Are vaccines (really) effective?
Vaccine efficacy, together with vaccine safety, is the foundation of the doctrine of mass vaccination to manage the covid crisis. Yet German and British epidemiological statistics show us that, on analysis, vaccine efficacy decreases sharply over time, particularly in risk groups. Should we then persist on this path by offering booster injections?
The Robert Koch Institut (RKI), the German counterpart of the American Center for Disease Control (CDC) or of Sciensano in Belgium, has been publishing weekly epidemiological reports since the start of the covid “pandemic”. Since this summer, these have also contained the rates of vaccinated people by age group and by type of covid case: symptomatic, hospitalised, hospitalised in an intensive care unit (ICU) and deceased. On the basis of Farrington's screening method, the RKI then uses them as an empirical basis to calculate its estimate of vaccine efficacy. In this article we will play devil's advocate and recalculate a vaccine efficacy estimate for Germany using the RKI's own figures and method. The results are most intriguing, which casts suspicion on Sciensano's censorship of statistics on the vaccination status of covid patients since the start of the high season.
Rise in illness among the vaccinated
A quick analysis [3], taking the RKI's raw figures as they are, does indeed reveal interesting results. The first is the continuous rise in the rate of vaccinated people among covid patients, so not simply positive cases, but cases that, in addition to a positive test result, show at least mild symptoms associated with the disease. This rise applies to almost all age groups and all levels of severity (cf. fig. 1-3). The rise is clearest for those aged 60 or over, and since these are risk groups, it is important to stress this. It is slighter for 18- to 59-year-olds. The rise remains almost indiscernible only for 0- to 17-year-olds, but it is present (22/07: 0.03% of symptomatic cases and 0% of hospitalised cases; 04/11: 4.2% of symptomatic cases and 3.4% of hospitalised covid cases of this age).
Figure 1: Changes in the vaccination rate and in the rates of vaccine failure among symptomatic, hospitalised, ICU and deceased cases in people aged 60 or over. Source: RKI.
Figure 2: Changes in the vaccination rate and in the rates of vaccine failure among symptomatic, hospitalised, ICU and deceased cases in 18- to 59-year-olds. Source: RKI.
Figure 3: Changes in the vaccination rate and in the rates of vaccine failure among symptomatic, hospitalised, ICU and deceased cases in 0- to 17-year-olds. Source: RKI.
From these rates, we can calculate an estimate of vaccine efficacy for each age group using, as the RKI does, Farrington's screening method. Following this model, we can estimate vaccine efficacy according to the RKI (cf. fig. 4-6). Several remarkable things then emerge.
Abnormality: inversion of vaccine failure rates
First, something we had already pointed out in an article on the ANSM report on vaccine failures: an inversion of vaccine failure rates according to the level of severity, i.e. among those aged 60 or over, the vaccine failure rate among deceased patients is higher than that among hospitalised patients (except between 07/10 and 14/10) and ICU patients. This opens up several conclusions for us: either the vaccine failure rates among hospitalised/ICU patients are underestimated, and vaccine efficacy against milder forms and severe, non-fatal forms would consequently be overestimated. Or the number of patients who died of covid is overestimated (by not distinguishing between “died of” and “died with” covid), artificially inflating the mortality rate.
Vaccine failures on the rise
Second, we can see a continuous increase in the vaccine failure rate, especially among those aged 60 or over, who, according to the authorities, nonetheless constitute the risk groups. In three and a half months, from the end of July to the beginning of November, the rate of vaccinated people among symptomatic patients increased by 43.6 points, climbing from 16.9% to 60.5%. This means that at the moment, among those aged 60 or over, more than one symptomatic case in two has been vaccinated! This goes hand in hand with a continuous decrease in vaccine efficacy, which, among those aged 60 or over, has fallen by 16.7 points, from around 89.5% (for mild symptomatic forms) to around 72.8%.
Figure 4: Changes in the vaccination rate and in vaccine efficacy for symptomatic, hospitalised, ICU and deceased cases in people aged 60 or over. Source: RKI.
Figure 5: Changes in the vaccination rate and in vaccine efficacy for symptomatic, hospitalised, ICU and deceased cases in 18- to 59-year-olds. Source: RKI.
Figure 6: Changes in the vaccination rate and in vaccine efficacy for symptomatic, hospitalised, ICU and deceased cases in 0- to 17-year-olds. Source: RKI.
