A Complete Refactor

After six months of intensive development and 597 commits, MortalityWatch 3.0 is here — a complete rewrite from the ground up. This isn’t just an update. It’s a reimagining of what mortality data visualization can be.

The Headline Feature: Age-Standardized Deaths

The biggest addition is the new Age-Standardized Deaths (ASD) metric. When comparing mortality across populations, raw death counts can be misleading — a country with an older population will naturally have more deaths, even if their healthcare is identical. ASD applies standard population weights, letting you make fair comparisons across countries, time periods, and demographic groups. This is the gold standard method used by epidemiologists worldwide, and now it’s available to everyone on MortalityWatch.

What Else Is New?

The Discover Page offers a visual way to browse all available charts and metrics — find interesting data you didn’t know existed. Life expectancy tracking now includes period life expectancy at birth, by age, with seasonal adjustments for accurate weekly and monthly trends. Z-Score analysis brings statistical significance testing right into the platform, so you can see whether a spike in deaths is meaningful or noise. Social login lets you sign in with X or Google, and QR code sharing compresses the entire chart state into the URL for instant sharing. Five baseline methods — from simple averages to advanced forecasting models — give you flexibility in how you calculate excess mortality.

For Pro subscribers, the release includes pre-2000 historical data going back to 1950 for some regions, plus resizable charts for custom dimensions.

Under the Hood

The entire platform was rebuilt on Nuxt 4, Vue 3, and TypeScript, covering 320+ regions updated daily from official government sources. It’s fully open source on GitHub.

Free Pro Access

To celebrate the launch, anyone who signs up with the code BMVE-2VQA gets 3 months of free Pro access. Just visit mortality.watch and create an account.

Thank You

This project wouldn’t be possible without your support — whether you’ve subscribed, donated, or simply shared the data. The goal remains the same: making mortality data accessible, transparent, and useful for researchers, journalists, and curious minds everywhere.

Try it out and let me know what you think.

Kevin McKernan - former R&D team leader for the Human Genome Project - attempted to dismiss concerns about the study’s primers cross-reacting with human DNA.

Curious response from a former Human Genome Project team leader - sloppy formatting (no spaces, inconsistent capitalization) paired with ad hominem attacks (”incoherent,” “never performed PCR,” “polarizing,” implying lack of intelligence) rather than addressing the actual science.

Besides this, his four arguments don’t hold up:

Point 1: “Primers don’t match on 3’ end”

Factually wrong. The nested forward primer (ACTTCACCGGCTGTGTGATT) shows a perfect 20/20 match to human Chromosome 1 (NC_000001.11) with q_start=1, q_end=20. The entire primer, including the 3’ end, aligns perfectly. E-value: 0.008.

Point 2: “Different chromosomes = impossible to amplify human DNA”

This assumes the SARS-CoV-2 genome is distinct from human sequences. But is it?

Edward Holmes - co-author of the original genome publication - confirmed that MN908947.1 contained human DNA:

The genome was assembled via MEGAHIT from mixed metagenomic reads - never from isolated viral material. Both the CDC (https://x.com/USMortality/status/1840501046910550277) and RKI (https://x.com/USMortality/status/1876650231850267129) have confirmed no viral genetic material was ever sequenced in isolation.

If “spike” sequences were derived from human reads misassembled as “viral,” then detecting them in unvaccinated humans isn’t cross-reactivity - it’s detecting endogenous human sequences that were mislabeled as viral from the start.

Point 3: “Sanger sequencing confirmed spike”

Circular logic. If spike was originally derived from human reads during genome assembly, confirming you found “spike” proves nothing about its origin. Also, where’s the raw data for public scrutiny?

Point 4: “Pre-2020 controls were negative”

Where’s the data? How many samples? What tissue types? Same PCR protocol?

Even if true - the sequence “appearing” in late 2019 coincides with when they created the reference genome from metagenomic assembly. That’s not proof of viral emergence.

The Real Question: Why Deleted?

The study wasn’t retracted - it was deleted. No DOI, no retraction notice, no explanation.

A retraction requires a formal process with a public record. Deletion leaves no trace.

If the problem were simply “bad primer design,” a retraction notice would suffice. But what if the study accidentally demonstrated something far more inconvenient?

50% of unvaccinated women tested positive for “spike” sequences. What if they weren’t detecting vaccine contamination or primer cross-reactivity - but rather human sequences that were incorrectly labeled as “viral” when the SARS-CoV-2 genome was first assembled in 2020?

Perhaps the study was deleted not because it was wrong, but because it was accidentally right about something it wasn’t supposed to find.

Background: Is the SARS-CoV-2 Genome Valid?

For those unfamiliar with my ongoing analysis:

1. The original SARS-CoV-2 genome (MN908947) was assembled from short metagenomic reads using MEGAHIT - a de Bruijn graph assembler that may create artificial contigs from k-mers. (more)

2. Early versions contained confirmed human DNA contamination.

3. No single long read (or at least several well overlapping long reads) spans the full 29,903 bp genome, that include the exact start and end.

4. All subsequent validation uses PCR primers derived from this unverified sequence (circular logic).

5. CDC and RKI confirm no viral material was ever sequenced in isolation.

👉 The deleted Israeli study may have inadvertently provided another data point supporting this hypothesis!

Before you dismiss this as “anti-vax,” read the methodology and the numbers. This is a data analysis, not an opinion piece.

TL;DR

WHO’s own data doesn’t clearly support vaccine efficacy for measles. The correlations that exist can be explained by confounding.

The Data

I used two datasets from the WHO Global Health Observatory:

- Measles cases (indicator WHS3_62)

- MCV1 vaccination coverage among 1-year-olds (indicator WHS8_110)

- Population data from mortality.watch

The analysis covers all 6 WHO regions: Africa (AFR), Americas (AMR), Eastern Mediterranean (EMR), Europe (EUR), South-East Asia (SEAR), and Western Pacific (WPR).

Finding #1: Only Half the Regions Show Any Correlation

If vaccines reduce measles, we’d expect a negative correlation between coverage and cases everywhere. Instead:

Three regions (Americas, Europe, Eastern Mediterranean) show no statistically significant correlation between vaccination and cases. Europe’s correlation is essentially zero (-0.0004).

Finding #2: The Confounding Pattern

Here’s what’s telling: the correlations appear only in developing regions (Africa, South-East Asia, Western Pacific) and are absent in developed regions (Europe, Americas).

This is exactly what you’d expect from confounding, not causation.

If vaccines caused the decline:

- We’d see negative correlations in ALL regions

- Europe and Americas have high vaccination rates - the effect should be visible

If development caused the decline:

- Developing regions: Sanitation, nutrition, healthcare ALL improved alongside vaccination programs

- Developed regions: These factors were already stable - no correlation appears

👉 The pattern fits confounding perfectly.

Finding #3: The Africa Anomaly

Africa provides the clearest test case. Coverage increased substantially over 24 years:

Cases in 2019 exceeded 2000 levels despite coverage rising from 50% to 71%.

If vaccines worked, this shouldn’t happen. You can’t have 21 percentage points more coverage and end up with more cases than you started with.

Finding #4: Year-Over-Year Analysis Shows Nothing

I tested whether years with increasing coverage correspond to years with decreasing cases:

Regression: Cases_change ~ Coverage_change

Slope: +4.12 (positive, not negative!)

p-value: 0.89 (nowhere near significant)

95% CI: [-52.4, +60.7]

R-squared: 0.00015 (essentially zero)

There is no detectable relationship between coverage changes and case changes. Years when vaccination increased were not associated with fewer cases.

Finding #5: Herd Immunity Threshold Not Observed

The theory says measles requires 92-95% coverage for herd immunity. Above this threshold, cases should plummet.

Reality:

Countries with 95%+ coverage have HIGHER case rates than 90-94% coverage.

Recent high-coverage “outbreaks”:

- Samoa (96% coverage): 27,341 cases per million

- Tonga (99% coverage): 2,130 cases per million

- Kyrgyzstan (96% coverage): 1,125 cases per million

The expected threshold effect simply isn’t visible in the data.

Finding #6: Lag Analysis Doesn’t Help

Maybe vaccines take time to show effect? I tested if coverage predicts cases 1-3 years later:

Africa’s correlation gets weaker with longer lags. This is the opposite of what you’d expect if vaccines caused a delayed reduction.

What This Analysis Shows

1. Temporal correlation exists between vaccination and case decline in developing regions

2. No correlation exists in developed regions despite high coverage

3. Pattern consistent with confounding by development (sanitation, nutrition, healthcare)

4. Africa anomaly - cases exceeded baseline despite coverage nearly doubling

5. No year-over-year relationship (p=0.89)

6. Herd immunity threshold not observed - 95%+ coverage countries still have major outbreaks

Limitations

To be clear about limitations:

- This cannot prove vaccines don’t work

- This cannot prove vaccines do work

- Ecological data cannot establish causation

- This doesn’t tell us what would have happened without vaccination

Alternative Explanation

The pattern is more consistent with:

👉 “Development reduces measles through improved sanitation, nutrition, and healthcare. Vaccination programs are part of development initiatives. The correlation is spurious.”

Than with:

👉 “Vaccines reduce measles cases.”

If vaccines were the causal factor, we’d expect to see the effect in ALL regions, not just those undergoing rapid development.

