See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/343775235
Evaluation of the virulence of SARS-CoV-2 in France, from all-cause mortality
Technical Report · August 2020
3 authors, including:
Some of the authors of this publication are also working on these related projects:
Science reviews relevant to COVID-19 View project
Ab Initio Mossbauer Parameter Calculations (MSc & PhD) View project
D. G. Rancourt
Ontario Civil Liberties Association
145 PUBLICATIONS4,293 CITATIONS
All content following this page was uploaded by D. G. Rancourt on 20 August 2020.
The user has requested enhancement of the downloaded file.
Evaluation of the virulence of SARS-CoV-2 in France,
from all-cause mortality 1946-2020
Denis G. Rancourt1,*, Marine Baudin2, Jérémie Mercier2
1 Ontario Civil Liberties Association (ocla.ca) ; 2 Mercier Production (jeremie-mercier.com) ;
Published at ResearchGate
20 August 2020
We analyzed historic and recent all-cause mortality data for France, and other
jurisdictions for comparison, using model fitting to quantify winter-burden deaths, and
deaths from exceptional events. In this way, COVID-19 is put in historic perspective. We
prove that the “COVID-peak” feature that is present in the all-cause mortality data of
certain mid-latitude Northern hemisphere jurisdictions, including France, cannot be a
natural epidemiological event occurring in the absence of a large non-pathogenic
perturbation. We are certain that this “COVID-peak” is artificial because it:
i. occurs sharply (one-month width) at an unprecedented location in the
seasonal cycle of all-cause mortality (centered at the end of March),
ii. is absent in many jurisdictions (34 of the USA States have no “COVID-peak”),
iii. varies widely in magnitude from jurisdiction to jurisdiction in which it occurs.
We suggest that:
• the unprecedented strict mass quarantine and isolation of both sick and healthy
elderly people, together and separately, killed many of them,
• that this quarantine and isolation is the cause of the “COVID-peak” event that we
• and that the medical mechanism is mainly via psychological stress and social
isolation of individuals with health vulnerabilities.
According to our calculations, this caused some 30.2 K deaths in France in March and
April 2020. However, even including the “COVID-peak”, the 2019-2020 winter-burden
all-cause mortality is not statistically larger than usual. Therefore SARS-CoV-2 is not an
unusually virulent viral respiratory disease pathogen. By analyzing the all-cause
mortality data from 1946 to 2020, we also identified a large and steady increase in all-
cause mortality that began in approximately 2008, which is too large to be explained by
population growth in the relevant age structure, and which may be related to the
economic crash of 2008 and its long-term societal consequences.
Résumé en français
Nous avons analysé les données historiques et récentes de mortalité toutes causes
confondues pour la France et d'autres juridictions à des fins de comparaison, en lissant
une courbe théorique pour quantifier les décès dus à la charge hivernale et les décès
dus à des événements exceptionnels. De cette façon, on peut observer le COVID-19
avec une perspective historique. Ainsi, nous prouvons que le « pic COVID » présent
dans les données de mortalité toutes causes confondues de certaines juridictions de
l'hémisphère Nord à moyenne latitude, y compris la France, ne peut pas être un
événement épidémiologique naturel ayant survenu de façon naturelle, en l'absence
d'une grande perturbation non pathogène. Nous sommes convaincus que le « pic
COVID » est artificiel car :
i. il s’est produit brusquement (largeur d'un mois) à une date sans précédent
dans le cycle saisonnier de mortalité toutes causes confondues (milieu du pic
à la fin mars),
ii. il est absent dans de nombreuses juridictions (34 des États américains n'ont
pas de « pic COVID »), et
iii. l’ampleur de ce pic varie considérablement d’une juridiction à l’autre.
Nous suggérons que :
• la quarantaine de masse et l'isolement strict sans précédent des personnes
âgées malades et en bonne santé, ensemble et séparément, a tué beaucoup
• que cette quarantaine et cet isolement sont la cause de l'événement « pic-
COVID » que nous avons quantifié,
• et que le mécanisme médical expliquant ce pic passe principalement par le
stress psychologique et l'isolement social des personnes vulnérables au niveau
de leur santé.
Selon nos calculs, ces mesures ont provoqué quelques 30,2 K décès en France en
mars et avril 2020. Cependant, même en incluant le « pic COVID », la charge hivernale
de mortalité toutes causes confondues pour l’hiver 2019-2020 n'est pas statistiquement
supérieure aux charges hivernales habituelles, ce qui nous amène à affirmer que le
SARS-CoV-2 n'est pas un virus responsable de maladies respiratoires inhabituellement
En analysant les données de mortalité toutes causes confondues de 1946 à 2020, nous
avons également identifié une augmentation importante et régulière de la mortalité
toutes causes confondues qui a commencé vers 2008, trop importante pour être
expliquée par la croissance de la population étant donné la pyramide des âges, mais
qui pourrait être liée à la crise économique de 2008 et à ses conséquences sociétales
sur le long terme.
