Cause deaths and Non-Natural Cause deaths. Old
Select Cause deaths follow the exact same patterns as the All Cause deaths, without COVID-19. Old Select
Cause are the main natural causes of death, and together they have consistently accounted for 70-71% of All
Cause deaths, so it makes sense that their deaths would drive the pattern of All Cause deaths. The Other Old
Natural Cause deaths seem to be showing a higher level of deaths than expected, for the second part of 2019-20
season (compared to the first part of the season). Non-Natural Cause deaths seem to be growing over the years,
with an end of the season’s level higher than the level of deaths experienced at the beginning of the season,
season after seasonbut that increase seems to be higher for the 2019-20 season.
The increase in Non-Natural Cause deaths over the years would be consistent with Roosen et al.’s remark about
“increasing trends in certain causes such as drug overdose mortality” (5). Would the higher increase of Non-
Natural Cause deaths in 2019-20 give support to the assessment that the cure has been worse than the virus? Or
only point to the increased number of deaths experienced year after year? Those are questions that will need to
be answered.
Page 39 of 112
Graph 13: U.S. Weekly Deaths per Season, for leading causes and main diseases of the respiratory system, 2014-20
All Cause deaths
Deaths due to heart diseases.
Deaths due to cancers.
Deaths due to influenza and pneumonia.
Deaths due to chronic lower respiratory diseases.
Page 40 of 112
Graph 13
Each of Graph 13 through Graph 15 first displays a copy of Graph 11 so the synchronicity of All Cause deaths
peak and other select causes deaths peaks, over the years, can more easily be identified. Below that, are plots of
deaths due to select causes from season 2014-15 to season 2019-20.
Graph 13 shows that the 2020 peak of deaths due to heart diseases is lower than its 2018’s peak, even though
heart diseases are the leading cause of deaths in the U.S. and heart diseases deaths peaks have followed the All
Cause deaths peaks pattern in the past (highs/lows and relatively magnitude across years). In other words, a
continuation of historical patterns would have dictated a higher peak of heart diseases deaths in 2020 than its
2018 peak. Now, the 2020 All Cause deaths peak occurred off season”—but so does the heart diseases one
and at exactly the same week (week ending April 11, 2020). Why? Note, after that “off-season” 2020 peak, the
level of deaths due to heart diseases does not go back down to its pre-March 24, 2020 level. Why?
Although a seasonality of deaths due to cancers was observed on Graph 5, it was not as pronounced as other
causes of death. This is also seen here, middle section of Graph 13. In addition, Graph 13 shows no noticeable
change in the pattern of deaths due to cancers, from previous seasons.
Below cancers, on Graph 13, deaths due to chronic lower respiratory diseases, not only, do not show a 2020
peak at all, but reach levels lower than they ever had before, come late spring. Because both COVID-19 and
chronic lower respiratory diseases are respiratory diseases, this would be evidence that some re-classification
between the two causes of deaths might have occurred.
Deaths due to influenza and pneumonia (last cause of death on Graph 13), like deaths due to heart diseases, also
display an “off-season” peak for 2019-20, which is much lower than its previous two peaks, and which is
synchronous with All Cause peak (and thus COVID-19’s). Unlike deaths due to heart diseases, after their peak,
the deaths due to chronic lower respiratory diseases go back to their previous seasons’ levels. Because COVID-
19, influenza and pneumonia are all respiratory diseases, and because historical patterns would have dictated a
higher peak of deaths due to influenza and pneumonia in 2020 than its 2018 peak, would this be evidence that
some re-classification between the two causes of deaths might have occurred?
Page 41 of 112
Graph 14: U.S. Weekly Deaths per Season, for select causes deaths between 1,000 and 3,500; 2014-15 to 2019-20
All Cause deaths
Deaths due to cerebrovascular diseases.
Deaths due to Alzheimer disease.
Deaths due to diabetes.
Page 42 of 112
Graph 14
Graph 14 reveals that during the 2020 deaths peak, deaths due to cerebrovascular diseases and Alzheimer
disease nearly reached their previously highest peak level, season 2017-18, while deaths due to diabetes well
surpassed it. Again, all those 2020 “off-season” peaks occurred over the same weeks. Note, after their “off-
season” 2020 peak, the levels of deaths due to those select causes do not go back down to their pre-March 24,
2020 level. Historical patterns would have dictated a higher peak of deaths due to cerebrovascular diseases and
Alzheimer’s in 2020 than their 2018 peak. Could this also be evidence of some re-classification between those
causes of deaths and COVID-19? Historical patterns would also have dictated a more decisive return to lower
level of deaths consistent with historical seasonal levels. Finally, why does the level of deaths due to diabetes
seem to be higher than historical patterns suggest it should have been?