Conclusion

The WHO’s own data doesn’t clearly support the claim that measles vaccination reduces cases. The correlations that exist can be explained by confounding. The Africa data directly contradicts efficacy claims. The herd immunity threshold isn’t visible.

This doesn’t mean vaccines definitely don’t work. But it does mean the ecological evidence cited to support their efficacy is weak and potentially spurious.

To properly establish efficacy requires double blinded randomized placebo controlled trials, case-control studies, or natural experiments - not ecological time-series correlations, that don’t exist.

Source & Code

• WHO Global Health Observatory, mortality.watch

• R Code

Even stranger: they claimed to find it in 3 of 6 unvaccinated pregnant women.

Then the study vanished - not retracted, just deleted.

I decided to check their primers. What I found illustrates a fundamental challenge with PCR and the so-called human genome…

TL;DR: The study used faulty primers that match the human genome, so they detected human DNA instead of vaccine mRNA - that’s why 50% of unvaccinated people tested “positive”.

All details below:

Analysis of the Deleted Mordechay et al. Study on Vaccine mRNA Detection

The Study’s Claim

A study published in October 2025 claimed to detect Pfizer vaccine mRNA in blood, placenta, and semen of vaccinated individuals - and remarkably, in 50% of unvaccinated women too. The study was completely deleted (not retracted) shortly after publication.

I analyzed their primer sequences against the human genome. Here’s what I found.

The Primers: 8 Out of 12 Were Unusable

The researchers designed 12 PCR primers to detect “vaccine-specific” sequences:

The authors admitted the UTR primers gave “non-specific amplification” - because Pfizer used human gene sequences for the vaccine’s regulatory regions. Only the 4 spike-region primers were used.

The Problem: Those 4 “Vaccine-Specific” Primers Match the Human Genome

I BLASTed all 4 primers against the human reference genome (GRCh38.p14):

The nested forward primer is a PERFECT match to human chromosome 1.

Why This Matters for PCR

The study used nested PCR - two rounds of amplification:

- Round 1: Outer primers (20 cycles)

- Round 2: Nested primers (30 cycles) - generates the final product

With 50 total cycles and primers that match human DNA:

- The nested forward primer (100% human match) will bind human DNA

- The 18/20 matches (90%) can still bind - PCR tolerates 1-2 mismatches, especially at the 5’ end

- Any human sample could potentially give a “positive” result

The Unvaccinated Results Make Sense Now

From the study’s own data:

The detection rate in vaccinated people drops to the same level as unvaccinated over time. This suggests the ~50% “baseline” is likely assay noise from human DNA cross-reactivity, not vaccine mRNA.

3 of 6 unvaccinated women tested “positive” - exactly what you’d expect if the primers are amplifying human sequences.

Conclusion

The study’s “vaccine-specific” primers are not vaccine-specific at all:

- 1 primer perfectly matches human chromosome 1

- 3 primers match human DNA at 18/20 bases (sufficient for PCR binding)

- The 50% positive rate in unvaccinated people likely reflects human DNA amplification, not vaccine detection

This is a fundamental primer design flaw that should have been caught by:

1. The researchers (basic BLAST check before experiments)

2. Peer reviewers

3. The journal editors

Why Deleted Instead of Retracted?

Speculation:

- Retraction requires a formal process - investigation, author response, public record

- Deletion leaves no trace - no DOI, no retraction notice, no explanation

- The flaw is so fundamental (primers match human genome) that it’s not a “mistake” that can be corrected - the entire study is invalid

- A retraction notice explaining “our vaccine-specific primers match human DNA” would be embarrassing and quotable

- Deletion allows the journal and authors to pretend it never happened

TL;DR

Researchers claimed to find vaccine mRNA in 50% of unvaccinated people. Their “vaccine-specific” detection method uses primers that match the human genome - including one perfect 20/20 match to chromosome 1. They were likely detecting human DNA, not vaccine mRNA. The study was deleted rather than retracted, leaving no public record of what went wrong.

---

Technical Details

BLAST Results Against GRCh38.p14

**Primer 1 - `CGTGATCCGGGGAGATGAAG` (1254 F outer):**

- NC_000016.10 (Chromosome 16), position 6,834,975-6,834,994: 90% identity (18/20, 2 mismatches)

**Primer 2 - `ACTTCACCGGCTGTGTGATT` (1337 F nested):**

- NC_000001.11 (Chromosome 1), position 33,048,488-33,048,507: **100% identity (20/20, 0 mismatches)**

**Primer 3 - `TGCTGGAATGGCAGGAACTT` (1745 R nested):**

- NC_000012.12 (Chromosome 12), position 55,432,051-55,432,070: 90% identity (18/20, 2 mismatches)

**Primer 4 - `GGGATCTCTAACGGCGTCTG` (1791 R outer):**

- NC_000003.12 (Chromosome 3), position 187,273,415-187,273,434: 90% identity (18/20, 2 mismatches)

Study Reference

Mordechay L, Baum G, Gabbay-Benziv R, Weinberger H, Morgenstern MF. (2025). Detection of Pfizer BioNTech Messenger RNA COVID-19 Vaccine in Human Blood, Placenta and Semen. Ann Case Rep 10: 2428. DOI: 10.29011/2574-7754.102428

Archived version: https://web.archive.org/web/20251205172931/https:/www.gavinpublishers.com/assets/articles_pdf/Detection-of-Pfizer-BioNTech-Messenger-RNA-COVID-19-Vaccine-in-Human-Blood-Placenta-and-Semen.pdf

CDC data reveals 785 deaths with vaccine-related complications during 2020-2022, based on Multiple Cause of Death (MCD) codes from US death certificates in the CDC's National Vital Statistics System.

I identified vaccine-related ICD-10 codes and analyzed their mortality patterns using a signal-to-noise approach, looking for codes that showed:

1. Significant excess deaths during 2020-2022 (vs. 2017-2019 baseline)

2. Return to baseline by 2023

This pattern isolates deaths temporally associated with the vaccination period but not persisting into 2023, potentially distinguishing vaccine-related mortality.

Thus, these codes can be identified as elevated and potentially related to vaccination:

Y590 - Poisoning by, adverse effect of and underdosing of other and unspecified drugs, medicaments and biological substances, undetermined intent

T887 - Other specified complications of surgical and medical care, not elsewhere classified

T881 - Complications following immunization, not elsewhere classified

T889 - Other specified complications of surgical and medical care, not elsewhere classified

Y14 - Poisoning by and exposure to other and unspecified drugs, medicaments and biological substances, undetermined intent

While this figure may represent a vast understatement due to underreporting (undetected, physician bias, etc. - and thus coded or other causes), these are documented cases coded by medical professionals on death certificates, establishing a conservative baseline estimate for direct vaccine harm.

It is evident, that the novel vaccines used during the vaccinations campaigns likely are responsible for this increase.

Note that if the CDC had coded these with the same methodology as COVID-19, these would all appear as Underlying Cause of Death (UCD) rather than contributing causes. These figures also suggest that injury from these products may be much higher than these documented vaccine deaths.

Historically, vaccines have been withdrawn from market with significantly fewer associated deaths - such as the 1976 swine flu vaccine, which was halted after approximately 45–50 deaths from Guillain-Barré syndrome.

The continued promotion of these vaccines despite accumulating safety signals and absence of demonstrable real-world clinical efficacy represents a critical regulatory failure requiring immediate reassessment.

Source & Code

The German RKI has confirmed they conducted no negative controls during genome sequencing of SARS-CoV-2 and measles viruses. Negative controls—where identical methodology is applied to samples without the target virus—are essential for validating results. Instead, they relied solely on unvalidated PCR testing to confirm presence prior to sequencing.

Here's the English translation of the relevant abstract:

“Both the cultivation of measles and SARS-CoV-2 viruses as well as the genome sequencing of such viruses is carried out at the RKI naturally in accordance with the standards of good scientific practice. However, these do not provide for negative samples to be included during virus cultivation or genome sequencing. Rather, virus detection already takes place beforehand by means of molecular detection via PCR and/or sequencing. Therefore, only pre-analyzed samples are used for virus cultivation or genome sequencing itself, where it has already been established that they contain the virus. Documentation and results for negative controls with healthy patient samples during the cultivation of measles viruses and SARS-CoV-2 or the genome sequencing of such viruses therefore do not exist.”

Key points from this RKI statement:

• They claim to follow "good scientific practice"

• They admit no negative controls are used in cultivation or sequencing

• They rely on prior PCR/sequencing to "confirm" virus presence

• Only "pre-analyzed" samples (already assumed to contain virus) are used

• No documentation exists of negative controls with healthy samples

This confirms circular reasoning - they're using PCR to "prove" virus presence, then using that same assumption to justify not including proper controls in sequencing.

This also directly contradicts findings from my CDC FOIA request, which revealed that no sample has ever been sequenced containing only purified viral genetic material. The CDC confirmed the presence of contaminating genetic material during SARS-CoV-2 sequencing, leaving the possibility wide open that the sequence could be a semi-random construct.

The methodological problem:
This creates a circular reasoning issue. To validate that a genomic sequence is accurate, researchers must either:

1. Sequence a purified sample containing only viral genetic material

   -or-

2. Obtain continuous sequencing reads spanning the entire claimed genome.

To my knowledge, neither approach has been accomplished & documented.

Here's a summary from Claude AI on why negative controls are crucial:

Link to full FOIA

A critical look at genome sequencing, triggered by Clare Craig’s claims.