France is said to be one of the five European countries most impacted by COVID-19,
with Belgium, UK, Italy and Spain.
France has applied broad response measures since the pandemic was declared by the
WHO on 11 March 2020, including national lockdown and systematic quarantine of sick
and healthy individuals together in care homes and facilities for elderly persons.
The question arises: Is there bias-free hard evidence that the extraordinary measures
were and are warranted? After all, if the pathogen is as contagious and virulent as
believed, then, irrespective of the array of efforts to mitigate spread of the epidemic, it
should be evident by now that the decisions to impose the measures were warranted.
Alternatively, if there is little evidence of an abnormal increase in mortality, then either
SARS-CoV-2 is not as dangerous as imagined, or the array of ad hoc mitigation
measures has been effective and should be considered proven.
2. Data and methods
2.1. Data selection
Cause-of-death assignation and COVID-19 mass “testing” are both susceptible to bias
(Cummins, 2020). All-cause mortality is not. Therefore, we use the extensive database
of all-cause mortality by month for metropolitan France 1946-2020, and other data (see
section 2.2), to cast recent deaths in their historical context. Here, “metropolitan France”
means continental France and Corsica (i.e. European France).
2.2. Data retrieval
Table 1 describes the data retrieved and which source it has been collected from.
France 1946-2020 Year Insee (2020c)
mortality France 1982-2019 Year Insee (2020a)
France 1946-2020 Month Insee (2020d)
mortality France 1994-2020 Month Insee (2020e)
1 March to 20
July for 2018,
2019 and 2020
Day Insee (2020b)
France 1968-2018 Day Insee (2019)
mortality Canada 2014-2020 Week StatCan (2020)
mortality USA 2013-2020 Week CDC (2020)
Table 1. Data retrieved. Metropolitan France means continental France and Corsica. France
means metropolitan France and overseas France.
2.3. Epidemiological data analysis
We chose not to analyse the data by the common method of using a sinusoidal signal
intended to separate viral respiratory disease deaths from other seasonally varying
deaths. We believe the latter method, although widely applied, is problematic for the
following main reasons:
• The assumed underlying sinusoidal component does not reliably separate deaths
assigned as being primarily caused by the viral respiratory disease of interest
and the deaths assigned as being primarily due to other seasonally varying (non-
• The sinusoidal model does not correctly fit the non-viral seasonal component of
all-cause deaths, since it has systematic residuals in those segments assumed to
be unaffected by the viral pathogen.
• There is no biological or medical reason that any seasonal component will have a
simple sinusoidal functional form, and many reasons that it would not.
Instead, we analyse the all-cause mortality by month data using a sum of one to three
Voigt lines for each peak or feature that rises above the assumed-linear summer
baseline for the fitting region. In practice, we select a fitting region over which the
summer baseline delimited by the bottoms of the summer troughs is approximately a
straight line with a given slope, and use the Voigt lines to fit the peaks that rise above
this summer baseline for the fitting region. In this way, the total area of all the Voigt lines
in a given winter peak, for example, is the winter-burden mortality for the given winter.
Figure 1 shows that whereas the winter-peak values vary somewhat erratically from
year to year, the summer-trough bottoms delineate linear trends with time (the “summer
baselines”), in distinct time periods. We delineated the data into five regions as:
2005-2020: linear with positive slope “region-I”
1994-2005: linear with near-zero slope “region-II”
1968-1994: linear with near-zero slope “region-III”
1958-1968: linear with positive slope “region-IV”
1946-1958: linear with near-zero slope “region-V”
Here, regions II and III both have essentially the same linear summer baselines
(Figure 1) but were divided into two regions to reduce the sizes of the fittings, and for
easier comparison with the 1994-2020 France data (Figure 2).
Within each such region (I through V), we fit the data with a linear summer baseline and
model peaks for each of the winters. The model peak for a given winter (or a given
anomalous peak, see section 3) was taken to be the sum of a variable number, Npeak, of
Voigt lines. The Voigt lineshape is a convolution between the Lorentzian lineshape and
the Gaussian lineshape, such that it can be varied to adopt any shape on a “Lorentzian-
Gaussian continuum” of shapes. This is convenient because, for a given lineshape-
area, the Lorentzian has broad wings (and a pointed head), whereas the Gaussian
shape has a crisp delineation with little wings (and a broad head). The Voigt lineshape
is symmetric about its center, whereas all-cause mortality peaks are not generally
symmetric, and contain structure such as shoulders, sharp rises, and asymmetric or
unequal decays on the two sides. We accommodate such structure by using as many
(Npeak) Voigt lines in a given all-cause mortality peak as are minimally needed to reduce
the residual (i.e. the difference between the data and the model function) to random
noise. With the France 1946-2020 data, this requires between 1 and 3 Voigt lines per
peak (Npeak = 1 to 3), excluding the anomalous peaks that each require their own Voigt
line (one per anomaly, in this case).
