Latest Stats (6 March 2023)...
NEW ZEALAND PFIZER VACCINE EFFICACY & SAFETY Cases, Hospitalisations & Deaths
The graphics on this page are produced using the regular updates on the Ministry of Health website. Where population data is required, it is sourced from Statistics NZ. The spreadsheets are here. The Seven Day Averages are updated weekly, and the remainder of the graphs are updated most days (update: in September 2022 the MoH dropped to weekly data updates instead of daily).
Note (29/10/22): The MoH stopped providing comprehensive data on the vaccination status of the population. Vaccination uptake had slowed sufficiently by this point, that the lack of updated data ought not make an appreciable difference to percentages in each group.
I HAD MY ANTIBODIES MEASURED & WROTE ABOUT SEROLOGY, & VACCINATED vs NATURAL IMMUNITY HERE.
READ ABOUT THE NZ MEDSAFE MYOCARDITIS DATA HERE.
OBSERVATIONAL DATA AND INTERVENTIONAL DATA
The graphs on the page use observational data provided by the MoH: We are watching this play out in real-time, we don’t get to decide who receives the vaccine, and we are unable to control for all the potential confounding variables which may affect the data.
An example of how confounding variables may affect the data: The elderly have the highest rate of receiving boosters, so this may make boosters look worse, because elderly people are more likely to die than young people, regardless of any effect from the booster. However, Māori and Pacific Islanders are less likely to be vaccinated. We are told that these communities have poorer health outcomes across the board, so we would expect this to create an appearance of poor outcomes in the unvaccinated. Confounding variables could be working both for and against the appearance of vaccine efficacy.
This MoH does not release sufficient data for us to adequately describe confounding variables, so we are left to interpret the data at face value - we must bear in mind that this interpretation is weakened by the lack of comprehensive data. If the vaccine were effectively reducing infection and severe illness to the degree that we were promised, one would expect the blue and red columns (cases and hospitalisation) to be smaller than the yellow columns (percentage of population by vaccination status) for the vaccinated.
Randomised controlled trials (RCTs) provide interventional data: A carefully selected group is given an intervention, and a comparable ‘control’ group is given a placebo, and the two are compared for outcomes. The groups are ‘blinded’ - individuals do not know whether they have received the intervention or the placebo. This is because, in most cases, knowing that you have received an intervention makes a significant difference to the outcomes. It is unfortunate that the original Pfizer trials were unblinded only a couple of months after they started, so we have no high quality long term efficacy or safety data. It is much easier to control for confounding variables in an interventional study.
At face value, it is hard to marry what we see in the data with the ‘pandemic of the unvaccinated’ rhetoric.
MISCATEGORISATION
Before we look at the MoH data, we need to understand how they miscategorise vaccine recipients. The Ministry does not count an individual as having received a vaccine until a week after they have had the injection.
Miscategorisation One: Stacking the data
In the first week after receiving a vaccination there is evidence that people become more likely to contract covid. This may be because of a transient vaccine-induced alteration of haematological parameters leading to immunosuppression. This involves a significant increase in the neutrophil:lymphocyte ratio - which is also a consistent finding in severe and fatal covid cases. It may be that catching covid during this period of immunosuppression could result in more severe outcomes, in which case it would be prudent to advise recently vaccinated individuals to isolate for a week or two. I am working on some further notes on this topic here.
Because the vaccination is not counted for the first week after it is administered, the individual will still be counted as a covid case in their previous category - thus unfairly stacking the lower categories.
For example: An individual is counted as unvaccinated until a week after their first dose; and an individual is counted in the two-dose category until a week after they receive their booster. Figures 1 and 2 illustrate how cases are moved down a category - leaving the boosted looking like they are doing better, and the unvaccinated like they are doing worse.
Figure 1
Figure 2
The justification for the week-long delay is that the immune protection from the vaccine does not kick in right away. But this only considers the benefits of the vaccines. As well as the potentially greater risk of infection in the first week, there will also be adverse events - the majority of which occur in the first 48 hours. The effects of the vaccine should be considered from the time of injection.
Miscategorisation Two: Ignoring the children
If a vaccinated 5-11 year old (of which there are over a quarter of a million) becomes a covid case, hospitalisation or death the MoH counts them all to the unvaccinated category.
Miscategorisation Three: Death data delay
So, we can see that the week-long delay stacks the data in favour of more doses of the vaccine. But the effect of the delay in categorisation can stretch out for much longer than a week. I have had some back and forth correspondence with the MoH regarding death categorisation, and they have confirmed the following via email:
Example: An individual receives dose one of the Covid-19 vaccination:
If they become a covid case before seven days have passed – they are classified as an unvaccinated covid case.
If they become a covid hospitalisation before seven days have passed – they are classified as an unvaccinated covid hospitalisation.
If they die from or with covid before seven days have passed - they are classified as an unvaccinated covid death.
But here is where the data really gets stretched:
If the individual becomes a covid case six days after they are vaccinated, get hospitalised two weeks later, and dies a week after that - four weeks after they were vaccinated - they are classified as an unvaccinated case, an unvaccinated hospitalisation and an unvaccinated death.
Thus, if someone dies with or from covid four weeks after they were vaccinated - the death can be added to the unvaccinated tally.
SEVEN DAY AVERAGES
(The MoH moved to weekly data releases in September 2022, so the 7 day averages are no longer needed and will be replaced with a 7 day snapshot)
Figure 3: Seven day averages snapshots
Figure 3 shows seven day averages snapshots for:
the proportion of the population in each vaccination group (yellow)
the proportion of total case numbers (blue)
the proportion of hospitalisations (red)
The group called ‘any vax’ includes anyone who has had at least one dose. In order to accurately measure how the unvaccinated are tracking, we need to be able to compare them against anyone who is not unvaccinated.