In a recent article, @ClareCraigPath responded to a comment by @Wood_House76, presenting a long list of reasons why she “believes that the virus travels rapidly across the world.”

In my opinion, framing this as a belief already undermines the argument—science should be grounded in verifiable facts, not beliefs. She then lists several claims to support her position. I’ve read the article, and I’d like to focus specifically on point 3, because I believe the topic of sequencing is widely misunderstood—not only by the public but, quite frankly, even by many geneticists. I’ve studied genomics and sequencing in depth, and have run these tools myself. So here’s what I can say based on verifiable facts:

Addressing Point 3: “Whole Genome Sequencing” as the apparent Confirmation of Spread

The term “whole genome sequencing” is problematic. It suggests that we have sequenced an entire genome from end to end. That is factually incorrect. In reality, the claimed genome has never been sequenced in a single read from start to end, and no method has ever produced a truly complete genome using a single approach.

1. The terminal regions of the genome are consistently missing. No existing algorithm has successfully reconstructed a genome in its entirety.

2. Even Sanger sequencing—often held up as the gold standard—has never yielded a complete genome with full raw data. Its assemblies also lack terminal sequences.

3. What is typically referred to as “the genome” is merely a candidate assembly, often derived from the longest contig (e.g., Wu et al., 2020). That assembly is not exactly reproducible from the raw data, and early genome versions even included confirmed human DNA.

4. The origin of many short reads is ambiguous due to unpurified samples that include human and bacterial material.

5. Tools like MEGAHIT use de Bruijn graph algorithms to assemble genomes from k-mers - substrings of sequencing reads. When applied to mixed samples, these tools can accidentally stitch together k-mers from different organisms, leading to chimeric assemblies that don’t even exist in nature.

6. Over 99% of “genomes” in public databases are not de novo assemblies but are instead aligned to the original Wuhan reference. Fewer than 1,000 genomes have ever been assembled de novo directly from patient samples.

7. Crucially, neither alignment nor assembly methods can mathematically verify that a sequence exists in nature. What we call a genome is, at best, a computational model—a hypothetical reconstruction of a pathogen.

8. Due to these limitations, the SARS-CoV-2 genome remains unvalidated. Proper validation steps - such as independent verification of the genome ends - have not been performed or at least publicly documented.

9. Proponents who defend the validity of the SARS-CoV-2 genome often point to the Human Genome as a “verified” benchmark. However, this is a fallacy. The Human Genome itself is not fully verified - it too is a composite model, stitched together from the DNA of multiple individuals, not from a single human or even a single human cell. As such, it is also incomplete and contains unresolved regions. This raises serious concerns about using the human genome as a reference to filter out ‘non-human’ reads, since the line between what is considered human and non-human in sequencing data remains ambiguous and questionable. After all, has any organism ever been fully recreated just from its claimed genome?

It’s also worth noting that some individuals actively dismiss or downplay these verifiable limitations, often while holding significant conflicts of interest. This should not be ignored!

In another part of her list (point 7), Clare claims that detecting viral material (via testing) in remote locations like Antarctica after long delays and without apparent transmission links supports the idea of rapid global spread.

However, reports of infections arising after long periods of isolation - 14 weeks for flu, 7 weeks for COVID, as seen in trips to Antarctica - could also suggest an endogenous origin, where the detected virus or viral-like genetic material is expressed internally without external infection (i.e. a phenomenon originating from the host itself). This raises a serious question: Are these tests detecting an external pathogen, or simply reacting to genetic material produced endogenously by a sick individual? In that case, what we’re calling “spread” might not be the primary mechanism at all - or at least not fully understood.

P.S. This opinion piece does not claim that viruses do not exist, nor does it endorse such views. It simply questions the validity of the current methods used in genome assembly and pathogen identification, and calls for greater transparency, validation, and scientific rigor. Personally, I am still actively researching and seeking to understand this complex topic. Many questions remain unresolved for me, and I do not claim to have all the answers.

Let me know your thoughts in the comments!

After vaccination, deaths dropped to nearly zero. However, proving causality would require long-term placebo-controlled trials. Charting the pre-vaccination trend from 1949 to 1962 shows that both cases and deaths followed the expected trajectory, meaning the decline might have continued without vaccination. The sharper drop in cases may be influenced by bias, as doctors and parents—assuming vaccination prevents measles—could have attributed symptoms to other causes.

As we can also observe from global epidemiological data, the decline in measles cases is very likely unrelated to vaccination rates.

• In 2008, Africa had its lowest incidence despite only 4% vaccination coverage.

• Similarly, South East Asia and the Western Pacific Region had 2–3% coverage with very low case rates.

No randomized placebo-controlled trials for measles vaccination appear to exist. The renowned Cochrane Institute, while assessing measles vaccination as effective, rates the evidence as only low to moderate certainty, relying solely on observational studies rather than the gold-standard placebo-controlled trials.

Furthermore, measles deaths have remained lower than ten since the early 1980s, often attributed to mass vaccination. However, as shown above, no solid evidence supports this claim. In contrast, an analysis of VAERS reports on measles vaccine-associated deaths compared to confirmed deaths coded as measles (B05) suggests that since 2000, up to 16 times more people may have died from the vaccine than from the disease.

Of course, this comes with the usual challenges of attribution and causality, as well as potential underreporting. A 2010 Harvard study estimated that “[…] fewer than 1% of vaccine adverse events are reported,” suggesting that actual measles vaccine-related deaths in the U.S. could be significantly higher—potentially ranging from 200 to 2,000 per year.

While measles vaccination is widely credited with reducing deaths, the absence of randomized placebo-controlled trials leaves room for uncertainty regarding its true impact. Historical data suggests that measles mortality was already in sharp decline before the vaccine’s introduction, and the Cochrane Institute itself acknowledges only low to moderate certainty in the evidence for its effectiveness. Furthermore, VAERS data indicates that vaccine-related deaths may far exceed those from measles in recent decades, with potential underreporting suggesting a much higher actual toll. Given these factors, a more critical examination of measles vaccination policies and long-term safety data is warranted.

Let me know what you think in the comments!

P.S. Both the AP and Robert F. Kennedy Jr. have falsely claimed that the recent death was the first in a decade. However, as shown above, the last recorded measles death occurred in 2019 and the data during the pandemic is not reliable. While AP has acknowledged this error to me, they have yet to correct their misleading headline.

Code & Sources

1. Lack of Evidence for the Spread of Artificially Created Viruses

   Is there any solid evidence that lab created viruses, particularly those produced through Gain-of-Function (GoF) research, can spread naturally? So far, no scientifically controlled animal experiments have convincingly demonstrated this.

2. Questionable Significance of DARPA Research Proposals

   The research proposals attributed to DARPA could have been deliberately placed and do not necessarily have a direct connection to SARS-CoV-2. These proposals provide no conclusive evidence regarding the origin of the virus and can just as easily be viewed as speculative conjecture.

3. Genetic Evidence is Not Sufficiently Robust

   1. Insufficient Sequence Validation:

      The virus and its sequence have never been fully validated. Neither the CDC, RKI, nor China CDC have sequenced the virus in true isolation (i.e., without other genetic material). Notably, the genome’s terminal sequences were never conclusively verified. The clinical relevance of the sequence is unknown.

   2. Furin Cleavage Site:

      The furin cleavage site, often cited as evidence of manipulation, can also be found in human genetic material. Its short length further limits its reliability as evidence.

   3. HIV Insertion Hypothesis:

      The theory that SARS-CoV-2 contains so-called HIV inserts has been debunked. These sequences are too short, may be coincidental, are incomplete, and are not supported by further studies. The original paper making this claim was retracted.

4. Questioning the Novelty of SARS-CoV-2

   1. The claim that the virus is novel is difficult to prove, as previous research did not use the same methods. Neither the primers nor the test volume used previously were designed to detect it.

   2. There is evidence that antibodies against SARS-CoV-2 reacted years before the official outbreak, and there may be SARS-CoV-2 sequences from as early as March 2019. These observations cast doubt on the assertion that SARS-CoV-2 is genuinely novel.

5. Causality Remains Unproven

   1. Causality cannot be scientifically established based on a single patient from whom the virus was sequenced. Estimates suggest that 60–80% of suspected causes of lung diseases generally remain unknown. How, then, can causality be proven solely based on the sequence from one patient?

   2. Additionally, the PCR test used as evidence is based on this sequence and was applied to sick individuals, with its results being treated as proof of causality. However, it is often overlooked that the test is not clinically validated, meaning its ability to predict the clinical etiology has not been established.

   → Under these circumstances, no reliable causality can be inferred.

Conclusion

While a lab leak remains unproven and may never be definitively confirmed, a far more pressing issue is the need for critical scrutiny of virology as a discipline and its methodologies. It is important to consider that findings based on sequencing and PCR may, in some cases, reflect methodological artifacts rather than biological reality.

Note

I want to clarify that I have no conflicts of interest. I have no relationship, contact, or financial connection with the individuals associated with the lab leak theory or any government agencies. These are solely my personal, well-informed thoughts, and any suggestions to the contrary are unfounded.

More Resources

Here’s a poster with an English and German Version of this article.

