Inadequate support: The analysis underlying the claim made use of several datasets that don’t provide the needed information. Statistics from the U.S. Bureau of Labor Statistics aren’t stratified by vaccination status, making it impossible to know whether absence from work or disabilities occurred preferentially among vaccinated people.
Flawed reasoning: Phinance Technologies relied heavily on correlations to support its claim. However, correlations aren't proof of a causal association.
FULL CLAIM: COVID-19 vaccines caused “26.6 million Injuries” and “1.36 million Disabilities” in 2022 and cost the economy billions of dollars
Deployment of COVID-19 vaccines began in December 2020 in several countries. However, groups opposing COVID-19 vaccination continue to claim that COVID-19 vaccines are dangerous for health and have negatively impacted society, which isn’t supported by the evidence. Iterations of such claims involve a variety of approaches, ranging from misrepresenting adverse event reports to misinterpreting published studies.
A more recent iteration used labor statistics to allege that COVID-19 vaccines caused an estimated 26.6 million injuries and 1.36 million disabilities in the U.S., resulting in an economic toll of almost 150 billion dollars due to lost work time. This claim was made by financier Edward Dowd, who co-founded the company Phinance Technologies. Phinance Technologies published the analysis on which the claim is based. Titled the “Vaccine Damage Project”, the analysis drew on the results of Phinance’s other projects.
Dowd’s claim was picked up by other websites like ZeroHedge, which promoted the claim. Edward Dowd already made unsupported claims regarding COVID-19 vaccines. He is also a board member of the Malone Institute, founded by Robert Malone. Malone is a known spreader of misinformation regarding COVID-19 vaccines and Health Feedback reviewed his claims on repeated occasions.
However, the analysis is flawed and its conclusions about vaccine damage unsubstantiated. In the following sections, we’ll first explain how their analysis has been done and then the main flaws and limitations that invalidate their conclusions on vaccine medical and economic negative outcomes.
What did the analysis do?
Phinance Technologies’ analysis consisted of collecting and estimating vaccine-related data, such as the number of administered vaccine doses and adverse events (AE), and correlating them with labor statistics from the U.S. Bureau of Labor Statistics (BLS) such as absence rates or number of workers with disabilities.
The authors first calculated the AE excess rate in the Pfizer vaccine clinical trial, that is, the additional number of people who experienced AEs in the vaccinated group compared to the placebo group. They then assumed that this represented people suffering from vaccine-related mild injuries that prevented them from working. To back their supposition, they correlated the excess AE rate to the increase in absence rate and lost work time in 2022 compared to 2021 reported by BLS.
Next, Phinance Technologies estimated the cumulative number of serious adverse events (SAE) from the Pfizer and Moderna clinical trials combined, and the cumulative number of severe adverse events (SevAE) from Pfizer’s clinical trial.
Although both SAE and SevAE sound similar, they are actually defined differently. SAEs are adverse events that are life-threatening or can cause permanent disabilities, or congenital anomaly in absence of medical intervention. SevAEs on the other hand, are adverse events with a high degree of intensity but not necessarily life-threatening. For example, a very painful sore arm could be regarded as a SevAE but is not life-threatening, so is not a SAE.
They then correlated these estimated rates of SAE and SevAE to the increase in people with disabilities among the civilian workforce in 2022 as provided by the BLS. They also reported correlations between the cumulative number of administered vaccine doses and the number of disabilities. From that, they concluded that their results “provide[d] a stronger case for establishing a causal relationship between disabilities and the Covid-19 vaccines”.
Finally, Phinance Technologies estimated the economic costs of these negative outcomes based on the total increase in lost work time and disabilities in 2022 compared to the previous years.
The analysis only reported correlations ; correlations alone are insufficient to demonstrate a causal association
As the previous section demonstrates, Phinance Technologies’ observations are purely correlative. The fact that two phenomena occur at the same time, or that variables increase at the same rate, doesn’t necessarily mean that one caused the other. In fact, many variables are correlated by pure chance. In other cases, both variables are actually affected by a third, unaccounted-for, variable, called a confounding factor. Therefore, the authors of the analysis have not established whether COVID-19 vaccination or adverse events following vaccinations are really the cause for workers disabilities or work time loss.
In fact, the manner in which the analysis was performed biased the analysis towards detecting correlation in the absence of causality. This is because it used metrics that, by their very nature, necessarily increased over time. We know from the BLS data, that the number of disabilities in the workforce increased over time. Therefore, any other metric that also increases over time is likely to correlate with the number of disabilities in the population.