Using this method, the winter-burden peaks are well represented and contribute little to
raising the summer-trough bottoms above the linear summer baseline. Thus our model
reliably captures the winter-burden deaths that occur above the summer baseline. In
other words, the winter-burden deaths of a season correspond to the area under the
winter-burden peak for that season.
The yearly all-cause mortality is calculated for two types of years: the cycle-year and the
Cycle-year: For a given winter-centered year (cycle-year), the all-cause mortality is
equal to the summer baseline value of mortality per month evaluated at the weighted
peak position (close to 1 January) times 12 plus the areas of all the Npeak Voigt lines in
the winter peak.
Calendar-year: The all-cause mortality is obtained by direct counting for the 12 months
in each calendar year.
Fitting and quantification are done with the Recoil spectral analysis software, adapted
as needed for the epidemiological context (Lagarec and Rancourt, 1998; Rancourt,
3. Analysis and discussion
3.1. France 1946-2020 data
France maintains a high-quality demographic database, from 1946 to present (Insee,
2020d). Figure 1 shows all-cause mortality by month for metropolitan France, from
January 1946 to June 2020:
Figure 1. All-cause mortality by month in metropolitan France from 1946 to 2020. Data are
displayed from January 1946 to June 2020. Data were retrieved from Insee (Insee, 2020d), as
described in Table 1.
The data shows the well-known and prominent winter peaks and summer troughs
(Dowell, 2001; Marti-Soler et al., 2014; Paules and Subbarao, 2017; Rancourt, 2020).
Such seasonal patterns of all-cause mortality occur in all mid-latitude countries. The
patterns are shifted by 6 months in the Southern-hemisphere mid-latitudes, where the
peaks again correspond to winters in that hemisphere.
Visual inspection of Figure 1 shows that the 2019-2020 winter mortality in France was
not obviously anomalous, at first sight. This is not surprising to us: most provinces in
Canada and most states in the USA have 2019-2020 winter-burden all-cause mortalities
that are smaller than for each of at least two other winters in the last decade
Figure 1 is a sobering result, which is in contrast to the focus of media coverage since
March 2020. There was not an extraordinary winter mortality in France in 2019-2020. In
light of 75 years of all-cause mortality data, death has continued its seasonal variation
without any remarkable event, remaining within the bounds of year-to-year statistical
variation, at least on the large scale of this figure.
In France, there have been five seasons over the last 75 years with a higher maximum
in all-cause mortality by month than the maximum of the 2019-2020 season: 1945-1946,
1948-1949, 1952-1953, 1969-1970 and 2016-2017 (Figure 1). The 2019-2020 seasonal
epidemic was not the worst in a century, as claimed by French president Emmanuel
Macron (see France 24, 2020, at 00:34).
3.2. France 1994-2020 data
France has also released “all-France” mortality data, which includes metropolitan and
overseas France, for the last nearly three decades (Insee, 2020e). Figure 2 shows all-
cause mortality by month for the whole of France, from January 1994 to June 2020:
Figure 2. All-cause mortality by month in France from 1994 to 2020. Data are displayed for
“all-France”, which includes metropolitan and overseas France, from January 1994 to June
2020. The arrows show the two anomalous peaks discussed in the text. Data were retrieved
from Insee (Insee, 2020e), as described in Table 1.
At this resolution (1994-2020, by month), two anomalies are recognized, which do not
conform to known seasonal-variation patterns for mid-latitude countries in the Northern
hemisphere: the August-2003 heat wave anomaly and the March-April-2020 anomaly,
which we name the “COVID-peak” (following Rancourt (2020)) and describe in the next
3.3. France August-2003 heat wave anomaly
The first anomaly is a single-month spike that occurred in August 2003 (“2003-08”),
which would normally be part of a trough in all-cause mortality by month, which rises
near the 58 K deaths/month mark in 2003 (Figure 2). This anomaly has conclusively
been attributed to an exceptional heat wave that hit nearly all of France in that month
and that killed approximately 15 K people (Evin et al., 2004; Hémon and Jougla, 2004).
It is an example of deaths that cannot be attributed to a pathogen acting on a population
in normal circumstances.
3.4. “COVID-peak” anomaly
The second anomaly is a narrow peak, having a width of approximately 1 month,
occurring at (centered on) the end of March 2020, which would normally be the
decaying shoulder of the recent winter peak. Winter peaks are always centered at the
beginning of January and by March are always in decay towards the next summer
trough in all-cause mortality. Rancourt has called the second anomaly the “COVID-
peak” and he has postulated that it was caused by the government responses that
followed the 11 March 2020 WHO declaration of the pandemic (Rancourt, 2020).
The all-cause mortality by day (Figure 3) shows that the said “COVID-peak” occurs on
the March-side decay of the preceding winter peaks. Figure 3 shows the all-cause
mortality by day for France, for the years 2018, 2019 and 2020, from 1 March through
Figure 3. All-cause mortality by day in France from March to June 2018, 2019 and 2020.