Sources

Here’s the answer from Germany (translated from German):

After filing a lawsuit against the German CDC equivalent, RKI (Robert Koch Institute) due to their failure to respond to my Freedom of Information request, the RKI provided a response 139 days late. In their response, the RKI entirely rejected my request, claiming either that the requested information does not exist or that all relevant information is already publicly available. This approach appears to aim at shifting at least part of the legal costs onto me, even though the response was issued only after the lawsuit was filed.

1. Question about the isolation of viral genetic material before sequencing

   My inquiry whether the RKI had sequenced a virus that was physically isolated from other genetic material prior to sequencing was simply sidestepped. The RKI stated that such information is publicly accessible. However, it can be confirmed that this is not described in any public publication. Thus, the RKI indirectly confirms that no sequencing of purified, exclusively virus-originating genetic material has ever been performed.

2. Examination of the reference genome and related genomes:

   1. In response to my question about whether the RKI had internally examined the SARS-CoV-2 genome and its close relative RaTG13, the RKI stated that no such investigations had been conducted at the institute, and therefore no records exist. This indicates that the RKI adopted the reference genome published by Wu et al. (2020) without conducting further verifications. The reference genome was originally derived from a single patient and constructed without additional scientific controls, such as sequencing PCR negative patients with similar etiology from the same hospital.

      1. It remains unclear whether the RKI refers to these specific lineages or generally rules out having examined SARS-CoV-2.

      2. According to the RKI, the protocol used does not allow sequencing of the virus’s terminal regions, although this is deemed irrelevant for lineage determination. While this may be true for lineage determination, it is relevant for the initial validation of the sequence (i.e., origin and clinical relevance).

      3. A search for de novo assemblers (e.g., SPAdes, MEGAHIT, Trinity) in the official RKI pipeline repository (https://github.com/rki-mf1/CoVpipe2) yielded no results. Instead, only the “mapper” or aligner bwa mem was found. The diagram in the README confirms that only mapping for variant determination is performed, not de novo assembly of raw data.

      4. RNA extraction in RKI studies includes total RNA, which includes host RNA and RNA from other organisms:

         1. PMC9278222

            

         2. PMC11074694

         3. PMC8355734

      5. All three studies exclusively use alignment methods (mapping) to reconstruct the SARS-CoV-2 genome, instead of performing de novo assembly.

3. Lack of de novo assembly

   The RKI has no samples or datasets where the reference genome could be perfectly reconstructed using a de novo assembler (e.g., MEGAHIT).

4. Absence of the RACE methodology

   The RACE (Rapid Amplification of cDNA Ends) method, which allows for the amplification and validation of genome ends, was not performed by the RKI.

Conclusion

The RKI indirectly confirms that no sequencing of genetic material exclusively of viral origin was conducted. As a result, the origin and validity of the sequence cannot be verified with certainty. Moreover, the RKI omitted critical validation steps required to ensure the structural integrity of the sequence as an independent single-stranded RNA genome of the SARS-CoV-2 virus.

Opinion

The methodological ambiguities raise questions about the origin of the genetic material as well as the clinical significance and validity of the identified sequences. It remains unclear whether these sequences are truly specific to the presumed pathogen or whether they may partially originate from other material present in the host or the environment. Such gaps could fuel speculation about the objectives of the RKI’s genomic activities.

Links

• German FOIA (English)

• Original FOIA (German)

General Remarks & Fallacies by Wilf

Wilf begins his essay by asserting that “vaccines work,” drawing a comparison to similar products from the past. However, this argument relies on a logical fallacy: it assumes that the assumed effectiveness (and mechanism) of other vaccines inherently proves the effectiveness of the current ones. This raises a critical question: what if past vaccines also didn’t work as claimed? Furthermore, Wilf argues that the majority of people believing in vaccines makes it unlikely for them to be wrong. Yet, this is another fallacy—an appeal to popularity—offered without supporting evidence.

Later, Wilf remarks that in the past, death signals from vaccination would have been detected. However, concluding that the same would inevitably occur this time is a flawed assumption and represents a logical fallacy.

In addition, he highlights that the mechanism of action causing harm or death would not be unknown. Contrarily, Pfizer explicitly states that ‘the exact immunologic mechanism that confers protection against SARS-CoV-2 is unknown.’ This implies they do not fully understand how their product is even supposed to work.

In contrast, there are at least two plausible mechanisms for harm:

1. Direct cell inflammation and potential cell death caused by lipid nanoparticles (LNP), as described by Ndeupen et al., 2021.

2. The immune system attacks cells that produce the “supposedly foreign” or harmful antigen. This occurs due to the widespread distribution of lipid nanoparticles (LNPs) throughout the body, leading to the destruction of cells in multiple organs, including the heart and brain.

Moreover, vaccine-critical studies have been subjected to significantly more scrutiny, suppression, and even outright banning by journals and certain authors. This double standard raises valid concerns about bias in the academic and scientific discourse surrounding vaccines.

Wilf also overlooks two critical motivations that shape the broader vaccine narrative:

1. Government Objectives: Governments primarily aim to calm the public and maintain order during crises. Tools like masks, tests, and vaccines serve as tangible measures to reassure the public. However, whether these tools are effective is often a secondary concern, as the primary motivation of governments is typically to control the narrative and preserve authority.

2. Pharmaceutical Industry Incentives: The incentives for pharmaceutical companies are clear. With liability protections in place, these companies effectively operate with a blank check, profiting immensely as long as public fear—and thus demand—can be sustained. This is facilitated by significant influence over the media, where the pharmaceutical industry is the major force in advertising spending. Compounding this issue, these companies often conduct their own trials, creating an inherent conflict of interest that undermines trust in the results.

These considerations highlight critical gaps in Wilf’s arguments and underscore the importance of scrutinizing both the evidence and the motivations driving the vaccine narrative.

Establishing the Meaning of Net Benefit

Before addressing and debunking each of the points raised by Wilf, it’s important to first revisit the objectives of this debate and clarify what constitutes valid proof in this context. The legal framework how cause can only be established by prospective double-blinded, placebo RCT’s - and not by con-founded observational studies - was recently laid out by vaccine attorney Aaron Siri on Tucker Carlson’s podcast.

The term Net Benefit implies that the benefits must significantly outweigh the harms, meaning notably more lives are saved than harmed.

There are three main types of evidence to consider to answer the question of net-benefit:

1. Studies that prove causality: These include prospective, double-blind, placebo-controlled trials, which are the gold standard for establishing cause and effect.

2. Confirmed reports and harm statistics: Documented cases and data that provide evidence of adverse outcomes.

3. Post-marketing studies: These studies assess real-world data but can only indicate correlation and cannot definitively establish causation.

So what are the results of the above when it comes to the COVID-19 vaccines:

1. COVID-19 Vaccine Results from Manufacturer RCTs

Efficacy

Wilf entirely overlooks the fact that the original RCTs achieved the “95%” efficacy label through the use of several methodological maneuvers. These trials had significant issues, including:

1. Non-representative study population

2. Testing discouraged <7 days after vaccination

3. Lack of relevant clinical endpoints

4. Early unblinding

5. High dropout rates

6. Possible unblinding via PCR tag

7. Unknown mechanism of immunity

8. Neutralizing antibody timeline mismatch

A detailed article including sources can be found here.

Wilf correctly points out that the studies show no sign. death signal. However, it is important to note that the study population was exceptionally healthy, as individuals at high risk for severe COVID-19 were explicitly excluded. This raises a significant concern: How can a study, primarily designed to evaluate protection against severe illness and death, demonstrate any meaningful effect when the very people most likely to benefit were excluded from participation?

The endpoints used in the studies were inadequate for establishing clinically meaningful efficacy. The only truly relevant endpoint was all-cause morbidity and mortality, yet only mortality was reported, leaving a significant gap in assessing the vaccine’s overall impact. Here’s an overview of the prospective, double-blind, placebo-controlled trials that were conducted for vaccines later administered in the West:

• mRNA Vaccines:

  • All-Cause Mortality: No benefit (RR = 1.03)

  • COVID-19 Mortality: Reduced (RR = 0.4)

  • Non-COVID Mortality: Elevated (RR = 1.16)

• Adenovector Vaccines:

  • All-Cause Mortality: Significant benefit (RR = 0.38).

  • COVID-19 Mortality: Strong reduction (RR = 0.18).

  • Non-COVID Mortality: Unexpected reduction (RR = 0.5), raising concerns of imbalance or trial irregularities.

Evident from the above data, the mRNA vaccines, the most commonly used in the West, show no benefit for all-cause mortality (RR = 1.03) and may increase non-COVID deaths (RR = 1.16). Adenovector vaccines report suspiciously large reductions in non-COVID mortality, raising concerns about potential trial issues. This has been confirmed by a peer reviewed article by Benn et al., 2023.

Notably, all trials were unblinded early, with observation periods lasting less than 3 months (e.g., Pfizer). This limited timeframe is inadequate for detecting medium- to long-term side effects, making it impossible to fully assess potential harms. While follow-up periods were longer, the early unblinding disrupted the balance between groups, rendering proper comparison of all-cause mortality unreliable.

Conclusion: None of the RCTs convincingly demonstrated lives saved, leaving the potential for harm unresolved due to the much healthier population and short observation.

2. Confirmed reports and harm

There are numerous documented reports and signals that highlight potential harms. Here is a brief overview of the evidence. Any such documentation is sufficient to establish that harm is possible:

Individual Case Reports

Examples such as Roy Butler, Natalie Boyce, who both passed away several weeks after vaccination, underscore potential risks. Numerous similar reports have been documented and are identifiable in public records.