Now is a good time to remember that the All Cause deaths level for 2019-20 is not unexpected nor alarming:
See earlier section about U.S. deaths put into its short-term historic context. Nonetheless, the changes in the
patterns of deaths are worth inquiring about.
One more thing to note on Graph 14, is about Alzheimer deaths2014-15 season. For the first part of this
2014-15 season, Alzheimer deaths are distinctively lower than other seasons. This was also visible on Graph 8,
where the earlier trough is much lower than later ones (see the Alzheimer series at the bottom of Graph 8).
Starting in 2015, the number of Alzheimer deaths is distinctively higher. Could this be due to the adoption of
new means (improved tests/methodology) that allowed better detection of Alzheimer at the time? This is very
important to note: An increased deaths number due to a specific cause is not necessarily solely due to the
disease being more prevalent but can also be due to the availability of new methodology and increased means
devoted to its detection and recording, as well as incentives to do both. Could the sharp increase in U.S.
COVID-19 deaths numbers, be partly explained by the corresponding sharp increase, in financial means and
manpower, devoted to developing and implementing methodology to detect them, and to recording them? As
well as the pressure to do so?
Page 43 of 112
Graph 15: U.S. Weekly Deaths per Season, for select causes with less than 1,500 deaths per week, 2014-15 to 2019-20
All Cause deaths
Deaths due to nephritis, nephrotic syndrome and nephrosis.
Deaths due to other diseases of the respiratory system.
Deaths due to septicemia.
Deaths due to symptoms, signs and abnormal clinical and laboratory findings, not elsewhere classified.
Page 44 of 112
Graph 15
Graph 15 reveals small and short lived “off-season” peaks of deaths due to nephritis, nephrotic syndrome and
nephrosis, as well as due to septicemia. After their peaks, those deaths go back to their previous seasons’ levels.
Their 2020 peaks did not reach their 2018 peaks though, which, again, is unexpected given historical trends.
Deaths due to other diseases of the respiratory system (also on Graph 15) reached an “off-season” peak on par
with their 2018’s one. After their peak, those deaths go back to their previous seasons levelsalthough at a
lower level than its pre-March 24, 2020 level (not as noticeable as it was for deaths due to influenza and
pneumonia on Graph 13). Because COVID-19 and other diseases of the respiratory system are all respiratory
diseases, and because historical patterns would have dictated a higher peak of deaths due to other diseases of the
respiratory system in 2020 than its 2018 peak, would this be evidence that some re-classification between the
two causes of death might have occurred?
The last plot on Graph 15 is that of deaths due to symptoms, signs and abnormal clinical and laboratory
findings, not elsewhere classified. The 2019-20 line distinctively departs from usual patterns in the second half
of the season. (This was visible on Graph 10, as well.) If a new COVID-19 category had not been implemented,
this could have been interpreted as deaths due to unfamiliar and/or not elsewhere classifiable COVID-19
symptomsthis cannot be the case since we have had a new code for COVID-19 deaths since March 24, 2020.
Dr. Rossen explains that many of these still provisional deaths due to symptoms, signs and abnormal clinical
and laboratory findings, not elsewhere classified, will go on to be coded as external causes of death. External
causes of death include accidents, assault (homicides), intentional self-harm (suicides) and injuries of
undetermined intent (21). Some will also be updated to various natural causes as well (4).
Could those deaths due to symptoms, signs and abnormal clinical and laboratory findings, not elsewhere
classified, reflect the unprecedented state of stress, anxiety, anger, depression, and incomprehension, that the
COVID-19 campaigns of fear and isolation has brought on individuals who usually find health, meaning, and
thrive, through positive interactions with others?
In the last section, the variability of weekly deaths across jurisdictions is explored.
.
Page 45 of 112
6 Weekly deaths per jurisdiction: Variability
Graph 16: All Cause Deaths per Week and U.S. Jurisdiction, from MMWR week ending 1/11/2014 to MMWR week ending 11/14/2020. Vertical axis scaled with
minimum bound of 0 and maximum bound of 8,000 on all plots.
Page 46 of 112
Graph 16 displays the plots of All Cause deaths per weekfor all U.S. jurisdictions, over the period from
MMWR week ending 1/11/2014 to MMWR week ending 11/14/2020, a total of 358 weeks. For additional ease
of comparison between jurisdictions, all plots used the same scale on their vertical axes, with a minimum bound
of 0 and a maximum bound of 8,000 deaths. The drawback from this graph’s format is that the seasonality of
deaths cannot be discerned as well as it could on Graph 4, especially for lower deaths numbers’ jurisdictions.
Nonetheless, Graph 16 gives a great overview of the variability of deaths across jurisdictions.
Where is Waldo? is usually a game of finding a needle in a hay stackbut, in this case, the 2020 deaths peak
anomaly clearly stands out. Which jurisdiction would that be? You are invited to consult Table 14, which gives
the jurisdictions’ names corresponding to the plots displayed on Graph 16, to find out.
