Stress and measures killed suddenly during the so-called COVID pandemic

CORRELATION responds to comments by Professor Devon Brewer on its latest research report

CORRELATION’s latest research report “Constraints from geotemporal evolution of all-cause mortality on the hypothesis of disease spread during Covid” has received some thoughtful commentary from the research community posted at preprints.org.

Using high-resolution geotemporal all-cause mortality data, CORRELATION researchers Joseph Hickey, Denis G. Rancourt, and Christian Linard showed that mortality patterns during March-May 2020 in Europe and the USA could not have been caused by person-to-person spread of a novel infectious virus. The authors proposed an “alternative hypothesis” that biological-stress-induced self-infection pneumonia and medical mistreatment of patients in hospitals and care homes were responsible for virtually all excess death in excess death hotspots such as New York City, USA, and Lombardy, Italy, during March-May 2020.

Professor Devon Brewer posted public comments on the preprints version of the report, on June 17 and June 23, 2025. Professor Brewer agreed with the main finding of the report (regarding incompatibility of excess death data with the paradigm of person-to-person spread of a novel virus), but raised several points questioning the authors’ alternative hypothesis.

Below, we copy Professor Brewer’s comments, followed by a detailed authors’ response by Hickey, Rancourt, and Linard, dated July 11, 2025.

Professor Brewer replied to the authors’ response with in additional comment dated July 15, 2025, which is copied below the authors’ response, in this post.

Devon Brewer’s Comment of June 17, 2025

Thank you for your detailed and extensive descriptive analyses. You have done Covid researchers a big favor by conducting this work, and I plan to rely on and cite it accordingly. I agree with your conclusion that the excess mortality you described is not consistent with person-to-person epidemic spread, especially in contrast to the expectations from various models.

However, I think the explanations in the Discussion section of what might have caused the spikes in excess mortality are also at odds with the evidence. Stress presumably was fairly constant across time in 2020, once lockdowns and related measures began. Likewise, I am aware of no evidence that harmful healthcare interventions varied over time and space in a way that corresponded to variation in excess mortality. In my quick reading of your lengthy report I did not see any mention of systematic analysis of the topic. Did you discuss any? Constants or variables that change slowly/have limited variation are inherently poor predictors of outcome variables that change rapidly/have large variation. Simply noting that hospitals in areas and times with high mortality used mechanical ventilation, for example, does not demonstrate an association, but merely a co-occurence. To assess an association, there must be a comparison with practices in low mortality areas and times. Clinicians generally select treatments in response to the conditions of their patients, often following fairly inflexible treatment guidelines/algorithms. Presumably, clinicians in low mortality areas and times would select similar treatments for patients with the same conditions even if they saw few such patients. That is, blaming the favored Covid treatments of the day (even if harmful) might be like blaming CPR for cardiac deaths. The treatment follows the condition, and many of the deceased patients might have died regardless of the treatment they received. In such circumstances, it is easy to mistake a treatment as the cause of death. Moreover, a substantial proportion of excess deaths occurred at home, spiking considerably in the spring of 2020 in NYC (https://www.amazon.com/Illustrated-Mortality-Guide-objective-eye-opening-ebook/dp/B0D4JF8QHB; p. 55). Harmful healthcare interventions cannot account for such excess mortality. It does not seem unusual that when many people feel ill (for whatever reason) that they go to hospitals, and that hospitals will account for an increased share of deaths as a result. In short, stress and harmful healthcare interventions don't explain why mortality spiked in one place and not another, nor at one time and not another. For those places and times where excess mortality occurred, these factors also don't explain the variation in the magnitude of excess mortality.

Virginia Stoner has raised many of these points before. See: https://www.virginiastoner.com/writing/2024/5/22/the-nyc-mass-casualty-event-litmus-test, https://www.virginiastoner.com/writing/2024/6/5/the-final-pandemic-or-how-to-stop-worrying-and-love-the-democide, and https://www.virginiastoner.com/writing/2024/7/28/researchers-discover-reason-for-31-million-excess-deaths-worldwide-mistakes-were-made-l2any-j8egn.