Myocarditis Autopsy Study

Among 25 individuals who died unexpectedly within 20 days of vaccination, autopsies identified myocarditis as the likely cause of death in five cases. Additional studies have examined other adverse effects, such as thrombosis and other potentially fatal side effects.

VAERS Data

The U.S. Vaccine Adverse Event Reporting System (VAERS) recorded a significant signal of adverse events following vaccine rollout. While these reports do not establish causation, the sheer magnitude of the signal provides a strong indication of potential risks that warrant further investigation.

Official German Cause of Death Statistics

The German Statistical Office in Germany has recorded 316 deaths under the underlying cause of death category for T88.1 “Other complications following immunization, not elsewhere classified” in 2021. This number was 255 in 2022, 47 in 2023, and a total of 316 over the past four years. Notably, there were no deaths recorded in this category in the previous several years.

The Statistics Office has confirmed that at least a significant portion of these cases can be directly attributed to COVID-19 vaccination. A more detailed FOIA request by me is currently pending.

Interim Conclusion

As you can see, section 1, the original randomized controlled trials, did not demonstrate any overall mortality benefit; section 2, harms, clearly shows evidence of several hundred deaths in Germany alone (1 in ~200k vaccinated); a rough global estimate (based on the confirmed German mortality incidence) brings the number of deaths to over 27,500 (5.65b vaccinated).

3. Observational Studies

This brings us to the third aspect: the observational studies, which Wilf frequently cites. These studies must clearly and unequivocally demonstrate a substantial positive effect to offset the observed net harm. However, from a technical standpoint, the debate should already be resolved, as observational studies cannot establish causality. Despite this limitation, let us take a closer look and dissect them further:

Analysis of the entire Hungarian population

Wilf mentions a study from Hungary, examining all-cause mortality over a brief 4-month period, only two of which experienced excess mortality. It is is so fundamentally flawed that the authors withheld both the underlying data and their calculations, and stopped responding to several inquiries:

• Exclusion of partially vaccinated adults: UK data shows highest mortality in this group.

• Health-related variability: Structural indicators vary across vaccination groups, confounded by general health status.

• Moderna lower survival probability: Highlighted in Figure 2.

• Short observation period: Only 4 months (Apr-Aug 2021), highest mortality observed in partially vaccinated.

• Person-year miscalculation: Potential errors if vaccination date reporting is inaccurate.

Conclusion: The study fails to provide valid evidence of vaccine efficacy due to its very short observation period (4 months), significant confounding, and incomplete cohort comparisons (e.g., ever vaccinated vs. unvaccinated). Similar to the recently published Norwegian study, it cannot demonstrate that mortality rates are indeed equal to or lower in the vaccinated cohort compared to those without the intervention.

Comparing ACM across countries and time

Wilf asserts that vaccination and COVID-19 incidence are the key determinants of excess deaths. However, vaccination has not demonstrated any positive impact on the presumed spread or incidence of the virus to begin with, as randomized controlled trials (RCTs) have not shown that vaccines have saved any lives. Moreover, COVID-19 data is neither reliable nor comparable, as it was not collected in a controlled manner from a statistically representative population sample.

When comparing all-cause mortality across countries, it becomes evident that while vaccines appeared to have an impact in 2021, several other confounders are the primary drivers of excess deaths.

For instance, after adjusting mortality rates for the confounding variable “extreme poverty,” any positive correlation with vaccination disappears entirely.

The significant impact of poverty on mortality has been well-documented, as highlighted by Ioannidis et al. (2023): “Excess deaths (as a proportion of expected deaths, p%) were inversely correlated with per capita GDP (r = -0.60) and positively correlated with the proportion living in poverty (r = 0.66).”

ED of Republicans and Democrats

The study’s findings may highlight real disparities in excess mortality, but its several major methodological issues severely limit the validity of its conclusions. The relevance of political affiliation is likely overstated or arbitrary. Instead, the differences likely stem from:

1. Movement During the Pandemic: Population movement, especially starting in mid 2020, after the initial lockdowns, may have severely affected the AI-based name-matching algorithm used to link voter registration and mortality records. This could lead to significant errors in matching mortality to the individuals.

2. Limited Timeframe: The analysis only extends until 2021. To robustly attribute the observed effects to vaccination, data beyond 2022 is essential. If mortality rates converge after 2021 back to no differences between R/D voters, it would indicate vaccination effectiveness; if not, it could suggest confounding factors or other effects.

3. Societal and Long-Term Factors: Stress, lockdown-related effects, economic hardship, drug overdoses, and other long-term societal impacts cannot be ruled out as contributors to excess mortality, disproportionally affecting R voters.

4. State-Level Discrepancies: The presumed effect is much less visible in Florida but more pronounced in Ohio, suggesting the observed differences may result from factors other than vaccination or party affiliation.

5. Age Adjustment Issues: The study only adjusts for broad age bins. Republican voters, being older on average, may exhibit higher relative effects of mortality impact, which could bias the results.

6. Short Baseline Fit: The baseline is based on a two-year fit (2018–2019) using a Poisson regression model at the county-by-party-by-age level. This short period risks producing artifacts, especially given seasonal or structural trends in mortality. Additionally, no confidence intervals (CIs) are provided for these baseline estimates. This fact alone should immediately dismiss the validity of the study.

7. No Adjustment for Sex or Race: The study does not account for sex or race, both of which are critical factors in mortality and could confound the observed results.

8. Timing of Divergence: The divergence in mortality begins much earlier than the smoothed line suggests, around mid-2020. This timing predates vaccine availability, suggesting that non-vaccine factors played a significant role.

9. Data and Code Not Publicly Available: The lack of public access to the dataset and analysis code prevents independent verification and replication of the study’s findings, reducing transparency and credibility.

The study’s findings are significantly undermined by its methodological flaws. The observed differences in excess mortality, attributed to political affiliation, are likely a proxy for other structural and behavioral factors, such as rural vs. urban residence, socioeconomic disparities, healthcare access, and vaccination behavior. A more robust analysis would require individual vaccination data, adjustments for key demographics, and a longer timeframe to draw valid conclusions. The focus on political affiliation appears overstated and overly politicized.

EDs are Covid

Wilf claims that all excess deaths in the US are caused by COVID-19 but overlooks critical nuances. Several regions have shown no significant excess mortality, even as the US recorded double-digit excess rates, bringing into question the underlying health emergency to begin with. Additionally, Wilf misunderstands how ICD-10 codes for COVID-19 are determined. There is a straightforward explanation for why COVID-19 deaths often align with excess deaths — but this phenomenon is primarily observed in western and wealthy countries.

1. COVID-19 Death Classification in Western Countries

   Most western countries have incentivized the classification of seasonal respiratory illnesses as COVID-19. Additionally, many apply their own death certificate modeling systems, such as the CDC’s NVSS/MMDS. Source

2. Origin & Clinical Relevance of Found Sequence

   The origin & clinical significance of the identified sequence remains unknown to this day. Source

3. PCR Primer Similarity to the Human Genome

   The PCR primers used for COVID-19 testing have a high similarity to the human genome. Source

4. Limitations of Wastewater Surveillance

   Even wastewater surveillance cannot substantiate claims that COVID-19 was a novel virus or provide precise assessments of viral levels. The method involves pooling genetic material from multiple strains and individuals, complicating attribution. Furthermore, no pre-2020 data exists to validate these techniques as a reliable control. Source

5. Clinical Validation of COVID-19 PCR Tests

   The COVID-19 PCR test has never undergone clinical validation to demonstrate specificity or predictive accuracy for diagnosing COVID-19. In hospitals, many positive results were incidental, meaning individuals tested positive but were not ill with respiratory conditions like COVID-19. Source

All of this raises significant doubts about whether the excess deaths can genuinely be attributed to an assumed novel pathogen.

ED in countries with Zero Covid

Wilf asserts that several countries were able to “stop COVID-19 for a long time,” but this claim does not hold up to scrutiny. No mitigation measures have been scientifically proven to significantly halt the spread of COVID-19 or any respiratory disease.

The countries cited are predominantly island nations, which may have created the appearance of controlling the virus through border closures and quarantine measures. However, scientific evidence does not conclusively support the effectiveness of these strategies in completely stopping viral spread. Instead, the observed outcomes may have simply sustained an illusion of control within the population.

Notably, after reopening, many of these highly vaccinated countries experienced significant excess mortality over multiple years. For example, Singapore, with a vaccination rate of 94%, recorded substantial excess mortality, especially in the elderly population—but only starting in 2021.

Similarly, Taiwan, another highly vaccinated nation, experienced a comparable trend.

These figures challenge the assumption of a highly effective vaccine. In contrast, some countries with lower vaccination rates—such as Luxembourg, Denmark, and Germany—reported little to no notable excess mortality. A full ranking can be found here.

Wilf should investigate whether these deaths could be linked to mechanisms such as Antibody-Dependent Enhancement (ADE) or other vaccine-related factors that may exacerbate disease outcomes. Alternatively, he should assess whether these excess deaths are genuinely attributable to SARS-CoV-2 or if other contributing factors better explain the observed trends. Data from the Mortality Watch Excess Ranking, based on age-standardized deaths and a conservative 3-year pre-pandemic baseline, reveals at least two dozen countries that did not exhibit statistically significant excess mortality during the 2020–2023 period. These findings challenge the notion of a universally impactful pandemic and warrant further scrutiny into regional disparities and underlying causes.