The authors compared the increase in disabilities with cumulative metrics: cumulative number of administered doses, SAE or SevAE. These numbers are cumulative by nature, and so can only increase over time. Therefore, these are likely to correlate with the number of disabilities, even without a causal association.
Such an approach is particularly problematic in the case of the SAE and SevAE metrics. Indeed, the authors estimated the number of people who experienced SAE and SevAE following vaccination by multiplying the number of people getting vaccinated each month by an adverse event rate derived from the literature[1,2].
By doing so, they implicitly assumed that SAE and SevAE are permanent conditions. Their approach relied on the assumptions that new people get vaccinated each month; that some will experience SAE or SevAE and will be added to the existing pool of people with SAE or SevAE; and that everyone who developed an SAE or SevAE will never stop suffering from them.
In reality, this isn’t the case. For instance, the SAE data they used included diarrhea, which usually isn’t a permanent medical condition.
In short, Phinance Technologies observed a correlation between datasets that have been built in such a way that the values increase over time. As a result, these metrics are likely to show a correlation with the number of disabilities purely by design. Thus, this data isn’t indicative of a causal association.
In fact, the authors acknowledged that bias several times by writing for example:
“We realise that performing the correlation of cumulative time series is misleading and the R2 should not be taken as an indication of establishing a statistically significant relationship as both time series have autocorrelation.”
However, they immediately disregarded their own caveat by concluding: Our results provide a stronger case for establishing a causal relationship between disabilities and the Covid-19 vaccines. The time series of SAEs that were computed based on the rates estimated during the mRNA clinical trials are shown to be of the same magnitude as the rate of increase in disability rates in the 16-64 Civilian Labor Force.”
In an email to Health Feedback, Phinance Technologies acknowledged that correlations aren’t sufficient to prove that COVID-19 vaccines caused injuries or disabilities, especially when using a cumulative time series as they did. However, this acknowledgement doesn’t alter the fact that their analysis relies solely on correlations to arrive at its conclusions.
In summary, Phinance Technologies conducted their analysis in such a way that observing correlations was likely, without proving any causal association between vaccination, injuries and reduced work productivity.
Phinance Technologies used flawed or inadequate datasets
Another problem in Phinance Technologies’ analysis is that they used datasets that are either flawed or unable to provide the information needed to make conclusions about the relationship between COVID-19 vaccination and the health of the workforce.
First, the BLS didn’t provide the vaccination status of people missing work or with disabilities. Therefore, it is impossible to tell whether vaccinated people are overrepresented among absentees or among people with disabilities. This seriously weakens Phinance Technologies’ conclusions. In an email to Health Feedback, Phinance Technologies acknowledged not having datasets stratified by vaccination status, which would be necessary for a proper analysis.
Second, they estimated the number of “injuries” by calculating an “excess rate of [vaccine-]related adverse events” in vaccinated people compared to unvaccinated people, using data from the Pfizer-BioNTech vaccine clinical trial. By doing so, they equated the excess rate of vaccine-related adverse events to injuries.This means that, according to them, every additional adverse event that occurred among vaccinated people but not in unvaccinated ones was an injury.
Considering all AE as “injuries” is inadequate. In fact, adverse events in the vaccinated group also included what are called “reactogenicity events”. These are benign, expected reactions to the vaccines, such as pain at the injection site, redness, fatigue or fever. They are actually the sign that the vaccine worked and triggered an inflammatory response. Considering them as “injuries” is therefore unjustified.
Furthermore, reactogenicity events mostly occur in the vaccinated group, by definition. Including reactogenicity events in the total number of AEs would necessarily cause the total number of AEs in the vaccinated group to exceed that of the placebo group. In fact, the clinical trial authors warned about that in their paper:
“Reactogenicity events among the participants who were not in the reactogenicity subgroup were reported as adverse events, which resulted in imbalances between the BNT162b2 group and the placebo group with respect to adverse events (30% vs. 14%), related adverse events (24% vs. 6%), and severe adverse events (1.2% vs. 0.7%)”
The consequence is that it inflated the excess rate of vaccine-related AE calculated by Phinance Technologies and led to an overestimation of their alleged “injuries”.
Third, the authors relied on a study with methodological flaws to estimate the number of SAE in the Moderna and Pfizer clinical trials. Health Feedback previously detailed those flaws. Briefly, there were inconsistencies with the kinds of adverse events the paper’s authors chose to include and exclude. For example, the authors included diarrhea, but excluded vomiting; they included hyperglycemia (high blood sugar), but excluded hypoglycemia (low blood sugar).