Also, your discussion of hydroxychloroquine seems incomplete, as it does not take into account the timing of treatment. A current meta-analysis, with consideration of timing of treatment, is available at https://c19hcq.org/.

Nonetheless, despite these criticisms, I am grateful for your very useful contribution.

Devon Brewer’s Comment of June 23, 2025

I neglected to mention that in your appendix, many/most states show mortality spikes for deaths at home and deaths in healthcare contexts that parallel each other. This further demonstrates that whatever caused the spikes had to be 1) general across all settings (thus seemingly eliminating the harmful healthcare interventions hypothesis) and 2) limited in time, rising and then falling in intensity within a short period (thus seemingly eliminating both the stress and harmful healthcare interventions hypotheses).

Authors’ Response of July 11, 2025

By Joseph Hickey, Denis G. Rancourt, and Christian Linard

Thank you for your thoughtful comments and suggestions, and links. We are grateful to have this opportunity to respond.

The main result of our paper is that excess mortality patterns during the “first-peak period” (March-May 2020) in the USA and Europe are incompatible with the paradigm of person-to-person spread of an infectious respiratory virus. We note that you do not make any criticisms of the said main result of our paper in either of your two comments. Rather, your comments relate to our proposed “alternative hypothesis” about the causes of excess mortality during the first-peak period.

Our alternative hypothesis is that the combination of respiratory illness induced by biological stress (due to the sudden and severe imposition of “lockdowns” and related public health measures in March 2020 in hotspots of excess mortality) and medical mistreatment in hospitals and nursing homes were the causes of essentially all excess mortality during the first-peak period, in jurisdictions with large excess mortality. We argue that our alternative hypothesis is supported by various reports, is consistent with the mortality data, and is not disproved by any feature of the mortality data. Furthermore, our alternative hypothesis has aesthetic value in that the observed jurisdictional inhomogeneity of mortality is naturally associated with the reasonable expectation that there will be significant jurisdictional differences (and discontinuities) in medical and public-health practice, down to institutional and individual levels.

You argue that our alternative hypothesis is incompatible with empirical data regarding mortality, medical treatments, and timing and severity of public health measures. In our view, your arguments in your two comments do not refute our alternative hypothesis.

In your first comment, you argue that:

“Stress presumably was fairly constant across time in 2020, once lockdowns and related measures began.”

We disagree that stress was constant across time in 2020 once lockdowns and related measures began. The stringency of COVID public health measures changed over time, with large changes in stringency occurring on the scale of weeks, including large decreases for many European countries and USA states from a peak in stringency at the end of March or beginning of April 2020 to a lower level in June/July 2020 (Financial Times, 2022; Hallas et al., 2021), and New York City began removing lockdown measures in June 2020 (Wikipedia, 2025).

Importantly, the measures were applied (“turned on”) suddenly in March 2020 in the excess mortality hotspots studied in our paper, increasing from zero intensity to essentially maximum intensity virtually instantaneously, while being amplified by mass media. This sudden unprecedented imposition of serious restrictions was a unique characteristic of the first-peak period, which arguably would have imposed a high level of biological stress on subjected populations.

In addition, virtually everyone in society received the hysteria of a global and lethal danger, including all caregivers and family members and service personnel in contact with those who died (the sick, frail and elderly), irrespective of the place of death. There are many reports of terrified health care workers and long-term care workers who abandoned their posts, and many families “protected” their elderly by reducing social contact, where social isolation is known to dramatically increase biological stress and susceptibility to respiratory infection (e.g., the extensive work on young adults of Professor Sheldon Cohen of the USA). The deadliness of biological stress and the dominant factors of hysteria, interrupted care and social isolation among the frail and elderly cannot easily be overestimated, as is well demonstrated in the scientific literature. The hysteria itself has a natural and rapid decay time, as it cannot be maintained for long periods without lethal consequences to the individual. We have cited the scientific literature in these regards in many papers about all-cause mortality.

Whether subsequent (post-March-May 2020) changes (increases or decreases) in the nature and severity of the measures had greater or lesser impact on biological stress experienced by different groups within society — including the role of chronic vs episodic stress, and adaptation vs failure of adaptation to stress (see Rancourt, 2024) — is an important topic for epidemiological research on COVID-period mortality.