Saved by the Vaccine

In-Vitro Studies

Wilf’s argument regarding “neutralizing antibodies,” which are demonstrated in vitro to neutralize the virus, is fundamentally flawed. Antibodies are not inherently specific to this virus, and their levels are merely determined by concentration (titers). In fact, most antibodies may be found in small quantities across all individuals, rendering their purported specificity questionable.

Moreover, no studies have conclusively shown a correlation between antibody levels and improved clinical outcomes. As such, the argument lacks empirical support and is ultimately moot.

Israeli data

The presented data is insufficient to establish any meaningful conclusions about vaccine efficacy, primarily due to the use of non-comparable cohorts and the presence of confounding factors. Even in the provided charts, there is no clear effect over time; instead, the lines remain consistently different. This pattern strongly suggests that the data is confounded, making it unreliable for drawing causal inferences about vaccine impact.

Effect on cases and CFR

Any calculation involving COVID-19 data is inherently flawed, as it does not originate from a statistically representative population sample. Furthermore, the data is highly confounded by factors such as the number of tests conducted, testing policies, and variations in health status among populations. As a result, these calculations cannot be considered reliable evidence for determining efficacy.

Covid Deaths by Vaccination Status

Wilf prominently references U.S. data on COVID-19 deaths by vaccination status, as presented on OWID. However, this data is fundamentally flawed, as it is not publicly available in its entirety. The data presented only includes mortality rates, without providing the actual death or population figures used for calculations, making it impossible to verify or properly analyze.

In summary, the chart suffers from at least three major issues:

1. Use of unreliable data

   The analysis relies on biased, uncontrolled, and unvalidated COVID-19 testing data, rather than objective, gold-standard metrics like all-cause mortality.

2. Exclusion of relevant data

   Vaccinated individuals are only considered “fully vaccinated” 14 days after their second dose. This approach omits at least five critical weeks of data, including the period immediately following vaccination, which could influence outcomes.

3. Misclassification of vaccination status

   The “unvaccinated” group likely includes a significant number of vaccinated individuals whose vaccination status could not be matched or verified. For example, in Germany, approximately 80% of individuals’ vaccination statuses are unknown, leading to severe misclassification.

These issues are not unique to U.S. data; similar problems exist in datasets from other countries. On the other hand, data from New Zealand provides a contrasting perspective. It shows that mortality rates for vaccinated and unvaccinated groups are similar, and the country experienced no statistically significant excess mortality during the pandemic. This, along with Sweden’s experience, serves as evidence that there was no widespread health crisis requiring extraordinary measures to begin with.

Covid Deaths by Vaccination Rates

The same fundamental issues outlined above apply here: COVID-19 data is inherently unreliable, and all-cause mortality metrics are heavily confounded. After accounting for these confounding factors through proper deconfounding methods, it is evident that no significant vaccination effect remains.

Calculating Lives Saved

Thus, the entire calculation for “lives saved” can be outright dismissed, as it relies on flawed models. These models, for example, assume a constant Case Fatality Rate (CFR) without considering whether the CFR is accurately measured in the first place. This methodology suffers from the same issues I outlined in my article “Did the vaccine save 20 million lives?”

Summary

Randomized Controlled Trials (RCTs) provided no evidence of efficacy for the general healthy population—no lives saved and no data reported on clinically relevant morbidity. Furthermore, these trials were not designed to adequately assess safety for older, frail individuals or those at higher risk, on top of having several methodological issues that may even have been designed intentionally to achieve the intended outcome (“the illusion of efficacy, through not clinically relevant data”).

Hard evidence, including ICD-10 coded vaccine-related deaths, confirmed autopsies, and detailed case reports, clearly demonstrates that these vaccines can cause harm and even death.

Observational studies, despite their volume, are riddled with methodological flaws that render their conclusions unreliable. Issues such as insufficient evidence, methodological weaknesses, and pervasive confounding are pervasive. Additionally, societal and lockdown measures disproportionately impacted unvaccinated individuals, especially from mid-2021, when significant societal pressure and mental stress were placed on them. Social exclusion and related stress led to further disruptions, including relocation (skewing population data/denominators), changes in life routines (e.g., commuting, eating habits), and increased drug use, complicating the interpretation of these studies even further.

These considerations underscore that there is no solid basis to claim that the net effect of vaccines can ever be shown to be causally positive. The belief that “vaccines work” appears to remain a matter of faith rather than scientific evidence. 😉

I disagree with these findings because they lack concrete evidence and sufficient methodology.

Here are my detailed comments:

1. Healthy User Bias

The analysis suggests that the lower mortality rate in vaccinated cohorts, as evidenced by the hazard ratio (IRR), may indicate a causal relationship with COVID-19 vaccination. However, the observed lower IRR in individuals with 3+ doses compared to those with 1-2 doses strongly suggests the influence of healthy user bias—a self-selection effect where health-conscious individuals are more likely to get vaccinated and maintain healthier lifestyles overall.

2. Lack of a Time Series Analysis

The study would have benefited greatly from presenting time series data, such as weekly or monthly mortality rates by vaccine dose. A temporal view would provide insights into how mortality rates fluctuate over time in relation to vaccination campaigns, pandemic waves, and other external factors. This is particularly important for understanding potential lagged effects or temporal biases.

3. Crude Age-Banding

The use of broad age groups limits the granularity of the analysis and may obscure differences in mortality risk across finer age brackets. Incorporating more detailed age stratification could enhance the precision and interpretability of the results. A blended age-standardized rate was not provided.

4. Misleading Comparisons

In the introduction, the study states: “Most countries worldwide have experienced excess mortality that coincides temporally with the COVID-19 mass vaccination campaigns. This has led to speculation on the potential long-term effects of the vaccines on mortality risk.” While this is an important concern, the analysis does not provide sufficient evidence to determine whether vaccinated individuals experienced lower (or higher) mortality compared to what they would have experienced with a placebo or no vaccination. (The methodology also does not address the temporal correlation; lack of time series analysis) The comparison of vaccinated versus unvaccinated individuals alone cannot fully disentangle the effect of vaccination from other underlying factors, such as baseline health differences or confounders.

5. Contextual Trends in Norway

The analysis could have benefited from situating the findings within broader national mortality trends. For example, data from Norway show no statistically significant excess mortality compared to the 2017-2019 baseline, as per Mortality Watch. Excess mortality remained within the 95% prediction intervals (PI) during the pandemic years:

• 2020: -4.1% [-8.7%, 0.4%]

• 2021: -2.9% [-7.4%, 1.7%]

• 2022: 3.4% [-1.1%, 8.0%]

• 2023: -3.3% [-7.8%, 1.3%]

These data highlight the importance of contextualizing results against national trends. For Norway, this suggests that mortality patterns during the pandemic largely remained within expected ranges, making causal claims about vaccination’s impact doubtful. Moreover, Norway did not appear to experience any significant excess mortality during the peak of the pandemic. The slight increase in 2022 might simply be a result of a “catch-up” effect or a lingering negative impact of the pandemic response.

6. Alternative Methodology

The authors adjusted for several confounding variables, including sex, calendar time, county of residence, and risk group status. However, an alternative approach could involve historical matching or synthetic control. By comparing vaccinated individuals’ mortality rates to historical mortality rates of individuals with identical attributes (e.g., matched on sex, age, risk group, and residence), they could have established a counterfactual baseline. This methodology would help mitigate biases and provide a more robust assessment of vaccination’s impact on mortality. Conclusion While the study addresses a crucial research question, I believe it lacks sufficient evidence to establish a net positive effect of vaccination or its safety. The absence of a time series analysis further hinders the validity and impact of the findings. To address these methodological and interpretational gaps, I suggest incorporating a time series analysis with a historically matched baseline.

7. Sweden vs. Norway

Epidemiological evidence from comparing Sweden and Norway directly contradicts the notion that these vaccines are both “safe and effective”. Despite having a higher vaccination rate, Norway experienced higher excess mortality compared to Sweden, in 2022, after vaccination!

8. Lack of Transparency in Model Calculations

The calculations used in the model are not publicly disclosed. This especially leaves the exact method for determining rates unclear. For instance, it is not specified how the rate for the unvaccinated group was calculated. If the unvaccinated population (denominator) was derived by subtracting the known vaccinated population from the total population, this residual group might not accurately represent the unvaccinated population. Such an approach introduces a risk of misclassification, where unconfirmed or unreported vaccinated individuals could be mistakenly included in the unvaccinated group. This misclassification risk is one-sided, potentially leading to over- or underrepresentation of the unvaccinated cohort. Additionally, no confidence intervals (CI’s) were provided to account for this uncertainty, further limiting the reliability of the conclusions.

9. Absence of Stratification by COVID-19 and Non-COVID-19 Mortality

The data were not stratified to differentiate between COVID-19-related and non-COVID-19-related mortality. Such stratification is essential for a more rigorous estimation of the potential effects of COVID-19 vaccination and to address other issues, such as attribution bias and possible over- or undercounting of deaths. By isolating non-COVID-19 mortality, it becomes possible to better assess confounding factors that primarily affect this category. Additionally, examining changes over time could offer valuable insights into whether vaccination has a protective or harmful effect, while also highlighting trends that might otherwise remain obscured.