Data are displayed for “all-France”, which includes metropolitan and overseas France, from 1
March to 30 June of 2018, 2019 and 2020. The black line is the data for 2018. The grey line is
the data for 2019. The green dashed line is the data for 2020. Data were retrieved from Insee
(Insee, 2020b), as described in Table 1.
There has never previously been a sharp (1 month width) prominent peak in all-cause
mortality, occurring at the end of March, such as this “COVID-peak”, in the 75 years of
all-cause mortality records for France, nor for available records for Canada and its
provinces, the USA and its states, England and Wales, and European countries
(Rancourt, 2020 and to be published).
In addition, the “COVID-peak” anomaly not only occurs at a unique time in the
epidemiological cycle but also varies widely in magnitude, from zero (e.g. California) to
overwhelmingly large (e.g. New York State), in going from one mid-latitude Northern-
hemisphere jurisdiction to another (manuscript in preparation). This is illustrated for
Canada, as follows. Figure 4 shows all-cause mortality by week (number of deaths per
week vs standard CDC weeks) from week-1 (first week of January) of 2014 to week-22
(last week of May) of 2020, for the provinces of Ontario and Quebec:
Figure 4. All-cause mortality by week in Ontario and Quebec, from 2014 week-1 to 2020
week-22. The grey line shows the data for Ontario. The black line shows the data for Quebec.
Data were retrieved from Statistics Canada (StatCan, 2020), as described in Table 1.
Ontario and Quebec are similarly populous East-West adjacent provinces of similar
sizes, having distinct medical systems (health is a provincial jurisdiction in the Canadian
constitution). As with virtually all mid-latitude Northern-hemisphere countries, the
epidemiological cycles (all-cause mortality curves) of Ontario and Quebec are virtually
identical, except for the “COVID-peak” anomaly. The “COVID-peak” is much larger in
Quebec than in Ontario, where Quebec was the first province to impose an aggressive
lockdown and close its provincial borders.
For decades the epidemiological cycles (all-cause mortality curves) in all mid-latitude
Northern-hemisphere jurisdictions have been virtually identical, and have never
displayed any peak centered at the end of March, until after 11 March 2020 when a
“COVID-peak” anomaly occurred in certain jurisdictions, which is widely variable in
magnitude. Therefore, the “COVID-peak” cannot be due to a natural progression of a
viral respiratory disease (regardless its virulence), in unperturbed societal structures.
Indeed, if this anomaly was due to virulence, it would be difficult to understand the large
time-lag between the first reported case in France (27 December 2019 according to
Deslandes et al., 2020) and the anomaly’s sudden rise starting in mid-March of the
“COVID-peak”. We postulate that the excess all-cause mortality captured by the
“COVID-peak” anomaly was caused by government responses to the declaration of the
“pandemic” by the WHO on 11 March 2020. It is not a natural epidemiological event,
irrespective of the underlying pathogenic and co-morbidity circumstances.
Indeed, the said “COVID-peak” is remarkable in epidemiological terms in that it is
entirely absent for many jurisdictions, where the absence appears to be tied more to
jurisdictional politics and policy rather than any epidemiological logic. For example, the
“COVID-peak” is entirely absent in 34 of the USA States, and varies dramatically in
intensity from state to state for those States in which it is present (manuscript in
preparation). Figure 5 shows a colour-coded map of the USA for “COVID-peak”
intensity. Darker green is increased degree of absence of the “COVID-peak”, and darker
grey is increased intensity of a discerned “COVID-peak”:
Figure 5. “COVID-peak” intensity map of the USA. States in green are the states where the
“COVID-peak” is absent. The darker the green, the more intense the absence. States in grey
are the states where the “COVID-peak” is present. The darker the grey, the more intense the
presence. Data were retrieved from CDC (CDC, 2020), as stated in Table 1.
Here, all the USA States have comparable infection rates, according to reported mass
testing results (Ioannidis, 2020). Such geographical variation in an all-cause mortality
peak that occurs simultaneously in various localities on two continents is unprecedented
in the natural history of human epidemiology.
Either SARS-CoV-2 is such a unique viral respiratory disease pathogen, unlike any
previously seen, that it can naturally cause a mortality peak at the end of March, across
the mid-latitude Northern-hemisphere world, solely in certain jurisdictions where it
occurs, or synchronous and local external (non-pathogenic) factors played a major role.
We conclude the latter.
3.5. Quantitative analysis of the all-cause mortality data
Next, we made a quantitative analysis of the all-cause mortality by month for
metropolitan France from January 1946 to June 2020 (Figure 1), as described in
Figure 6 shows our fit, and its residual, for region-II (1994-January through 2005-
Figure 6. Fit of the monthly all-cause mortality data of metropolitan France 1946-2020,
region-II. Region-II corresponds to the period between January 1994 and September 2005, as
defined in section 2.3. The y-scale is millions of deaths per month. The x-scale is in months.
The blue line is the fitted function. The residual is shown at the bottom.