Where are the most populated jurisdictions?
Do all jurisdictions show a 2020 deaths peak?
Do all jurisdictions show a seasonality of deaths?
How many jurisdictions show an alarmingly high 2020 deaths peak?
How many jurisdictions do not show any 2020 deaths peak at all?
Graphs 17-69 (see Appendix) are individual plots of All Cause weekly deaths per season, from season 2014-15
to season 2019-20, for each U.S. jurisdiction. Jurisdictions include the 50 U.S. states, plus District of Columbia,
New York City and Puerto Rico territory. The graphs are presented in alphabetic order of the U.S. jurisdictions’
names. Readers are encouraged to consult the graph for their jurisdiction of residence or any jurisdiction(s) of
special interest to them. Graphs 17-69 take a closer look at the variability first outlined by Graph 16.
Did you find the anomaly on Graph 16? Or maybe a couple of them?
Were the numbers of COVID-19 weekly deaths, ever, at a disturbingly high level, for jurisdictions other than
New York City and the State of New Jersey? Would the variability of weekly deaths and 2020 peaks, observed
across states within the United States, also be observed across counties within each state? This is something that
every state and every county would need to investigate, if they have not done so already.
Page 47 of 112
Table 14: U.S. Jurisdictions corresponding to Plots in Graph 16.
Page 48 of 112
Table 15: Weekly deaths per jurisdiction, week ending 01/11/14-03/21/20 and 04/11/20-04/25/20 periods
Pre-March 24, 2020: Week ending 01/11/14-03/21/20
3-week peak deaths 2019-20: Week ending 04/11/20-04/25/20
United States
53,275
United States
76,505
California
5,073
California
6,037
Florida
3,871
New York City
5,936
Texas
3,805
Florida
4,505
Pennsylvania
2,583
New Jersey
4,491
Ohio
2,311
Texas
4,339
Illinois
2,019
New York
4,289
New York
1,900
Pennsylvania
3,671
Michigan
1,836
Michigan
3,112
North Carolina
1,775
Illinois
2,930
Georgia
1,590
Ohio
2,705
New Jersey
1,398
Massachusetts
2,354
Tennessee
1,394
North Carolina
2,049
Virginia
1,280
Georgia
2,031
Indiana
1,244
Virginia
1,617
Missouri
1,214
Indiana
1,577
Massachusetts
1,124
Tennessee
1,512
Arizona
1,107
Missouri
1,418
Washington
1,069
Maryland
1,416
New York City
1,043
Louisiana
1,376
Alabama
982
Arizona
1,368
Wisconsin
968
Connecticut
1,324
Maryland
935
Washington
1,262
South Carolina
919
Alabama
1,138
Kentucky
909
Colorado
1,089
Louisiana
868
South Carolina
1,070
Minnesota
837
Wisconsin
1,015
Oklahoma
746
Kentucky
986
Colorado
735
Minnesota
965
Oregon
693
Oklahoma
810
Arkansas
602
Oregon
768
Mississippi
599
Mississippi
727
Connecticut
596
Arkansas
629
Iowa
573
Iowa
605
Puerto Rico
570
Puerto Rico
568
Kansas
499
Nevada
566
Nevada
477
Kansas
537
West Virginia
430
West Virginia
436
Utah
355
New Mexico
408
New Mexico
344
Utah
391
Nebraska
324
Nebraska
350
Maine
275
Maine
300
Idaho
260
Idaho
284
New Hampshire
231
Rhode Island
272
Hawaii
218
New Hampshire
262
Rhode Island
198
Delaware
244
Montana
192
District of Columbia
212
Delaware
172
Hawaii
211
South Dakota
154
Montana
198
North Dakota
135
South Dakota
159
District of Columbia
115
North Dakota
158
Vermont
110
Vermont
125
Wyoming
86
Wyoming
96
Alaska
82
Alaska
75
Page 49 of 112
Prior to March 24, 2020, the U.S. jurisdictions with the highest average weekly deaths were California, Florida
and Texas (see Table 15). That makes sense. These correspond to the three most highly populated US
jurisdictions and the higher the population, the higher the weekly deaths. Deaths, after all, are part of life.
Similarly, some of the least populated U.S. jurisdictions experience the lowest average weekly deaths numbers.
Those are Vermont, Wyoming and Alaska.
From week ending April 11, 2020 to week ending April 25, 2020, the three weeks with the highest weekly
deaths in season 2019-20, the states of California, Florida and Texas kept their ranking relative to one another.
However, New York City and the State of New Jersey became 2nd and 4th, respectively, in the ranking of U.S.
jurisdictions for average weekly deaths, while they ranked 19th and 11th, respectively, prior to March 24th 2020.