It is plausible that the intensity of biological stress induced by lockdowns was much higher during March-May 2020 than in the remainder of 2020, in the states with the highest first-peak period excess mortality.

Furthermore, regarding sharpness of a large mortality peak, on the timescale of weeks there would be an important “die off” or “dry tinder” effect in the populations most at risk. In prior and ongoing papers and in presentations (e.g., Rancourt, 2025) we demonstrated such strong effects in which mortality is abnormally small (below the historic trend) immediately following a peak in excess mortality in the older age groups.

In your first, comment, you also state:

“Likewise, I am aware of no evidence that harmful healthcare interventions varied over time and space in a way that corresponded to variation in excess mortality.”

To the contrary, the medical system tempered its initially reckless application of mechanical ventilation and associated pharmaceutical treatments (e.g. midazolam) as data and studies emerged demonstrating the dangers of these treatments. Also consider the case of hydroxychloroquine, which was widely (and rashly) adopted and then abandoned as a recommended treatment by many high-profile institutions.

However, we admit that the quantitative magnitude of mortality directly induced or accelerated by harmful healthcare interventions, as opposed to primarily biological stress or other causes, is not presently known and remains a hypothesis for future attempts at quantification.

In your first, comment, you also state:

“Simply noting that hospitals in areas and times with high mortality used mechanical ventilation, for example, does not demonstrate an association, but merely a co-occurence. To assess an association, there must be a comparison with practices in low mortality areas and times. Clinicians generally select treatments in response to the conditions of their patients, often following fairly inflexible treatment guidelines/algorithms. Presumably, clinicians in low mortality areas and times would select similar treatments for patients with the same conditions even if they saw few such patients. That is, blaming the favored Covid treatments of the day (even if harmful) might be like blaming CPR for cardiac deaths. The treatment follows the condition, and many of the deceased patients might have died regardless of the treatment they received. In such circumstances, it is easy to mistake a treatment as the cause of death.”

In particular, you state:

“Presumably, clinicians in low mortality areas and times would select similar treatments for patients with the same conditions even if they saw few such patients.”

We do not disagree with this point (which does not disprove our hypothesis), and we would add the following.

In our paper, we pointed out that areas with low first-peak period excess mortality in Italy (section 4.4.1 of our paper) and the USA (section 4.4.2 of our paper) adopted approaches in which individuals were essentially encouraged to remain home when sick rather than attend hospitals, whereas the opposite was done in Lombardy and New York City. Fewer patients being treated with killer treatments would result in fewer patients killed overall. Also, high first-peak period excess mortality areas such as Lombardy and New York City were sites with intense governmental and media messaging about the threat of the virus, resulting in a higher level of biological stress for all residents in hotspots as compared to other locations, for the same reasons that we mention above.

Basically, your point describes a particular difficulty in demonstrating a high likelihood of the medical harm part of our hypothesis (versus the biological stress part) using available all-cause mortality data and reports about medical practice, which we admit. The said difficulty will be particularly large given the medical establishment’s proven systemic aversion to admitting iatrogenic harm (e.g., James, 2013).

In your first, comment, you also state:

“Moreover, a substantial proportion of excess deaths occurred at home, spiking considerably in the spring of 2020 in NYC (https://www.amazon.com/Illustrated-Mortality-Guide-objective-eye-opening-ebook/dp/B0D4JF8QHB; p. 55). Harmful healthcare interventions cannot account for such excess mortality.”

You are correct that many USA states, including those with large first-peak period excess mortality, had a higher than normal number of deaths that occurred at home during March-May 2020. However, for all US states with large first-peak period excess mortality (large “F-peaks” using our terminology), the increase in deaths in hospitals and/or nursing homes in March-May 2020 compared to March-May 2019 was significantly larger than the increase in deaths at home in March-May 2020 compared to March-May 2019. This can be seen in the figures in Appendix B of our paper; for example, see the figures in Appendix B for the five states with the largest first-peak period excess mortality: New York, New Jersey, Connecticut, Massachusetts, and District of Columbia.