The single-month spike that corresponds to the August-2003 heat wave is seen at
month number 116, and, in our fit, corresponds to a spike area (heat wave deaths) of
19 K deaths. Note that our goal here was not to determine an accurate number of
deaths for the heat wave itself but rather to correctly represent the total mortality profile
in this period. We obtain a more accurate value of 15.3 K deaths for this heat wave by
our analysis of the higher resolution all-cause mortality by day (Insee, 2019) (not
shown). The difference (19 K versus 15.3 K) occurs because higher resolution data
provides greater power to separate overlapping contributions in a given region of the
Figure 7 shows our fit, and its residual, for region-I (2005-August through 2020-June):
Figure 7. Fit of the monthly all-cause mortality data of metropolitan France 1946-2020,
region-I. Region-I corresponds to the period between August 2005 and June 2020, as defined
in section 2.3. The y-scale is millions of deaths per month. The x-scale is in months. The blue
line is the fitted function. The residual is shown at the bottom.
The month-wide “COVID-peak” is seen, centered at the end of March 2020, straddling
March and April, as seen in Figure 3. In this fit (Figure 7), the “COVID-peak” has an
estimated area of 41 K deaths. This estimate is limited in accuracy by two main factors:
(i) the low temporal resolution of the mortality by month data, which limits the power to
separate overlapping contributions, and (ii) the missing mortality by month data beyond
June 2020. These problems are resolved in our analysis of the mortality by day data, as
Accurate quantification of the deaths in the complete “COVID-peak” is obtained by fitting
the all-cause mortality by day for France for 1 March 2020 through 30 June 2020,
shown in Figure 3. The fit uses a linear sloped background for the non-COVID-peak
components and two Voigt lines (Npeak = 2) for the “COVID-peak”, as shown in Figure 8:
Figure 8. Fit of the daily all-cause mortality data of France (metropolitan + overseas),
from 1 March to 30 June 2020. The y-scale is millions of deaths per day. The x-scale is in
days. The blue line is the fitted function. The residual is shown at the bottom.
This fit gives an accurate “COVID-peak” area equal to 30.2 K deaths, which is
approximately double the deaths from the August-2003 heat wave in France, and which
we attribute to the total deaths in France due to government interventions responding to
the declared “pandemic”.
3.6. Graphical analysis of the model-fitting results
In examining our fit results for metropolitan France 1946-2020, we first calculate the all-
cause mortality per cycle-year, as defined in section 2.3.
Figure 9 shows the all-cause mortality per cycle-year for metropolitan France 1946-
2020, compared to the all-cause mortality per calendar-year for the same data:
Figure 9. All-cause mortality by cycle-year and by calendar-year in metropolitan France
from 1946 to 2020. The grey line shows the data per cycle-year (centered in January), meaning
that the year of the month of January in the winter peak is used on the x-axis. The black line
shows the data per calendar-year (direct sum). The cycle-year values were obtained by fitting,
as described in section 2.3.
The break that occurs between 1986 and 1987 is probably an artifact of the data
collection method. There may be another such break between 1961 and 1962. Overall,
there is a decline of mortality per year after the Second World War and up to 1961,
plateaus in mortality per year for the periods 1962-1986 and 1987-2008, and a steady
and steep increase starting at approximately 2008 through to the present. The latter
steady and steep increase is essentially the same as reported by Insee (2020a) for the
yearly mortality data for France, 1982-2019.
The latter 2008-present rise in all-cause mortality per year is remarkable, approximately
double than can be accounted for by the increasing population with a constant age
structure. How is this dramatic break and increase, which also occurs in Canada and
the USA, not a “pandemic”? It has not attracted any media attention, to our knowledge.
Was it caused by the global economic crash of 2008, which many economists compare
to the Great Depression (Bordo and James, 2009; Shaikh, 2010; Chang et al., 2013;
O’Brien, 2018)? There is a surprising media and academic-research relative silence
regarding this compelling public health phenomenon (Figure 9), although some
research for other countries is tangentially relevant (e.g., Falagas et al., 2009; Stuckler
et al., 2009; Ruhm, 2016).
Figure 10 shows the all-cause mortality in metropolitan France per cycle-year (as
defined in section 2.3), as a percentage of the population of metropolitan France
evaluated on 1 January of each year, for the 1946-2020 period:
Figure 10. All-cause mortality by cycle-year in metropolitan France from 1946 to 2020, as
a percentage of French metropolitan population over the same period. The x-axis year is
the year of the January in the cycle-year (the January of the winter season). The population is
for 1 January of each year. The population data was retrieved from Insee (Insee, 2020c), as
stated in Table 1.
Again, we note the dramatic upturn at approximately 2008. Mortality on a per capita
basis decreases steadily after the Second World War, and then the trend is reversed to
increasing mortality, starting at approximately 2008.
The estimate of cycle-year mortality for nominally 2020 is expected to be fairly good
because the fit (Figure 7) reasonably completes the 2019-2020 winter peak, down to
the expected 2020 summer trough (and see Figure 3).