What happened?
New York City’s average weekly deaths increased from 1,043, pre-March 24 2020, to an average of 5,936 dead
per week, from week ending April 11 to April 25, 2020. That is a 469% increase! How is that even possible?
Does this imply that deaths were shifted across weeks (as in, accelerated in time) for New York City? Those are
questions New York City and families of the deceased will have to grapple with.
The five jurisdictions that experienced a percentage change in average weekly deaths greater than 100%, from
the pre-March 24 2020 period to the three-week peak deaths period from week ending 4/11/20 to 4/25/20, are:
New York City: 469%
New Jersey: 221%
New York: 126%
Connecticut: 122%
Massachusetts:110%
Page 50 of 112
Graph 70: All Cause Deaths per Week, U.S. versus U.S. without New York City and the State of New Jersey, MMWR week ending 1/11/2014 to 11/14/2020
Page 51 of 112
Graph 71: All Cause Weekly Deaths per Season, U.S. versus U.S. without NY City and the State of New Jersey, 2014-20
U.S. All Cause deaths
U.S. without New York City and the State of New Jersey
All Cause deaths
Page 52 of 112
Graph 70 displays All Cause weekly deaths, from MMWR week ending 1/11/2014 to MMWR week ending
11/14/2020, a total of 358 weeks, for the U.S. as a whole (like Graph 4 did), and for all U.S. jurisdictions,
except New York City and the State of New Jersey.
Graph 71 displays All Cause weekly deaths per season, from 2014-15 to 2019-20, for the U.S. as a whole (like
Graph 11 did), and for all U.S. jurisdictions, except New York City and the State of New Jersey.
Graphs 70 and 71 reveals that New York City and the State of New Jersey’s numbers sorely skewed the picture
of the U.S. COVID-19 situation. Did this, in turn, drive the states and counties health officers’ decisions, all
across the United States? Or did health officers diligently analyze the situation at their own state or county
level? Those are questions that residents across states and counties would surely want to be answered.
Page 53 of 112
7 Conclusion
The CDC data on U.S. deaths used in this analysis are the best available data at one point in time. Each and
every recorded death is given one and only one underlying cause of death. Old age is not an underlying cause of
death. U.S. total deaths’ increases over the years have followed U.S. population’s increases. U.S. total deaths
for calendar year 2020 or season 2019-20 are normal death numbers, as shown by death rate statistics. Death
rate statistics account for the increase in population over the years—standard “excess deaths estimates” do not.
The U.S. age distribution of deaths did not significantly change following COVID-19. This suggests the reason
a higher number of COVID-19 deaths has been reported among older individuals than younger individuals, is
simply because, every day in the U.S., COVID-19 or not, those aged 85 or older are at a higher risk of dying
than those aged 84 or younger. This should come as no surprise to anyone.
Deaths have followed a pattern of highs and lows, year after year, and the magnitude of those highs and lows,
one year relatively to another, has been consistent for each and every main cause of death. In other words, the
peaks of deaths, from all those different causes, have coincide (happened over the same weeks)year after
year, and the magnitude of those peaks for one year relative to another, for each cause, has also been consistent
across causes. But not for 2020.
Respiratory diseases had previously been classified into one the following three cause of death categories:
Chronic lower respiratory diseases, influenza and pneumonia, and other diseases of the respiratory system. The
historically low levels of deaths due to these old respiratory diseases points to a reclassification of deaths into
the newly and specially introduced respiratory disease category for COVID-19. Level of deaths departing from
their historical levels, such as deaths due to non-natural causes (suicides) and natural cause (diabetes) points to
mitigation efforts (isolation, lockdowns) being the cause of death, rather than COVID-19 itself.
March 24th 2020 is the date the new International Classification of Diseases (ICD) code issued by the World
Trade Organization (WHO) for COVID-19 was introduced and implemented in the United States. Up until then,
the CDC’s deaths data provided a treasure trove of information, with easily recognizable seasonal patterns and
the U.S. weekly deaths were following those historical patterns, exactly. On week ending March 28th 2020, the
U.S. weekly deaths sharply departed from them. Are the unfamiliar patterns seen since due to COVID-19 or the
reoccurring fear campaigns?
On March 13, 2020 the President of the United States declared a national emergency concerning COVID-19
that would legally allow the distribution of funds. The unprecedented level of efforts, pressure, and means
devoted to COVID-19 since, rule out any concerns of underreporting of COVID-19 deaths. What role have
those different incentives play in the potential overreporting of COVID-19 deaths?
Page 54 of 112
The picture of the U.S. COVID-19 situation was heavily skewed by deaths numbers from jurisdictions such as
New York City and the State of New Jersey. At the state level, many have shown no change in deaths numbers
and pattern of deaths in 2019-20, compared to prior seasons.