We believe that respiratory disease such as pneumonia induced by biological stress could be responsible for many of the excess deaths that occurred at home during March-May 2020. Again, see: Rancourt (2024). To this we can add the known denial of treatment with antibiotics associated with a documented significant drop in non-hospital prescriptions of antibiotics despite proven high mortality certified as associated with bacterial pneumonia in the USA (Rancourt et al., 2021). The large drop in antibiotic prescriptions starting when a pandemic was declared appears to have been systematic and global, as described in our paper and in several reports. The first months of the pandemic was a period of highly reduced contacts with non-hospital treating physicians.

Also, given government pronouncements and media reports about hospitals and contagion in the first months, we can expect that many individuals would decide to stay home even if they were very ill with any condition rather than travel to enter the hospital system, more so than in non-pandemic times, while not having normal access to out-patient or non-hospital care for any illness.

In your first comment, you also state:

“It does not seem unusual that when many people feel ill (for whatever reason) that they go to hospitals, and that hospitals will account for an increased share of deaths as a result.”

We might add that beyond going to a hospital because one “feels ill”, during March-May 2020 many individuals would have gone to a hospital following a positive COVID test (regardless of presence or absence of any respiratory symptoms).

In your first comment, you also state:

“In short, stress and harmful healthcare interventions don't explain why mortality spiked in one place and not another, nor at one time and not another. For those places and times where excess mortality occurred, these factors also don't explain the variation in the magnitude of excess mortality.”

On the contrary, we would expect the type and magnitude of biological, interventional and psychological stress imposed on sick, frail and elderly subjects to have a large institutional, organizational, cultural and jurisdictional variability, consistent with the jurisdictional boundary effects that we quantify. Management is a lead determinant of iatrogenic misbehaviour. The societal (Western World) pre-Covid context is one where at least one third of deaths are iatrogenic, at base level (James, 2013).

In your second comment, you state:

“I neglected to mention that in your appendix, many/most states show mortality spikes for deaths at home and deaths in healthcare contexts that parallel each other. This further demonstrates that whatever caused the spikes had to be 1) general across all settings (thus seemingly eliminating the harmful healthcare interventions hypothesis) and 2) limited in time, rising and then falling in intensity within a short period (thus seemingly eliminating both the stress and harmful healthcare interventions hypotheses).”

We disagree with these two points, as follows.

The fact that “whatever caused the spikes [in excess mortality] had to be 1) general across [many] settings” does not “eliminate” the “harmful healthcare interventions” aspect of our alternative hypothesis, because our hypothesis is that a combination of pneumonia induced by biological stress and mistreatment of vulnerable individuals in hospitals and care homes is responsible for the excess death.

The argument that “whatever caused the spikes [in excess mortality] had to be […]2) limited in time, rising and then falling in intensity within a short period” does not refute our alternative hypothesis. In our view, an imposition of sudden, unprecedented, and unexpected serious restrictions (such as: cutting contact with family and support networks, confining people to small living quarters, subjecting people to intense media and governmental messaging about imminent danger, etc.) could be sufficient to subject many people to severe biological stress, potentially causing many deaths from conditions such as pneumonia shortly following the initiation of the restrictions (rise side of the excess mortality peak). Prolongation of the measures for several months would not necessarily produce a maintained (constant) high level of induced biological stress, given adaptation responses.

Also, regarding the “die off” effect mentioned above, many of the most vulnerable people in the hotspot areas would have already died during the first few months following the WHO’s March 11, 2020 pandemic declaration, leaving a smaller population of vulnerable people who could die as time progressed. Furthermore, the stringency of measures in excess mortality hotspots generally fell from a peak of stringency in late March 2020 / early April 2020 to lower values as of June/July 2020 (fall side of the excess mortality peak).

Basically, excess mortality (signal) is a convolution between the epidemiological landscape of sick, frail and elderly populations (response function) and the external causes and their time series (drivers). Cause-specific excess mortality is not necessarily expected to have the same temporal variation as the driver.

We also note that there are several jurisdictions in the USA that do not have clearly defined “peaks” of excess mortality during March-May 2020 (with clearly defined rise to a maximum then fall from the maximum), but rather exhibit an initial rise followed by a plateau of excess death. Examples include the states of California, Ohio, Indiana, Kentucky, and New Mexico (see Appendix A in our paper).