10. Exclusion of Data for Cross-Vaccinated Individuals

The dataset completely removed (“censored”) cross-vaccinated individuals without providing a clear rationale. This lack of explanation raises questions about the potential impact of this exclusion on the results, and whether it could introduce bias or reduce the generalizability of the findings.

11. Population Data of Study Appears 7% Higher than Adult Population from Stats Norway Indicates!

The population data in the study appears to be incomplete or inconsistent.

The study reports: “The study included 4,645,910 individuals (49.8% women) with 132,963 deaths during follow-up.”

DEATHS:

I couldn't find deaths by single year ages at Stats Norway; but I found it at Eurostat, however 2023 is not posted, thus I simply doubled 2022 (to account for 2023):

Thus reported deaths for 18+ sum to 41,781 + 45,570 + 45,570 = 132,921 which is roughly in line with what the study reports.

POPULATION:

Similarly, the mean population 18+ calculated as (4,279,679+4,316,747+4,376,793)/3 = 4,324,406 from Stats Norway is 6.9% lower than what the study reports.

It is possible that the 7% difference may be explained by "lost to follow up" (death, migration, etc.) - however that seems quite high, and it remains unknown if this disproportionately affects the unvaccinated cohort in the rates calculations.

The discrepancy in the population figures are unexplained and raise significant concerns about the accuracy and reliability of the study’s findings.

Code

12. Conclusion

The authors’ conclusion, “In conclusion, there was a reduced rate of death among individuals vaccinated with mRNA vaccines during 2021-2023 in Norway,” is clearly flawed. This conclusion was drawn by comparing the mortality rates of vaccinated individuals to those of the unvaccinated cohort. However, as outlined in the 10 points above, this approach is severely flawed and lacks proper scientific rigor. Such a conclusion cannot be legitimately inferred.

What the study actually demonstrates is that the mortality rate in the self-selected vaccinated cohort appears lower, primarily due to confounding factors. The unvaccinated cohort represents the remainder of the population, including individuals at the end of life receiving palliative care, who were unlikely to receive vaccination.

A proper comparison would require establishing a baseline mortality rate for the vaccinated cohort using historical, cohort-specific data that accounts for confounders. Only by comparing the observed mortality rate against this true baseline could one accurately determine whether the rate among vaccinated individuals is genuinely lower or higher.

Note: 7-12 were added after the initial publication of this article.

I have sent this friendly letter, which includes most of this feedback, to the authors:

Please let me know your thoughts in the comments!

Below is a prime example of why reading the sources and methods sections is crucial. Without doing so, it’s easy to be misled by deceptive pharma propaganda charts like this spread by Our World in Data (OWID).

The underlying CDC dataset explicitly acknowledges that it did not track data from day one, even in cases where vaccination status could potentially be matched—though it is doubtful they were able to do so for the majority of deaths. This is primarily because the CDC lacks a comprehensive registry linking individuals to their vaccination status by name or social security number, making it challenging to accurately capture data across large segments of the population. Furthermore, it is worth noting that Our World in Data (OWID) receives funding from organizations such as the Bill & Melinda Gates Foundation and pharmaceutical companies, which raises questions about the objectivity of the narratives they promote.

So in summary, the chart contains at least three major issues:

1. Use of unreliable data – It relies on biased, uncontrolled & unvalidated COVID-19 testing data rather than, objective gold-standard all-cause mortality.

2. Exclusion of relevant data – Individuals are only considered vaccinated 14 days after their second dose, omitting at least five weeks of critical data.

3. Misclassification of vaccination status – The “unvaccinated” group likely includes a significant number of vaccinated individuals whose status could not be matched or verified. (In Germany, ~80% vaccination status is unknown!)

In contrast, gold-standard age-standardized all-cause mortality data obtained from Health NZ through a FOIA request shows that unvaccinated individuals did not have a higher all-cause mortality risk than vaccinated people.

2) Correlations of Vaccine doses administered vs Excess Deaths

Top: Simple correlation of excess deaths vs vaccinations, shows a spurious correlation.

Bottom: Correctly adjusted for confounders like age and poverty, shows no (positive) vaccination effect!

The observed correlation in 2021 is driven by confounding factors, such as extreme poverty and cardiovascular death rates, making it a spurious correlation. Once mortality rates are properly adjusted for these confounders, the correlation disappears entirely.

Read more here:

3) UK ONS All-Cause Mortality data by Vaccination Status

1. The perceived benefit of vaccination over non-vaccination can simply be explained by examining non-COVID-19 mortality or focusing on a period when COVID-19 had minimal impact—for instance, May 2021, when COVID-19 accounted for only 1.4% of all deaths. This analysis reveals that overall mortality was higher among the unvaccinated, not due to COVID-19, but as a result of other confounding factors, such as poverty, preexisting health conditions, and socioeconomic disparities.

2. Fenton & Neil highlighted that the initial spike in mortality was merely an artifact caused by delays in reporting vaccination status. In response, the ONS removed the problematic data, effectively acknowledging that their initial analysis contained flaws.

Once adjusted by the confounding effects, this spurious effect also completely disappears.

Summary

All three charts can be attributed to flawed methodologies—such as failing to match entire cohorts, not starting observations from day one, and misclassification—or to confounding factors like underlying health conditions, the impact of lockdowns and socioeconomic factors (with unvaccinated individuals often excluded), poverty, and more.

As a result, I have yet to find credible evidence that these vaccines have saved a single life. On the contrary, we already have confirmed cases of severe harm, with thousands injured or killed, even acknowledge official by governments!

Please don’t forget to like & leave a comment about your thoughts. Most importantly, if you enjoy my work & want to help, please share & consider upgrading to a paid subscription!

Background

In 2024, many countries have reported detailed all-cause mortality data by age, allowing for a comprehensive retrospective analysis of the impact of lockdowns and vaccinations. By utilizing Age-Standardized Mortality Rates (ASMR), we can better assess how government interventions, such as lockdowns and vaccination campaigns, correlate with mortality outcomes during the COVID-19 pandemic, independent of age—a major confounding factor. This approach provides clearer insights into how these measures influenced public health across different populations.

Data & Methods

For the mortality data, I used age-standardized values provided by Mortality.Watch and applied a simple 3-year pre-pandemic average. While there are various methods to estimate excess mortality, as Levitt et al. have argued, there is no universal standard for extrapolating mortality rates from trends. Using trends, as often done, may “result in generating an unrealistically low number of expected deaths.”

To address this, I applied a straightforward 3-year average of the 2017-2019 age-standardized mortality rate, following the cautious, conservative methodology used by Levitt et al.

The government response was measured using the Oxford Government Stringency Index, which quantifies the strength of government interventions on a daily scale from 0% to 100%. For this analysis, I aggregated this data annually.

COVID-19 vaccination was measured as the percentage of the population that received at least one dose. While further analysis based on the number of doses is possible, it is important to acknowledge the influence of confounding factors, such as the “healthy user” and placebo effects. Individuals who are more health-conscious and confident in vaccines may be likely to receive more doses, while those who are skeptical or in poorer health may receive fewer or none. If the vaccine’s effectiveness is as high as reported (e.g., “95%”), we should observe a significant impact even after a single dose. Notably, the difference in coverage & efficacy between one and two doses was often marginal.

Possible confounding variables were included from the Our World in Data (OWID) dataset & the CIA World Factbook.

Results

Government Stringency

When analyzing the correlation between government stringency and excess mortality, a significant relationship is observed in 2020 and 2022. However, this correlation disappears in 2021, likely due to the higher overall stringency levels during that year.

When grouping the data by levels of government intervention for 2020, we find that the group with ‘low’ intervention group also experienced the lowest excess mortality.

This suggests a potential relationship where higher levels of government intervention were associated with higher excess mortality.

An often-cited argument—that governments responded because COVID-19 levels were higher—is flawed. A novel pathogen, alleged to cause COVID-19, should theoretically spread equally across all populations.

COVID-19 Vaccination

The correlation plots between vaccination coverage and excess mortality reveal a distinct pattern in 2021 - it appears as if the vaccine significantly reduces excess mortality. However, this correlation disappears in 2022 and 2023. But let’s take a closer look…

Naturally, I remained skeptical and suspected potential confounding, as previously highlighted by Ulf Lorre, given that this is high-level population data. To address this, I included 17 additional variables and plotted them against excess mortality.

Highlighted are all Pearson correlations above (or below) ±0.4, indicating at least a moderate relationship. The year 2021 is heavily influenced by at least six confounding variables: Tobacco/Smoking, Life Expectancy, HDI, GDP, Poverty, and Cardiovascular Deaths.​

Then I adjusted the excess mortality rates to account for the impact of the confounders. Below is an overview of the adjusted excess mortality rates for the past three years, considering six moderate or higher confounders.

And voila, the correlation observed in 2021 completely disappears, revealing itself as a spurious relationship!

This is how the adjusted chart looks like when we just adjust by Extreme Poverty:

And here adjusted by Cardiovascular Deaths:

Conveniently OWID has a map of Cardiovascular deaths, which confirms there’s a steep slope between rich and developing countries.

OWID summarizes what causes these kinds of deaths: “Cardiovascular diseases encompass all conditions affecting the heart and blood vessels, including heart attacks, strokes, atherosclerosis, ischemic heart disease, hypertension, cardiomyopathy, rheumatic heart disease, and more. These conditions typically develop gradually with age, particularly in individuals with risk factors like high blood pressure, smoking, alcohol use, poor diet, and air pollution.”