We need to get back to reality
“72.8% is still not bad”, one might say, and “Look at the high vaccine efficacy among young people, it's proof that it works!”. But on closer inspection, that is not quite the right conclusion. Farrington's screening method has its limits, because it does not take certain realities into account.
First, there is little point in basing vaccine efficacy on groups that are not at risk to begin with, i.e. where severe forms are rare or even non-existent. This does not allow conclusions to be drawn about vaccine efficacy for risk groups. Conversely, it means that if we know which the risk groups are and they are fairly small, vaccine efficacy can be manipulated by vaccinating as many people as possible who, in any case, are not at risk of falling seriously ill. In the end, it will no longer be possible to determine whether they did not fall ill thanks to the vaccine or thanks to natural immunity. This is therefore more a matter of statistical manipulation than of serious epidemiology. It would in fact be playing on small numbers and on a lack of understanding of the difference between relative risk reduction (RRR) and absolute risk reduction (ARR): if we imagine that a group's risk is 5%, a decrease of 1 or 2 points makes it possible to claim a decrease of 20% or 40%, and therefore impressive efficacy, whereas the same decreases of 1 to 2 points in a group whose risk was, say, 30% would amount to only 3% or 6%, and would therefore be almost insignificant.
Second, the testing strategy inflates vaccine efficacy: unvaccinated people were - and still are - forced to get tested much more often than vaccinated people (in summer to go on holiday, to take a plane or train, when returning from red zones, for venues requiring the health pass, etc.). It is therefore possible that the number of unvaccinated people in the patient groups is clearly artificially over-represented because of testing strategies and regulations, thus overestimating vaccine efficacy. Moreover, we note that nowhere in the RKI reports is there any mention of vaccine failures among “asymptomatic vaccinated people”. Whereas for a year and a half we have been bombarded with “asymptomatic cases” and the danger they represent and why one must get tested even when not ill, this suddenly seems to be of no importance whatsoever in the case of vaccinated people, even though it has been established that they are just as contagious , (if not more so) as the unvaccinated. That is an undeniable double standard.
Third, the definition of the vaccinated underestimates vaccine failures. By considering as fully vaccinated only people whose last dose was 14 or more days ago, the RKI's definition of the vaccinated systematically underestimates the number of positive cases among the vaccinated. Since at least 3 weeks are needed between the two doses of Pfizer and Moderna, plus 2 weeks after the second to be labelled “vaccinated”, this means that for 5 weeks people vaccinated with these two vaccines will not be considered fully vaccinated. All positive cases occurring during these 5 weeks will therefore be subtracted from the count of vaccine failures, thereby de facto exaggerating vaccine efficacy. But by a perverse effect, since these people are included in the unvaccinated group, this increases the pathological problems among the latter in an entirely artificial way.
What supports this hypothesis are the official English figures, which show that the partially vaccinated account (on average) for around 5% of all positive cases. If we (theoretically) increase the RKI's official failure rates by 5 points for the 60-or-over risk group alone, vaccine efficacy is even lower (fig. 7). This would give a vaccine efficacy of only between 66.2% and 88.4% depending on the level of severity, with a downward trend. As recently as 2 months ago, this share of partially vaccinated people was not only larger (up to 25%), but it was not the same for all age groups or levels of severity. Moreover, it varies over time (at most up to 15% as at 19 September 2021). The decrease over time in the share of partially vaccinated people is simply explained by the fact that fewer and fewer people can still belong to it, since fewer and fewer people are available for vaccination. Either because they are already vaccinated, or because they have not done so until now and never will. However, the fact that the share of partially vaccinated people varies suggests that vaccine efficacy would be variable. Yet normally, for the same risk groups, efficacy should be (if the vaccine works in the long term) more or less stable, which suggests that we are dealing here with an artefact of the testing strategy or of the counting, or of both, casting doubt – surprise, surprise – on the validity of the figures and on vaccine efficacy as such.
Figure 7: Changes in the vaccination rate and in vaccine efficacy for symptomatic, hospitalised, ICU and deceased cases in people aged 60 or over, each theoretically increased by 5 points. Source: RKI.
That said, the English figures reveal something else too: as at 31 October 2021, more than four-fifths of patients aged 60 or over who were hospitalised (83.8%) or died (83%) because of covid were vaccinated , , [13], [14]. Set against the vaccination rate among the English population aged 60 or over (estimated as at 31 October 2021 at 13,089,249 out of 13,660,832, i.e. 95.8%), this corresponds to a vaccine efficacy against severe forms in the risk group of only 77.5% to 78.7%. This is not yet confirmed by the RKI figures, but given the rising vaccine failure rates, it could be confirmed in the coming months. It could moreover be explained by the lead the English have over the Germans in terms of vaccination, which would favour the spread of resistant variants.