With Figure 10, it is difficult to see the latest winter cycle that includes the “COVID-
peak” as extraordinary. The value does not appear to warrant any extreme reaction, in
the context of the entire 1946-2020 trend and its both regular and statistical variations.
By comparison, the upturn in yearly all-cause mortality, which is initiated at
approximately 2008, is real and does warrant public concern and a public-health
investigation. It seems unreasonable to concentrate on an external-event disaster
(“COVID-peak”), while ignoring a massive and systematic health issue easily detected
after analyzing all-cause mortality data.
Figure 11 shows the numbers of winter-burden deaths for metropolitan France 1946-
2020, which result from our fits of the data for all-cause mortality by month:
Figure 11. Winter-burden mortality in metropolitan France from 1946 to 2020. The data
results from the fit of monthly all-cause mortality in metropolitan France, 1946-2020. The x-axis
year is the year of the January in the cycle-year (the January of the winter season).
In Figure 12, the same numbers of winter-burden deaths for metropolitan France 1946-
2020, which result from our fits of the data for all-cause mortality by month, are
expressed as percentages of the total all-cause mortality per cycle-year, for each given
cycle-year having its own winter-burden mortality:
Figure 12. All-cause winter-burden mortality as a percentage of yearly all-cause mortality
in metropolitan France from 1946 to 2020. The data are cycle-year based (see section 2.3).
The x-axis year is the year of the January in the cycle-year (the January of the winter season).
The anti-correlation in time for year to year values (a low year is followed by a high year,
and a high year is followed by a low year), especially prominent in the early years
following the Second World War, seen in both Figure 11 and Figure 12, is real and can
be interpreted as follows: winter-burden mortality is a convolution between the
prevailing pathogenic conditions and the population of immune-vulnerable individuals
(i.e. population of fragile mostly elderly persons). A winter that relatively devastates the
fragile-person population leaves a relatively small such population for the following
winter, and vice versa. The year-to-year effect is greatest to the extent that the mean
lifetime of a concerned fragile person is one year. In other words, the one-year time
anti-correlation is predominantly from the number of individuals having a one-year mean
lifetime or life expectancy.
This shows that it would be ill-advised to assign such year-to-year variations in winter-
burden mortality to virulence of the particular year’s seasonal viral pathogens. The
changes are more a function of the general health status of the population, and the
population numbers of the most vulnerable individuals, rather than virulence of a
particular pathogen. It would be incorrect to postulate that viral virulence progressively
decreased after the Second World War in France, just as it would be incorrect to
interpret relatively small variations occurring in recent decades as being due to year-to-
year changes in virulence of the seasonal pathogens.
Figure 12 shows that 2019-2020 was not a statistically unusual cycle-year in France, in
terms purely of the total number of winter-burden deaths, which include the anomalous
“COVID-peak” deaths. Is this because mitigation measures were effective in the
presence of an exceptionally virulent pathogen? On the contrary, as explained above,
the “COVID-peak” anomaly must be interpreted as the result of an exceptional imposed
perturbation in the society. The “COVID-peak” would not have occurred in the absence
of the said perturbation, and some 30.2 K lives would have been saved in France.
4. Mechanistic causes for “COVID-peak” deaths
In light of epidemiological history, we have proven that the “COVID-peak” feature that is
present in the all-cause mortality data of certain mid-latitude Northern hemisphere
jurisdictions, including France, cannot be a natural epidemiological event occurring in an
absence of an external non-pathogenic perturbation. This is true because the “COVID-
i. occurs sharply (one-month width) at an unprecedented location in the
seasonal cycle (centered at the end of March),
ii. is absent in many jurisdictions (34 of the USA States have no “COVID-peak”),
iii. varies widely in magnitude from jurisdiction to jurisdiction in which it occurs
(such as the example of Ontario and Quebec, Figure 4).
Such a feature in all-cause mortality by week or month has never previously occurred in
known epidemiological data, except with exceptional events such as the August-2003
heat wave in France, or regional earthquakes. Barring such exceptional events, the
known all-cause mortality curves for populations in the entire mid-latitude Northern
hemisphere are remarkably the same; without disappearing or appearing peaks in
different geographical locations, and without peaks occurring at unusual times in the
We end this article by outlining a mechanism wherein one aspect of government
responses could have caused the excess 30.2 K deaths in the “COVID-peak”.
We believe that the unprecedented strict mass quarantine and isolation of both sick and
healthy elderly people, together and separately, would have killed many of them, and is
the main cause of the “COVID-peak” event that we have identified.
By the said mass quarantine in care homes and establishments, the State isolated
vulnerable elderly persons from their families, limited movements within establishments,
often confining individuals to their rooms or beds for days and weeks if not months,
reduced the staff and allowed staff to take extended or frequent sick leaves, forced staff
to adopt extreme measures such as masks, shields and gloves, which can induce a
measure of fear or terror, created a general atmosphere of danger, and prevented air
circulation by locking doors and windows, and by preventing ingoing and outgoing traffic
except for essential services (Campbell, 2020; Comas-Herrera, Fernandez, et al., 2020;
This would have both: retained the pathogen-bearing aerosol particles suspended in the
air without their evacuation (Morawska and Milton, 2020); and induced psychological
stress in the residents.