This suggests that excess mortality was largely driven by the impacts of lockdowns, which exacerbated existing health & environmental issues, such as cardiovascular diseases linked to unhealthy lifestyles, drugs & alcohol, and possibly pollution.

Summary

The analysis of government stringency and excess mortality reveals significant correlations in 2020 and 2022, which weaken in 2021. This change is likely linked to widespread lockdowns, which led to increased excess mortality across countries. Grouping the data by levels of government intervention shows that countries with the lowest levels of intervention experienced the least excess mortality, suggesting a potential association where stricter interventions corresponded with higher excess mortality.

When examining the relationship between COVID-19 vaccination and excess mortality, a noticeable correlation appears in 2021 but fades in 2022 and 2023. To account for potential confounding, 17 variables were included in the analysis, such as tobacco use, life expectancy, GDP, extreme poverty, and cardiovascular death rates. After adjusting for these strongest confounders, the previously observed correlation in 2021 disappears completely, indicating that the initial relationship may have been driven by these confounding factors.

Contrary to prior assumptions, stringent government interventions may have contributed to higher excess mortality, and vaccination did not show the anticipated reduction in all-cause excess mortality. There remains a possibility that vaccines may have increased mortality due to known side effects, including heart and neurological issues. Countries with minimal interventions appeared to have lower excess mortality from 2020 to 2023 compared to those with stricter measures. Vaccination did not show a clear positive impact on all-cause mortality.

Code & Data

• Github: 1, 2, 3, 4

• OWID

Hence, I’ve plotted the returns of the index for each presidential term.

While the above chart is pretty noisy, grouping it by party, we can see that there’s a pretty big lead for Democratic presidents, however it is not statistical significant.

Since party policy might have changed over time, let’s just compare the last two presidents, against the mean of all:

Both, Biden & Trump performed very similar. Here’s a list of all presidents and their annualized stock market performance (taking into account the different presidential terms):

There’s no clear significant pattern, that we can conclude.

Moreover, this model does not take into account the starting valuation, in relationship to e.g. the baseline - so more room for future improvements here - but let me know what you think in the comments :)

Code: 1 2

Singapore is 94% vaccinated against COVID-19 and saw massive excess mortality, but only starting in 2021. Mortality.Watch does not have age stratified data available, thus no age standardized rate can be calculated. However, a reader pointed me to the official government page, where I have found official age stratified mortality rates from the Singaporean government.

The impact is most significant among older age groups, particularly those aged 60 and above, with the 80+ demographic being especially affected.

Overall, Singapore has observed an excess mortality rate ranging between 12-20%. In certain age groups, excess mortality has exceeded 30%.

The monthly data is also quite telling - it looks almost if someone had turned on a switch and increased the mortality level by ~20%:

If it’s not the vaccines, as some still claim (“look at Sweden”) what is it, then? Masks?

I have reached out to Singapore’s MoH for further insights.

Sources & Calculations:

• Github

• Mortality.Watch: 1, 2, 3

• Our World In Data

…, in which I have asked for records related to these four points:

1. Records on single virion sequencing of SARS-CoV-2 that ensured the virion was physically isolated from any other genetic material before sequencing.

2. Records of a single sequencing (long-)read from the first position [..] to the last position [..] of the genome. [..]

3. Records of a perfectly assembled contig [..] by a metagenomic assembler like Megahit (as used by Wu et al, 2020) [..].

4. Records of RACE (Rapid Amplification of cDNA Ends) processes carried out on a patient sample or viral culture [..]

The CDC stated that they lack the capability to isolate a single virion from all other genetic material present in a sample, and that they don’t have records that this has ever been performed or even is possible.

They also indicated that they do not possess or have knowledge of any single read or complete assembly that would mathematically confirm the complete genome of a virion.

Furthermore, the CDC relied exclusively on publicly available sequence data and did not employ the RACE (Rapid Amplification of cDNA Ends) technique to verify the manually identified genome ends by Chinese researchers. Verification of these ends is crucial, as the presence of additional bases beyond the first and last positions of the reference genome would suggest that RACE amplification does not capture the true terminal sequences. Although RACE amplicon data for SARS-CoV-2 has not been published, findings from the same researchers who studied the RaTG13 virus suggest that this may be the case, highlighting the need for independent validation.

Long-read sequencing has not yet convincingly demonstrated the complete sequence of the genome, as errors and missing sequence ends prevent the generation of a high-confidence genomic model.

The De Bruijn graph approach to short-read assembly may generate multiple genome candidates by combining various segments of genetic material. This raises the possibility that the SARS-CoV-2 genome could be, a semi-random construct, formed through the recombination of genetic sequences using this methodology.

This hypothesis remains unchallenged, as the reference genome has never been fully sequenced from start to finish, nor assembled using the overlap-layout-consensus (OLC) method which could possibly offer more confidence in the reference genome.

You can try out how the De Bruijn graph approach works, on my web tool at https://debruijn.mortality.watch/

X thread: https://x.com/USMortality/status/1825961166667133203

Let me know your thoughts in the comments!

++ Please consider subscribing to support my work! I provide all my work transparent and for free, but any contribution helps! ++

Births

The absolute number of live births occurring in a population within a specified period, usually a year.

Crude Birth Rate (CBR)

This rate expresses the number of live births per 1,000 people in the total population per year.

General Fertility Rate (GFR)

This rate expresses the number of live births per 1,000 women of childbearing age (15-49 years) per year.

All three previous charts show that current levels are similar to the early 2010’s, so nothing too unusual.

Age-Specific Fertility Rate (ASFR)

This rate expresses the number of live births per 1,000 women in a specific age group per year.

Unfortunately, only data from 2009 onwards is available. Here are the individual rates per age:

I don’t see any obvious signals from either COVID or the vaccinations.

It seems that the current situation might just be within the range of previous declines - likely due to social and economic uncertainty. While it appears they panicked unnecessarily and the injections were unnecessary, we can’t observe anything devastating either.

I might be wrong, though. Let me know what you think!

Code can be found here!

🔗 debruijn.mortality.watch

How the Algorithm Works

The De Bruijn graph algorithm is a widely used method for genome assembly, particularly used for handling short reads from sequencing technologies. Here's a step-by-step breakdown of the process:

1. Reads Splitting into k-mers

The short genomic reads are divided into smaller, overlapping sequences of length (k), known as k-mers.

2. Constructing the Graph

Each (k-1)-mer represents a node in the graph. Edges are created between nodes based on the k-mers. Specifically, an edge is formed from one (k-1)-mer to another if there is a k-mer that bridges them (i.e., the suffix of the first (k-1)-mer matches the prefix of the second (k-1)-mer).

3. Traversing the Graph

The graph is traversed to find all possible paths, known as contigs. These paths represent all possible sequences of DNA that can be assembled from the k-mers, but do not necessarily have to exist in nature.

4. Result

The tool shows if the original sequence (i.e. that of a bacterium or virus) was found in the results (contigs) and if it matches the longest found contig.

Input Modes

The tool offers two input modes for flexibility:

1. Reads Mode

Users can define a list of genomic reads. This mode is suitable for working with sequencing data where the target genome is unknown and needs to be assembled from the provided reads.

2. Genome Mode

Users can define a reference genome. Based on this reference, the tool generates reads with added random noise to simulate additional genomic material (bacterial, human, etc.). This mode is useful for testing the assembly process and understanding how other genetic material might impact the assembly.

Data Download

Reads and contigs can be downloaded in FASTA format for further bioinformatic processing.

Overview of my Seven Part Genomics Series:

Summary

This tool provides a powerful way to visualize and understand the genome assembly process and its potential issues using the De Bruijn graph algorithm.

Please let me know your thoughts and feedback!

1. Excess Mortality in the EU during COVID-19

With a three-year average pre-pandemic baseline, Sweden saw the lowest Excess Mortality in the European Union.

2. Mortality in the USA & Sweden since 2000/'01

Age-standardized mortality by midyear seasons (Jul-Jun) shows no significant mortality increase in Sweden, in contrast to the USA.

3. Mortality in Germany since 2001

Yearly age-standardized mortality shows no significant increase in Germany during 2020, in contrast to the following two years of 2021/'22.

4. Excess Deaths in the USA during COVID-19

During the COVID-19 Pandemic, the USA recorded over a million more deaths than projected.

5. Florida vs California during COVID-19

Florida and California had very different public health approaches, but no significant difference in excess mortality.

6. Excess Mortality by Age Group in the USA during COVID-19

Contrary to the risk profile of COVID-19, the USA recorded the highest relative increase in excess mortality in the working age population.

7. Excess Mortality in Scandinavia during COVID-19

While Sweden saw most of its excess mortality during 2020, its two Scandinavian neighbors saw much larger excess mortality later on.

8. North vs South Dakota during COVID-19

Despite North Dakota's strict lockdowns and mask mandates, and South Dakota's lack thereof, both states saw similar COVID-19 outcomes.

9. Excess Mortality Ranking in the EU during COVID-19

With a three-year average pre-pandemic baseline, only 11 of 23 countries recorded statistical significant excess mortality cumulatively during the pandemic years.

All charts were generated using the interactive tool Mortality.Watch where these charts are also featured on the front page now.

Let me know if you have other great charts to add!

If so, please share, title, description & URL in the comments below!