Fourth, there are methodological criticisms of the Farrington screening method used by the RKI. First, the results of this method depend on the observation period. Thus, since week 5 of 2021, vaccine failures among those aged 60 or over represent only 18.9% according to the RKI, which would amount to a vaccine efficacy of 95.9% over the entire vaccination period. However, as we have seen, vaccine failures are increasing, and vaccine efficacy over the 4 weeks before the publication of the latest RKI report is only 72.8% for symptomatic cases among those aged 60 or over. It may be even lower than that if we had only the figures for the last 2 weeks. The method also has an unavoidable, if rather theoretical, flaw, which is that as soon as the vaccination rate reached 100%, vaccine efficacy estimates would become impossible to calculate, because one would have to divide by 0.
Another, less theoretical flaw: the rarer a disease, the more sensitive it is to minor absolute variations. Let us imagine a situation in which there were very few unvaccinated people against a rare disease. For example, in a population of 10,000,000 inhabitants with 9,900,000 vaccinated (99%), suppose an outbreak of 100 sick people occurs and the distribution of cases equals the vaccination rate of the population: 1 unvaccinated sick person (0.001% of 100,000) versus 99 vaccinated sick people (0.001% of 9,900,000). This would amount to a vaccine efficacy of 0%. However, if 2 more unvaccinated people fell ill (thus going from 0.001% to 0.003% of 100,000), it would not change much, either in the distribution of the sick (97.1% would be vaccinated versus 2.9% unvaccinated) or in the epidemiological situation, and would rather be anecdotal or a matter of daily variability, but vaccine efficacy (relative risk reduction) would then rise to 66%, while the absolute risk reduction would go from 0 points to 0.002 points (the difference between 0.003% and 0.001%).
It must also be stressed that this method can produce unreliable estimates. If, for example, we take the English figures and calculate the estimates not according to Farrington's screening method, but on the basis of incidence among the vaccinated and the unvaccinated (case-control method ), then as at 31 October 2021 vaccine efficacy for mild cases among those aged 60 or over is only -18.3%! Vaccinated English seniors are therefore 18.3% more likely to test positive than unvaccinated English seniors and to share their fate: isolation, quarantine, denial of access, etc. And as for severe forms, hospitalisations and deaths, with 73.7% and 74.1% efficacy, we are already a long way from the 95% announced 11 months ago, with a downward trend.
Fifth, we must hunt down confounding factors before jumping to conclusions with the raw figures. The WHO has clear guidelines for evaluating vaccines. These confounding factors represent possible explanations that must be eliminated at all costs if one wants to measure the variable under study. A dozen of them can be found here.
Basically, this amounts to setting up control cohorts. These cohorts do not need to contain tens of thousands of individuals; a few hundred or thousand are enough. The control consists in creating two cohorts with more or less comparable profiles in terms of demographics (age, sex, income, education, place of residence, etc.), in terms of risk factors (e.g. number of smokers and non-smokers, use of contraceptives, etc.), and of underlying conditions (e.g. hypertension, diabetes, cancers and other comorbidities). That is to say, the cohorts must on average be as identical as possible; then one is entitled to carry out a multifactorial analysis that has a chance of proving anything at all. Relying on rules of three or on methods closer to political polling or opinion surveys gives no information that could justify decisions.
Exclusion will also apply to vaccinated people who test positive
Let us also not forget that isolation and quarantine measures, denial of access, etc. apply to all positive cases, regardless of the level of severity or symptoms. Yet if one hopes to escape them by getting vaccinated, or even got vaccinated mainly in order to “regain one's freedom”, those chances are melting away before our eyes. For what will this trend look like in two months? If the increase in the rate of vaccinated people among symptomatic patients aged 60 or over continues at the same pace as over the last two months – as the RKI figures suggest – then by mid-December at the latest it will equal the rate of vaccinated people in the population aged 60 or over. When that happens, vaccine efficacy for symptomatic cases will fall to 0% (this is already the case in England). It will fall below 50% well before that – Israeli figures already estimated it at only 41% in the summer. For severe forms it will take longer, but according to current official trends, it will happen sooner or later.