Psychological stress is known:
i. to be a major factor causing diseases, including immune response
dysfunction, depression, cardiovascular disease and cancer (Cohen, Janicki-
Deverts and Miller, 2007),
ii. to be a dominant factor in making an individual susceptible to viral respiratory
diseases, in terms of intensity of the infection (Cohen, Tyrrell and Smith,
iii. to have more deleterious effects in elderly persons than in younger persons
(Prenderville et al., 2015).
Furthermore, social isolation itself, in addition to individual psychological stress, is
known to have an added impact on the said susceptibility to viral respiratory disease
(Cohen et al., 1997).
In addition, there is a longer term “abandonment of life” phenomenon that occurs with
imposed extended isolations of elderly persons, the so-called “glissement” syndrome (or
“slipping away syndrome” or “geriatric failure to thrive”), which is analogous to
depression (Robertson and Montagnini, 2004; Clegg et al., 2013; Steptoe et al., 2013;
Ong, Uchino and Wethington, 2016).
The suddenly applied national policy of forced quarantine and the psychological stress it
generated on fragile elderly people was certainly a major contributor in the decrease of
efficiency of immune system response to a viral respiratory disease (Comas-Herrera,
Zalakaín, et al., 2020) and this is today the most probable explanation for the most part
of the sharp and narrow mass excess death peak that occurred in March-April 2020 in
France. The same mechanism would operate in any setting (facility, group home, home,
hospital) where persons with health vulnerabilities are isolated and susceptible to
We claim that this mechanism is what occurred, as first suggested by Rancourt (2020),
and that this caused some 30.2 K deaths in France in March and April 2020, not any
viral respiratory disease or combination of such acting naturally in an unperturbed
Bordo, M. D. and James, H. (2009) The Great Depression Analogy. SSRN Scholarly
Paper ID 1522373. Rochester, NY: Social Science Research Network. Available at:
Campbell, A. D. (2020) ‘Practical Implications of Physical Distancing, Social Isolation,
and Reduced Physicality for Older Adults in Response to COVID-19’, Journal of
Gerontological Social Work, pp. 1–3. doi: 10.1080/01634372.2020.1772933.
CDC (2020) National Center for Health Statistics Mortality Surveillance System.
Available at: https://gis.cdc.gov/grasp/fluview/mortality.html (Accessed: 29 July 2020).
Chang, S.-S. et al. (2013) ‘Impact of 2008 global economic crisis on suicide: time trend
study in 54 countries’, BMJ. British Medical Journal Publishing Group, 347. doi:
Clegg, A. et al. (2013) ‘Frailty in elderly people’, Lancet (London, England), 381(9868),
pp. 752–762. doi: 10.1016/S0140-6736(12)62167-9.
Cohen, S. et al. (1997) ‘Social Ties and Susceptibility to the Common Cold’, JAMA,
277(24), pp. 1940–1944. doi: 10.1001/jama.1997.03540480040036.
Cohen, S., Janicki-Deverts, D. and Miller, G. E. (2007) ‘Psychological Stress and
Disease’, JAMA, 298(14), pp. 1685–1687. doi: 10.1001/jama.298.14.1685.
Cohen, S., Tyrrell, D. A. J. and Smith, A. P. (1991) ‘Psychological Stress and
Susceptibility to the Common Cold’, New England Journal of Medicine. Massachusetts
Medical Society, 325(9), pp. 606–612. doi: 10.1056/NEJM199108293250903.
Comas-Herrera, A., Fernandez, J.-L., et al. (2020) ‘COVID-19: Implications for the
Support of People with Social Care Needs in England’, Journal of Aging & Social Policy,
32(4–5), pp. 365–372. doi: 10.1080/08959420.2020.1759759.
Comas-Herrera, A., Zalakaín, J., et al. (2020) ‘Mortality associated with COVID-19
outbreaks in care homes: early international evidence’.
Cummins, I. (2020) Crucial Viewing - to truly understand our current Viral Issue
#Casedemic. Available at: https://youtu.be/FU3OibcindQ (Accessed: 17 August 2020).
Deslandes, A. et al. (2020) ‘SARS-CoV-2 was already spreading in France in late
December 2019’, International Journal of Antimicrobial Agents, 55(6), p. 106006. doi:
Dowell, S. F. (2001) ‘Seasonal variation in host susceptibility and cycles of certain
infectious diseases.’, Emerging Infectious Diseases, 7(3), pp. 369–374.
Evin, C. et al. (2004) N° 1455 - 01 - Rapport de la commission d’enquête sur les
conséquences de la canicule (MM. Claude Evin, François d’Aubert) (tome I),
Assemblée Nationale. Available at: http://www.assemblee-nationale.fr/12/rap-
enq/r1455-t1.asp (Accessed: 4 August 2020).