Again, thank you for your comments and for studying our paper. We continue to have the opinion that our alternative hypothesis (sudden increased medical harm and sudden fatal consequences of increased biological stress) is reasonable, likely, and not disproved by any known fact. We have not identified other alternative hypotheses (alternative to a viral pandemic) that would be better, and it appears that you have not expressly suggested a specific different hypothesis that you would consider to be better.

References for Authors’ Response of July 11, 2025

Financial Times, 2020: Financial Times, “Lockdowns compared: tracking governments’ coronavirus responses” 23 December 2022: https://ig.ft.com/coronavirus-lockdowns/. [See the embedded interactive graph using data from the Oxford Stringency Index, with Source: Blavatnik School of Government, Oxford University.]

Hallas et al., 2021: Hallas L, Hatibie A, Koch R, et al. Variation in US states’ COVID-19 policy responses. BSG Working Paper Series. BSG-WP-2020/034, Version 3.0, 7 May 2021. https://www.bsg.ox.ac.uk/sites/default/files/2021-05/BSG-WP-2020-034-v3.pdf.

James, 2013: James JT. A new, evidence-based estimate of patient harms associated with hospital care. J Patient Saf. 2013;9:122-188. https://doi.org/10.1097/PTS.0b013e3182948a69.

Rancourt, 2024: Rancourt, DG. Medical Hypothesis: Respiratory epidemics and pandemics without viral transmission. CORRELATION Research in the Public Interest, Report, 02 December 2024. https://correlation-canada.org/respiratory-epidemics-without-viral-transmission/.

Rancourt, 2025: Rancourt, DG. Towards understanding the government assault and persistent excess mortality. Presentation at An Injection of Truth: Healing Humanity, 03 March 2025, Calgary, Alberta, Canada. https://denisrancourt.ca/videos.php?id=135.

Rancourt et al., 2021: Rancourt DG, Baudin M, Mercier J. Nature of the COVID-era public health disaster in the USA, from all-cause mortality and socio-geo-economic and climatic data. ResearchGate. 25 Ocbtober 2021. Accessed on 2025-03-12 from: http://dx.doi.org/10.13140/RG.2.2.11570.32962. Also available at: https://correlationcanada.org/Mortality-public-health-disaster-USA/.

Wikipedia, 2025: COVID-19 pandemic in New York City. https://en.wikipedia.org/wiki/COVID-19_pandemic_in_New_York_City.

Devon Brewer’s comment of July 15, 2025

Thank you for your reply to my comment.

You are correct that my inferences that stress was general and that treatment protocols were also general do not falsify these respective hypotheses. However, these common sense inferences are grounds to be skeptical of the hypotheses in the absence of any systematic evidence that variations in stress and treatment protocols correlated with variations in excess mortality. You provided no such empirical evidence of these correlations in your report or in your response. Therefore, I conducted a 20-minute, incomplete, and somewhat informal search of Google Scholar with respect to invasive mechanical ventilation. I found several articles with results consistent with my inferences that do indeed falsify the harmful healthcare interventions hypothesis, at least with respect to invasive mechanical ventilation.

Douin and colleagues assessed invasive mechanical ventilation use for Covid and non-Covid patients in intensive care units (ICUs) at 13 academic medical centers distributed across the USA in the spring of 2020 (https://pmc.ncbi.nlm.nih.gov/articles/PMC7994039/pdf/cc9-3-e0361.pdf). One of the centers was Montefiore Medical Center in the Bronx (New York City), the county in the USA with the highest excess mortality in this period by your estimates (P = 233). Montefiore did indeed use invasive mechanical ventilation for a very high percentage (~ 90% by my visual estimate) of the ICU patients in April. The corresponding percentages in April for the other 12 medical centers ranged from approximately 25% to 90%, with a rough median estimate of about 60%. The other medical center with an approximately 90% invasive mechanical ventilation rate was Hennepin County Medical Center. Hennepin County (Minneapolis) had dramatically less excess mortality (P = 32) than the Bronx (P = 233), according to your appendix.