Has any consideration ever been given to the relevance of the national picture and national “solutions, to state,
county, local and individual situations? Or have COVID-19 measures been blindly adopted and applied, top-
down, in haste, obsessively and short sightedly? What role did groupthink play into this? Those questions will
need to be answered.
A more holistic approach to health seems to be needed to address each individual’s unique circumstances, needs
and preferences. Only such an approach can minimize all the suffering and inequities at the individual level, and
maximize welfare at the society level. This individual centered approach to health needs not to harm one, for
others to benefit. It fosters unity, by being compassionate to all.
Page 55 of 112
8 Testimonials
Page 56 of 112
9 References
(1) Briand, G. (2020) “COVID-19 Deaths: A Look at U.S. Data,” JHU Advanced Academic Programs, AAP
Events webinars, Nov. 11th 2020. Available at: https://www.youtube.com/watch?v=3TKJN61aflI
(2) Gu, Y. (2020) “A Closer Look at U.S. Deaths due to COVID-19,” The Johns Hopkins News-Letter, Nov. 27,
2020. Available at: https://drive.google.com/file/d/1Tnb1a8TXHj_jJCM2BDfGSriUgdn-2gec/view
(3) Parayil, T. and B. Murray (2020) “Public Health experts and biostatisticians weigh in on “COVID-19
Deaths: A Look at U.S. Data” webinar”, The Johns Hopkins News-Letter, Dec. 14, 2020. Available at:
https://www.jhunewsletter.com/article/2020/12/public-health-experts-and-biostatisticians-weigh-in-on-
covid-19-deaths-a-look-at-u-s-data-webinar
(4) Rossen, L. M. (2021) National Center for Health Statistics (NCHS), Centers for Disease Control and
Prevention (CDC). February 2021. Email communications.
(5) Rossen, L. M., A. M. Branum, F. B. Ahmad, P. Sutton, and R. N. Anderson (2020) “Excess Deaths
Associated with COVID-19, by Age and Race and EthnicityUnited States, January 26-October 3, 2020”
Morbidity and Mortality Weekly Report, 69(42), Oct. 23rd 2020. Available at:
https://www.cdc.gov/mmwr/volumes/69/wr/pdfs/mm6942e2-H.pdf (.pdf file)
https://www.cdc.gov/mmwr/volumes/69/wr/mm6942e2.htm (webpage)
(6) U.S. Census Bureau. U.S Population Clock. Consulted February 1st 2021. https://www.census.gov/popclock/
(7) U.S. Centers for Disease Control and Prevention (CDC), National Center for Health Statistics (NCHS),
COVID-19 Data from NCHS: COVID-19 Death Data and Resources, “Excess Deaths Associated with
COVID-19” webpage, https://www.cdc.gov/nchs/nvss/vsrr/covid19/excess_deaths.htm
(8) U.S. Centers for Disease Control and Prevention (CDC), National Center for Health Statistics (NCHS),
COVID-19 Data from NCHS: COVID-19 Death Data and Resources, Weekly Updates by Select
Demographic and Geographic Characteristics, data from page 2 of 2 of the “Provisional Death Counts for
Coronavirus Disease 2019 (COVID-19) By Week of Death” table. To get a deaths number for calendar year
2020, the 2020 Total Deaths data were downloaded again, on 01-28-21, and exclusively used for Tables 4,
7. Available at: https://www.cdc.gov/nchs/nvss/vsrr/covid_weekly/index.htm
(9) U.S. Centers for Disease Control and Prevention (CDC), National Center for Health Statistics (NCHS),
COVID-19 Data from NCHS: COVID-19 Death Data and Resources, Weekly Updates by Select
Demographic and Geographic Characteristics, data from page 2 of 2 of the “Provisional Death Counts for
Coronavirus Disease 2019 (COVID-19) By Week of Death” table. Downloaded 12-26-20. Available at:
https://www.cdc.gov/nchs/nvss/vsrr/covid_weekly/index.htm
(10) U.S. Centers for Disease Control and Prevention (CDC) Data Catalog, National Center for Health Statistics
(NCHS), “Weekly Counts of Deaths by State and Select Causes, 2014-2018” dataset. Downloaded 12-26-
20, with CDC last updated date of Nov. 10th 2020. Available at: https://data.cdc.gov/NCHS/Weekly-Counts-
of-Deaths-by-State-and-Select-Causes/3yf8-kanr
Page 57 of 112
(11) U.S. Centers for Disease Control and Prevention (CDC) Data Catalog, National Center for Health
Statistics (NCHS), “Weekly Counts of Deaths by State and Select Causes, 2019-2020” dataset.