The high season will reveal true vaccine efficacy
Finally, and this is the most important point, this may happen sooner than expected. For the winter seasonality of sars-cov-2 is, after some initially sceptical voices, increasingly recognised and accepted , , [24], [25]. Vaccine efficacy calculations are therefore based on “off-season” infections. From the absence, or rather the summer scarcity, of the virus, it would be fallacious to conclude any vaccine efficacy whatsoever. It is a bit like clapping your hands in the middle of Brussels and concluding, since there are no elephants around, that it is an effective gesture for keeping elephants away. Only in the season of increased transmission will vaccine efficacy really show itself. However, since it is already decreasing now, even though we are not yet in the “high season”, it seems more than uncertain.
Don't be taken for a ride
It is moreover likely that the authorities will twist this fact once the declining, or even non-existent, efficacy has become undeniable, and will use it in a purely opportunistic and cynical way to promote or even impose booster doses, along the lines of: “It is because the vaccine loses efficacy after a few weeks or months that regular boosters are needed. And if you choose not to have them, the choice to have them will be imposed on you.” This does bear a strong resemblance to the update system for protecting a certain computer system belonging to a certain American multi-billionaire against viruses (computer ones, in this case)…
However, it should be noted that ineffective vaccination will not become more effective by increasing the dose or the frequency of injection – 1 times 0 equals 0, and 2 or 3 or 10 times 0 still equals 0. Frequent boosters therefore cannot be justified (and certainly even less imposed) on the basis of the ineffectiveness of previous doses. For nothing certifies or proves that the efficacy of these vaccines, which, let us remember, are still in phase 3 trials, increases with the dose. On the other hand, each injected dose mechanically increases the chances of suffering side effects that could prove more serious than the disease itself.
By Colin Meier, citizen journalist at BAM!
While knowing full well that these figures are certainly not correct (see the close correlation between testing strategy and number of cases, false positive tests, dubious definitions of cases and of the vaccinated, such as “asymptomatic cases” or the “partially vaccinated”, omission of confounding factors such as the comorbidities of those hospitalised and deceased, etc.). In any analysis of health data, we would point out that it is important to exclude what are known as confounding factors, and we will come back to this.
Link to Excel file (in French)
https://academic.oup.com/ije/article-abstract/22/4/742/664122?redirectedFrom=fulltext: Basically, this method relates two ratios: it divides the rate of vaccinated by the rate of unvaccinated among patients of a given level of severity, and the rate of vaccinated by the rate of unvaccinated in the general population. The first ratio is then divided by the second, the result subtracted from 1 and the whole multiplied by 100. This gives an estimate of vaccine efficacy as a percentage. If among patients of a certain level of severity the rate of vaccinated is 0, then vaccine efficacy will be 100%, since the first ratio would be 0 and (1-0) * 100 = 100. The sick would consist only of unvaccinated people. If, on the other hand, the rate of vaccinated among these patients is equal to the rate of vaccinated in the general population, then vaccine efficacy will be 0%, since dividing the two ratios would give 1 and (1-1) * 100 = 0. The composition of the patient group would not differ from that of the general population. If, however, the rate of vaccinated in the patient group were even higher than the rate of vaccinated in the general population, then vaccine efficacy would even be negative, since the vaccinated would be over-represented in the patient group compared with the general population. Warning! We would point out that this method does not take into account the causes of hospitalisation, false positives, debatable definitions or confounding factors.
https://bam.news/revue-de-presse/la-charge-virale-des-personnes-vaccinees-est-tout-aussi-elevee/
https://www.nature.com/articles/d41586-021-02187-1
https://www.md-universal.eu/images/Vaccine-surveillance-report-week-42.pdf (13pp.)
https://www.md-universal.eu/images/Technical_Briefing_20.pdf
https://www.md-universal.eu/images/Vaccine_surveillance_report_-_week_38.pdf
https://www.md-universal.eu/images/Vaccine-surveillance-report-week-41.pdf
https://www.md-universal.eu/images/Vaccine-surveillance-report-week-42.pdf
https://www.md-universal.eu/images/Vaccine-surveillance-report-week-43.pdf
https://www.md-universal.eu/images/Vaccine_surveillance_report_-_week_44.pdf
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3411566/
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3411566/
https://www.nature.com/articles/s41467-020-18150-z#Sec6
https://www.md-universal.eu/images/jiaa436.pdf
https://agupubs.onlinelibrary.wiley.com/doi/epdf/10.1029/2021GH000413
https://www.md-universal.eu/images/main.pdf