Also, this study counted the number of SAE, whereas the BLS registered the number of people with disabilities. Given that a person can experience more than one SAE at the same time, the number of SAE is an overestimate of the number of people with SAE and thus cannot be directly compared with the number of people with disabilities, contrary to what Phinance Technologies did. Altogether, these methodological flaws call into question the accuracy of the SAE rate used by Phinance Technologies.
The economic impact calculation isn’t supported by data
The final step of Phinance Technologies analysis consisted in calculating an economic cost of the alleged injuries and disabilities. Here again, some flaws called into question the validity of their conclusions.
First, they calculated the economic cost by multiplying the 2022 gross domestic incomes by the increase in lost worktime or in disabilities in 2022 compared to the previous years. For example, there was a 0.8% increase in lost worktime in 2022 compared to the previous year. Phinance Technologies thus concluded that the cost of their so-called “injuries” amounted to 0.8%*11.238 trillion dollars, that is 89.5 billion dollars. An earlier version of the analysis stated a cost of 79.5 billion dollars, but Phinance Technologies told Health Feedback that this was a mistake that has later been corrected.
Such a calculation means that the authors considered that the entire increase in lost worktime of 2022 would be due to the vaccine-related injuries. Similarly, they considered that all new disabilities registered in 2022 must have been due to vaccination. Yet no evidence supports such a radical assumption.
Many other factors may have caused the increase in work lost time or disabilities. For instance, one of the disabilities registered by the BLS is “serious difficulty walking or climbing stairs” which could be a sequelae of COVID-19 itself.
In fact, the BLS wrote that “the increase in absences during January 2022 partly reflects the increase in COVID-19 cases, along with seasonal illnesses”. In an article from January 2023, The Guardian explained that long COVID was likely responsible for the increase in absence rate in 2022, detailing that “71% of long Covid patients [in New York] who filed for worker’s compensation still had symptoms requiring medical attention or were unable to work completely for at least six months”.
A compilation of data by the Brookings Institution showed that “3 million full-time equivalent workers are out of the labor force due to long COVID, and this without taking into account those who are still working but at low efficiency due to illness”.
Contrary to Phinance Technologies’ claim, research suggested that COVID-19 vaccination was actually linked to an improvement of the economy. Data from the International Labor Organization showed that vaccination was associated with reduced working hours losses.
Using credit card spending as an indication of economic health, economists from the International Monetary Fund observed that “an additional 1 percentage point in initiated vaccination rates increases weekly credit card spending by 0.6 percentage points and reduces initial unemployment claims by 0.2 percentage points of the 2019 labour force”, confirming that vaccination didn’t appear to jeopardize the economy.
Researchers also turned to indirect indicators of economic activity such as gas emissions and mobility data. They found that an “increase in vaccination per capita is associated with a significant increase in economic activity”, suggesting that vaccination was in fact beneficial to economic recovery during the pandemic.
Other studies found that vaccines saved millions of lives. One study concluded that vaccinations prevented 3 million deaths and 18 million hospitalizations in the U.S., saving 1.15 trillion dollars in medical costs. Another study estimated the number of averted deaths in a group of 185 countries at 14 to 19 million.
In summary, claims that vaccination cost billions to the economy due to injuries and disabilities are inaccurate and based on a flawed analysis by Phinance Technologies. The analysis showed no causal association, relying exclusively on detecting correlations. The analysis also used datasets inadequate to their objective and research actually points to COVID-19 vaccination being associated with economic benefits, not the opposite as claimed by Dowd.
UPDATE (17 April 2023):
This review was updated to include Phinance Technologies’ responses to Health Feedback’s request for comment. Their statement that correlation doesn’t equate to causation was added to the twenty-first paragraph. Phinance’s acknowledgement that they didn’t have information regarding vaccination status in their dataset was included in the twenty-fourth paragraph, and their acknowledgement of an error in their estimated economic cost was included in the thirty-third paragraph.
- 1 – Thomas et al. (2021) Safety and Efficacy of the BNT162b2 mRNA Covid-19 Vaccine through 6 Months. The New England Journal of Medicine.
- 2 – Fraiman et al. (2022) Serious adverse events of special interest following mRNA COVID-19 vaccination in randomized trials in adults. Vaccine.
- 3 – Hervé et al. (2019) The how’s and what’s of vaccine reactogenicity. Vaccines.
- 4 – Deb et al. (2022) The effects of COVID-19 vaccines on economic activity. Swiss journal of economics and statistics.
- 5 – Watson et al. (2022) Global impact of the first year of COVID-19 vaccination: a mathematical modelling study. The Lancet infectious diseases.