Downloaded 12-26-20, with CDC last updated date of Dec. 23rd 2020. Available at:
https://data.cdc.gov/NCHS/Weekly-Counts-of-Deaths-by-State-and-Select-Causes/muzy-jte6
(12) U.S. Centers for Disease Control and Prevention (CDC), Diseases & Conditions: Flu (Influenza), Flu
Activity & Surveillance, Overview of Influenza Surveillance in the United States, “U.S. Influenza
Surveillance System: Purpose and Methods” webpage, “Influenza Surveillance Considerations” bottom
section of the webpage. Consulted 02-02-21, with CDC last updated date of Oct. 6th 2020.
https://www.cdc.gov/flu/weekly/overview.htm#Mortality
(13) U.S. Centers for Disease Control and Prevention (CDC), National Center for Health Statistics (NCHS),
Classification of Diseases, Functioning and Disability, International Classification of Diseases Tenth
Revision (ICD-10). Consulted 02-20-21. https://www.cdc.gov/nchs/icd/icd10.htm
(14) U.S. Centers for Disease Control and Prevention (CDC), National Center for Health Statistics (NCHS),
Classification of Diseases, Functioning and Disability, International Classification of Diseases Ninth
Revision (ICD-9) webpage. Consulted 02-20-21. https://www.cdc.gov/nchs/icd/icd9.htm
(15) U.S. Centers for Disease Control and Prevention (CDC) Data Catalog, National Center for Health
Statistics (NCHS), National Vital Statistics System (NVSS), COVID-19 Death Data and Resources
webpage https://www.cdc.gov/nchs/nvss/covid-19.htm, NVSS COVID-19 Alerts section of the
webpage, Notification of new ICD code introduced for COVID-19. Released 3/24/2020. Available at:
https://www.cdc.gov/nchs/data/nvss/coronavirus/Alert-2-New-ICD-code-introduced-for-COVID-19-
deaths.pdf
(16) U.S. Centers for Disease Control and Prevention (CDC), National Center for Health Statistics (NCHS),
National Vital Statistics System (NVSS), Deaths/Mortality Data, Results and Publications, Mortality
Tables webpage, Leading Causes of Death-Combined Race and Hispanic Origin section of the
webpage, LCWK5_HR tables: Deaths, percent of total deaths for the 15 leading causes of death, United
States and each State, 2015-2017. https://www.cdc.gov/nchs/nvss/mortality/lcwk5_hr.htm
(17) U.S. Centers for Disease Control and Prevention (CDC), National Center for Health Statistics (NCHS),
National Vital Statistics System (NVSS), Deaths/Mortality Data, Results and Publications, Mortality
Tables webpage, Leading Causes of Death section of the webpage, Leading Causes of Death section of
the webpage, LCWK9 tables: Deaths, percent of total deaths, and death rates for the 15 leading causes
of death, United States and each State, 1999-2015. https://www.cdc.gov/nchs/nvss/mortality/lcwk9.htm
(18) U.S. Centers for Disease Control and Prevention (CDC) WONDER Underlying Cause of Death, All
Ages, Deaths online databases, 1999-2019 Underlying Cause of Death by Bridged-Race Categories,
table layout’s group results organized by year. Tables generated February 1st 2021.
https://wonder.cdc.gov/ucd-icd10.html
(19) Varian, Hal R. (1997) “How to Build an Economic Model in Your Spare Time,” The American
Economist, 41(2), Fall 1997, pp. 3-10. Available at:
https://people.ischool.berkeley.edu/~hal/Papers/how.pdf
Page 58 of 112
(20) World Health Organization (WHO), Classifications, Classifications of Diseases (ICD) webpage.