Falagas, ME et al. (2009) 'Economic crises and mortality: a review of the literature.' Int J
Clin Pract. 2009;63(8):1128-1135. https://doi.org/10.1111/j.1742-1241.2009.02124.x
France 24 (2020) REPLAY - Coronavirus : Allocution d’Emmanuel Macron à propos du
Covid-19 en France. Available at: https://youtu.be/uSZFA0xLQsQ (Accessed: 16 August
Hémon, D. and Jougla, E. (2004) ‘Surmortalité liée à la canicule d’août 2003 - Rapport
remis au Ministre de la Santé et de la Protection Sociale - INSERM’, p. 76.
Insee (2019) Les décès en 2018 - Tableaux de séries longues - État civil - Insee
Résultats. Available at: https://insee.fr/fr/statistiques/4204054?sommaire=4204068
(Accessed: 17 August 2020).
Insee (2020a) Décès et taux de mortalité - Données annuelles de 1982 à 2019.
Available at: https://www.insee.fr/fr/statistiques/2383440 (Accessed: 30 June 2020).
Insee (2020b) Nombre de décès quotidiens - France, régions et départements.
Available at: https://www.insee.fr/fr/statistiques/4487854 (Accessed: 04 August 2020).
Insee (2020c) Série 000067670 Population totale au 1er janvier - France métropolitaine.
Available at: https://www.insee.fr/fr/statistiques/serie/000067670 (Accessed: 05 August
Insee (2020d) Série 000436394 Démographie - Nombre de décès - France
métropolitaine. Available at: https://www.insee.fr/fr/statistiques/serie/000436394
(Accessed: 30 July 2020).
Insee (2020e) Série 001641603 Démographie - Nombre de décès - France (inclus
Mayotte à partir de 2014). Available at:
https://www.insee.fr/fr/statistiques/serie/001641603 (Accessed: 30 July 2020).
Ioannidis, J. (2020) ‘The infection fatality rate of COVID-19 inferred from seroprevalence
data’, medRxiv. Cold Spring Harbor Laboratory Press, p. 2020.05.13.20101253. doi:
Lagarec, K. and Rancourt, D.G. (1998) Recoil User Manual -- Mossbauer spectral
analysis software for Windows.
Marti-Soler, H. et al. (2014) ‘Seasonal Variation of Overall and Cardiovascular Mortality:
A Study in 19 Countries from Different Geographic Locations’, PLOS ONE. Public
Library of Science, 9(11), p. e113500. doi: 10.1371/journal.pone.0113500.
Morawska, L. and Milton, D. K. et al. (239 signatories) (2020) ‘It is Time to Address
Airborne Transmission of COVID-19’, Clinical Infectious Diseases. doi:
O’Brien, M. (2018) ‘The 2008 crisis really did start off worse than the Great Depression’,
The Washington Post, 15 September. Available at:
Ong, A. D., Uchino, B. N. and Wethington, E. (2016) ‘Loneliness and Health in Older
Adults: A Mini-Review and Synthesis’, Gerontology, 62(4), pp. 443–449. doi:
Paules, C. and Subbarao, K. (2017) ‘Influenza’, The Lancet. Elsevier, 390(10095), pp.
697–708. doi: 10.1016/S0140-6736(17)30129-0.
Prenderville, J. A. et al. (2015) ‘Adding fuel to the fire: the impact of stress on the ageing
brain’, Trends in Neurosciences, 38(1), pp. 13–25. doi: 10.1016/j.tins.2014.11.001.
Rancourt, D.G. (2019) ‘QUICK INSTALLATION GUIDE AND RECOMMENDATIONS for
Recoil’, Technical report, ResearchGate.
Rancourt, D.G. (2020) ‘All-cause mortality during COVID-19: No plague and a likely
signature of mass homicide by government response’, ResearchGate. doi:
Robertson, R. G. and Montagnini, M. (2004) ‘Geriatric failure to thrive’, American Family
Physician, 70(2), pp. 343–350. https://pubmed.ncbi.nlm.nih.gov/15291092/
Ruhm, CJ. (2016) 'Health Effects of Economic Crises'. Health Econ. 2016;25 Suppl 2:6-
Shaikh, A. (2010) The First Great Depression of the 21st Century. The Merlin Press.
(Accessed: 19 August 2020).
StatCan (2020) Weekly death counts: interactive tool. Available at:
03 August 2020).
Steptoe, A. et al. (2013) ‘Social isolation, loneliness, and all-cause mortality in older
men and women’, Proceedings of the National Academy of Sciences of the United
States of America, 110(15), pp. 5797–5801. doi: 10.1073/pnas.1219686110.
Stuckler, D. (2009) 'The public health effect of economic crises and alternative policy
responses in Europe: an empirical analysis'. Lancet. 2009;374(9686):315-323.
Wu, B. (2020) ‘Social isolation and loneliness among older adults in the context of
COVID-19: a global challenge’, Global Health Research and Policy, 5, p. 27. doi:
View publication statsView publication stats