Oliviera and colleagues reported on invasive mechanical ventilation use for Covid patients in ICUs at all 9 hospitals in the largest healthcare system in Orange County (Orlando), Florida, from March to May, 2020 (https://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0249038&type=printable). Eight-three percent of Covid patients during this period received invasive mechanical ventilation. Orange County had essentially no excess mortality during the spring of 2020 (P = 0.7 per your appendix), yet invasive mechanical ventilation was almost as common there as at Montefiore in the Bronx.

Mesotten and colleagues reported on invasive mechanical ventilation in Covid patients admitted to ICUs between March and August 2020 in the Euregio region spanning parts of Belgium, Germany, and the Netherlands. The percentage of patients who received invasive mechanical ventilation ranged from 53% to 100% across the three countries. In fact, these percentages were _inversely_ related to your estimates of excess mortality: Dutch Limburg (89% ventilation rate, excess mortality P = 45); Belgian Limburg (53% ventilation rate, excess mortality P = 46); and Aachen/Koeln region of Germany (100% ventilation rate, excess mortality P = 1.2 for Nordrhein-Westfalen).

These results demonstrate that variations in treatment protocols (for ventilation at least) do not account for the extreme spatial differences in excess mortality in spring of 2020. These reports are not obscure. Douin and colleagues published in 2021 and their article has been cited 52 times. Oliviera and colleagues also published in 2021 and their article has been cited 177 times. Mesotten and colleagues published in 2022 and their article has been cited 30 times. There are likely many other reports that show ventilation was common throughout North America and Europe, regardless of whether there was excess mortality locally. I suspect, but don't know, that the ventilation practices in ICUs during 2020 were not much different in prior years for patients with similar conditions as those in 2020. That's a further matter to investigate for those who think the harmful treatment protocol hypothesis still has value, especially since you have stated that these practices represented sudden changes.

In your reply you wrote "In our paper, we pointed out that areas with low first-peak period excess mortality in Italy (section 4.4.1 of our paper) and the USA (section 4.4.2 of our paper) adopted approaches in which individuals were essentially encouraged to remain home when sick rather than attend hospitals, whereas the opposite was done in Lombardy and New York City. Fewer patients being treated with killer treatments would result in fewer patients killed overall. Also, high first-peak period excess mortality areas such as Lombardy and New York City were sites with intense governmental and media messaging about the threat of the virus, resulting in a higher level of biological stress for all residents in hotspots as compared to other locations, for the same reasons that we mention above." As I noted in a comment on your Substack 5 days ago (https://unbekoming.substack.com/p/hospitals-not-viruses-what-really/comment/133807894), there are no systematic data in the sources you cited about hospital admission policies in Italy. In that Substack commnent, I wrote "Likewise, for the USA, you note that areas that didn't expand ICUs had less excess mortality. But based on the information available, that could easily be explained by those areas not needing to expand ICUs because they were not hit with many patients in severe condition. An analogy might be communities that experience a significant natural disaster opening emergency shelters for displaced persons. Communities not affected as much by the disaster might not open emergency shelters. In short, a correlation in this case -- even if reflecting a causal relationship -- doesn't indicate the direction of that relationship. Do emergency shelters cause natural disasters or the severity of their local effects?" Your assertions about differences in hospital admission policies and practices are speculative and not based on solid evidence.

I myself lived in an area locked down as or more extremely than New York City, yet the area had little excess mortality. I think most people in North America and Europe can attest that the national and international media messaging created widespread and simultaneous stress, both in areas that had excess mortality and those that did not. To override those common subjective experiences, it is necessary to have objective systematic data on the matter, which you have not presented or cited. I leave it to you to conduct and report on a careful search and review of the literature on the relationship (if any) between variations in stress and excess mortality in the spring of 2020.

My not proposing an alternative hypothesis has no bearing on the evaluation of the hypotheses you've proposed. Evaluating an hypothesis does not require proposing a replacement. Indeed, your showing that the patterns of excess mortality are inconsistent with expectations of person-to-person spread is an important result in itself.

Please note that I am not defending any treatment or practice, or promoting any hypothetical cause for the excess mortality. In my opinion, that remains a puzzle. But work like yours helps to winnow the possibilities, and I am grateful for your very useful contribution.