Consulted 02-20-21. https://www.who.int/classifications/classification-of-diseases
(21) World Health Organization (WHO), ICD-10 2019 online version. https://icd.who.int/browse10/2019/en
Page 59 of 112
10 Appendix
Page 60 of 112
Graph 17: AL (Alabama) All Cause Weekly Deaths, from season 2014-15 to season 2019-20
Page 61 of 112
Graph 18: AK (Alaska) All Cause Weekly Deaths, from season 2014-15 to season 2019-20
Page 62 of 112
Graph 19: AZ (Arizona) All Cause Weekly Deaths, from season 2014-15 to season 2019-20
Page 63 of 112
Graph 20: AR (Arkansas) All Cause Weekly Deaths, from season 2014-15 to season 2019-2
Page 64 of 112
Graph 21: CA (California) All Cause Weekly Deaths, from season 2014-15 to season 2019-20
Page 65 of 112
Graph 22: CO (Colorado) All Cause Weekly Deaths, from season 2014-15 to season 2019-20
Page 66 of 112
Graph 23: CT (Connecticut) All Cause Weekly Deaths, from season 2014-15 to season 2019-20
Page 67 of 112
Graph 24: DE (Delaware) All Cause Weekly Deaths, from season 2014-15 to season 2019-20
Page 68 of 112
Graph 25: DC (District of Columbia) All Cause Weekly Deaths, from season 2014-15 to season 2019-20
Page 69 of 112
Graph 26: FL (Florida) All Cause Weekly Deaths, from season 2014-15 to season 2019-20
Page 70 of 112
Graph 27: GA (Georgia) All Cause Weekly Deaths, from season 2014-15 to season 2019-20
Page 71 of 112
Graph 28: HI (Hawaii) All Cause Weekly Deaths, from season 2014-15 to season 2019-20
Page 72 of 112
Graph 29: ID (Idaho) All Cause Weekly Deaths, from season 2014-15 to season 2019-20
Page 73 of 112
Graph 30: IL (Illinois) All Cause Weekly Deaths, from season 2014-15 to season 2019-20
Page 74 of 112
Graph 31: IN (Indiana) All Cause Weekly Deaths, from season 2014-15 to season 2019-20
Page 75 of 112
Graph 32: IA (Iowa) All Cause Weekly Deaths, from season 2014-15 to season 2019-20
Page 76 of 112
Graph 33: KS (Kansas) All Cause Weekly Deaths, from season 2014-15 to season 2019-20
Page 77 of 112
Graph 34: KY (Kentucky) All Cause Weekly Deaths, from season 2014-15 to season 2019-20
Page 78 of 112
Graph 35: LA (Louisiana) All Cause Weekly Deaths, from season 2014-15 to season 2019-20
Page 79 of 112
Graph 36: ME (Maine) All Cause Weekly Deaths, from season 2014-15 to season 2019-20
Page 80 of 112
Graph 37: MD (Maryland) All Cause Weekly Deaths, from season 2014-15 to season 2019-20
Page 81 of 112
Graph 38: MA (Massachusetts) All Cause Weekly Deaths, from season 2014-15 to season 2019-20
Page 82 of 112
Graph 39: MI (Michigan) All Cause Weekly Deaths, from season 2014-15 to season 2019-20
Page 83 of 112
Graph 40: MN (Minnesota) All Cause Weekly Deaths, from season 2014-15 to season 2019-20
Page 84 of 112
Graph 41: MS (Mississippi) All Cause Weekly Deaths, from season 2014-15 to season 2019-20
Page 85 of 112
Graph 42: MO (Missouri) All Cause Weekly Deaths, from season 2014-15 to season 2019-20
Page 86 of 112
Graph 43: MT (Montana) All Cause Weekly Deaths, from season 2014-15 to season 2019-20
Page 87 of 112
Graph 44: NE (Nebraska) All Cause Weekly Deaths, from season 2014-15 to season 2019-20
Page 88 of 112
Graph 45: NV (Nevada) All Cause Weekly Deaths, from season 2014-15 to season 2019-20
Page 89 of 112
Graph 46: NH (New Hampshire) All Cause Weekly Deaths, from season 2014-15 to season 2019-20
Page 90 of 112
Graph 47: NJ (New Jersey) All Cause Weekly Deaths, from season 2014-15 to season 2019-20
Page 91 of 112
Graph 48: NM (New Mexico) All Cause Weekly Deaths, from season 2014-15 to season 2019-20
Page 92 of 112
Graph 49: NY (New York) All Cause Weekly Deaths, from season 2014-15 to season 2019-20
Page 93 of 112
Graph 50: NYC (New York City) All Cause Weekly Deaths, from season 2014-15 to season 2019-20
Page 94 of 112
Graph 51: NC (North Carolina) All Cause Weekly Deaths, from season 2014-15 to season 2019-20
Page 95 of 112
Graph 52: ND (North Dakota) All Cause Weekly Deaths, from season 2014-15 to season 2019-20
Page 96 of 112
Graph 53: OH (Ohio) All Cause Weekly Deaths, from season 2014-15 to season 2019-20
Page 97 of 112
Graph 54: OK (Oklahoma) All Cause Weekly Deaths, from season 2014-15 to season 2019-20
Page 98 of 112
Graph 55: OR (Oregon) All Cause Weekly Deaths, from season 2014-15 to season 2019-20
Page 99 of 112
Graph 56: PA (Pennsylvania) All Cause Weekly Deaths, from season 2014-15 to season 2019-20
Page 100 of 112
Graph 57: PR (Puerto Rico) All Cause Weekly Deaths, from season 2014-15 to season 2019-20
Page 101 of 112
Graph 58: RI (Rhode Island) All Cause Weekly Deaths, from season 2014-15 to season 2019-20
Page 102 of 112
Graph 59: SC (South Carolina) All Cause Weekly Deaths, from season 2014-15 to season 2019-20
Page 103 of 112
Graph 60: SD (South Dakota) All Cause Weekly Deaths, from season 2014-15 to season 2019-20
Page 104 of 112
Graph 61: TN (Tennessee) All Cause Weekly Deaths, from season 2014-15 to season 2019-20
Page 105 of 112
Graph 62: TX (Texas) All Cause Weekly Deaths, from season 2014-15 to season 2019-20
Page 106 of 112
Graph 63: UT (Utah) All Cause Weekly Deaths, from season 2014-15 to season 2019-20
Page 107 of 112
Graph 64: VT (Vermont) All Cause Weekly Deaths, from season 2014-15 to season 2019-20
Page 108 of 112
Graph 65: VA (Virginia) All Cause Weekly Deaths, from season 2014-15 to season 2019-20
Page 109 of 112
Graph 66: WA (Washington) All Cause Weekly Deaths, from season 2014-15 to season 2019-20
Page 110 of 112
Graph 67: WV (West Virginia) All Cause Weekly Deaths, from season 2014-15 to season 2019-20
Page 111 of 112
Graph 68: WI (Wisconsin) All Cause Weekly Deaths, from season 2014-15 to season 2019-20
Page 112 of 112
Graph 69 WY (Wyoming) All Cause Weekly Deaths, from season 2014-15 to season 2019-20
Preprint
Full-text available
This paper presents plots of U.S. monthly deaths per age group for each of the six main causes of death. Publicly available CDC data from 1999 to 2021 are used. The plots are presented at the request of readers and are meant to complement Briand's February 2022 working paper. These plots provide further evidence of reclassification of deaths across categories and evidence consistent with the vaccine deaths hypothesis.
Article
What is already known about this topic? As of October 15, 216,025 deaths from COVID-19 have been reported in the United States; however, this might underestimate the total impact of the pandemic on mortality. What is added by this report? Overall, an estimated 299,028 excess deaths occurred from late January through October 3, 2020, with 198,081 (66%) excess deaths attributed to COVID-19. The largest percentage increases were seen among adults aged 25-44 years and among Hispanic or Latino persons. What are the implications for public health practice? These results inform efforts to prevent mortality directly or indirectly associated with the COVID-19 pandemic, such as efforts to minimize disruptions to health care. © 2020 Department of Health and Human Services. All rights reserved.
Article
Editor’s Introduction Originally published in Volume 41, Number 2, Fall 1997, pages 3-10. Hal Varian (born 1947) is widely known by professional economists for his pathbreaking work in the economics of information and networks. Many more know him as the author of two bestselling microeconomics textbooks, one written for undergraduate college students and one designed for advanced graduate students. Through his research and his books, Professor Varian’s ideas have influenced a generation of economists. In this paper, Professor Varian outlines how he approaches the task of building an economic model to explain an observed phenomena or solve a problem. His words are encouraging advice for graduate students and young economists learning how to “practice the art” of economics. Professor Varian offers a number of tips ranging from how to choose a topic, when to read the literature, and even to how to effectively manage your bibliographic citations. Professor Varian’s advice has passed the market test as this paper remains one of the most referenced and downloaded papers in The American Economist’s backfile. However, after including the paper on a course reading list several years ago, one doctoral student pointed out to this editor that Professor Varian fails to explain how to find the “spare time” that he references in the title!
COVID-19 Deaths: A Look at U.S. Data
  • G Briand
Briand, G. (2020) "COVID-19 Deaths: A Look at U.S. Data," JHU Advanced Academic Programs, AAP Events webinars, Nov. 11 th 2020. Available at: https://www.youtube.com/watch?v=3TKJN61aflI
A Closer Look at U.S. Deaths due to COVID-19
  • Y Gu
Gu, Y. (2020) "A Closer Look at U.S. Deaths due to COVID-19," The Johns Hopkins News-Letter, Nov. 27, 2020. Available at: https://drive.google.com/file/d/1Tnb1a8TXHj_jJCM2BDfGSriUgdn-2gec/view
Public Health experts and biostatisticians weigh in on
  • T Parayil
  • B Murray
Parayil, T. and B. Murray (2020) "Public Health experts and biostatisticians weigh in on "COVID-19 Deaths: A Look at U.S. Data" webinar", The Johns Hopkins News-Letter, Dec. 14, 2020. Available at: https://www.jhunewsletter.com/article/2020/12/public-health-experts-and-biostatisticians-weigh-in-oncovid-19-deaths-a-look-at-u-s-data-webinar
Weekly Counts of Deaths by State and Select Causes
U.S. Centers for Disease Control and Prevention (CDC) Data Catalog, National Center for Health Statistics (NCHS), "Weekly Counts of Deaths by State and Select Causes, 2014-2018" dataset. Downloaded 12-26-20, with CDC last updated date of Nov. 10 th 2020. Available at: https://data.cdc.gov/NCHS/Weekly-Countsof-Deaths-by-State-and-Select-Causes/3yf8-kanr