Example: For the ‘dose2’ group: 29% of the population have had 2 doses, they make up 3.5% of covid positive hospitalisations, and 22.1% of cases are in people who have had 2 doses. A group is doing well when its red and blue columns are shorter than its yellow column (i.e. they are under-represented in hospitalisations and cases).
Figure 3
Figure 4: Cases and hospitalisations rates /100,000 in each category
Case rates are highly unreliable, because some people are more or less likely to test, or to report an infection than others. Hospitalisation rates are more reliable, because everyone who goes into hospital gets tested. The MoH does not tell us how many people are being admitted because of covid, rather than with covid, so the actual numbers are a misrepresentation, however these graphs are still useful for:
identifying trends,
and comparing the performance of the different categories.
Figure 4
Figure 5: Unvaccinated vs vaccinated
The unvaccinated are compared with everyone who has received at least one dose.
Figure 5
DAILY SNAPSHOTS & TRENDS
The MoH shares the data in a cumulative fashion. Snapshots can be elicited by subtracting each day from the next.
Note: The MoH changed some of their classification methods in July - it is unclear exactly how the parameters were moved, but you can see that the trends shift somewhat following the change. Around 6000 hospitalisations were removed from the vaccinated tallies, and just under 500 were removed from the unvaccinated. The boosted group was diverging significantly for cases and hospitalisations, but they are now tending towards convergence. This is at odds with the covid mortality data which shows the convergence continuing at an increasing pace.
Figure 6: Fluctuation in case numbers by vaccination status
Initially, data stacking (described above) enhanced any transient vaccine efficacy, meaning that the boosted category looked like it was doing really well, but the dose-two category looked bad.
Figure 6
Figure 7: Unvaccinated vs vaccinated
Trend lines are plotted so that we can see more clearly where the data is heading. They crossed in early-mid April at which point vaccine efficacy for cases became negative (the vaccinated became more likely to catch covid). The same thing happened with hospitalisations about ten days later.
Figure 7
Figure 8: Fluctuation in hospitalisation numbers by vaccination status
Note: MoH found a lot of ‘missing’ hospitalisation data in early October, hence the uptick.
Figure 8
Figure 9: Unvaccinated vs vaccinated
Note: MoH found a lot of ‘missing’ hospitalisation data in early October, hence the uptick.
Figure 9 includes trend lines. The trend lines crossed in mid-late April at which point vaccine efficacy for hospitalisations became negative (the vaccinated became more likely to become hospitalised with/from covid).
Figure 9
DEATHS
Deaths WITH covid
The MoH counts all deaths as covid-deaths if they occur within 28 days of being ‘reported as a case.’ Deaths from unrelated conditions (for example, they include a man who was shot in his driveway whilst being covid positive) are included in the count as long as the individual was reported as a case within the last 28 days. The media and the MoH routinely refer only to the deaths with covid when addressing the public. This creates an impression of greater mortality due to covid. (N.B. The MoH recently announced that they would shift the focus to deaths from covid. Despite this announcement, they still only provide vaccination status for deaths with covid - so I will continue to use these deaths in my graphs - the raw numbers will be debatable, however the graphs are still useful for monitoring trends).
For deaths with covid the MoH combines the unvaccinated group with the one-dose group. These two groups are not equivalent. It is possible that people in the one-dose group did not get further vaccinations because they suffered more adverse reactions than the other groups - thus they may be sicker overall. This would make the unvaccinated group appear to do worse. Regardless, it is unnecessary and inaccurate to put vaccinated deaths into the unvaccinated group.
Deaths FROM covid
The MoH provides the deaths from covid: the number of deaths for which Covid-19 is coded as the underlying cause (currently running at a little under half the ‘deaths with covid’ count). Deaths from covid can occur more than 28 days after being reported as a case.
An OIA request revealed that, prior to the August 2021 cluster, 5 of the 22 covid deaths never tested positive for covid (one was not tested, and four tested negative). Thus, it appears that an individual can be coded as a death from covid even if they have tested negative to covid.
Figure 10: Proportion of deaths relative to proportion of people in each vaccine category
Figures 10 shows:
The vaccine appeared to have initial efficacy against death, especially back in March 2022. This efficacy dropped over the next few months, and by the end of June, unvaccinated people are underrepresented in covid-related deaths.
The boosted are showing increasing negative efficacy. Boosted people are the most likely to die from/with covid. This may be due in part to a higher number of elderly people in the boosted group, however this effect will likely be somewhat tempered by the overrepresentation of Māori (who have higher all-cause mortality than other groups) in the unvaccinated group. Detailed data would be required to analyse the extent to which confounding variables are affecting the outcomes.
Figure 10
Figure 11: Unvaccinated deaths
The MoH combined the deaths of the unvaccinated with the one-dose deaths. In figure 11, the two groups are separated out by taking the proportion of the population in each group, and assigning that proportion to the deaths - thus shifting some deaths from the ‘unvaxed + dose 1’ group into a ‘dose 1 + dose 2’ group, and leaving a clean ‘unvaccinated group.’ This is a crude attempt to make up for the data fudging, and cannot be relied upon to be wholly accurate. It does, however, demonstrate that combining the unvaccinated and the one-dose group is likely to be somewhat misleading.
Figure 11
Figure 12: Deaths per 100,000 in each category
Figure 12
WHAT EFFECT DOES VARYING VACCINE EFFICACY (AGAINST CASES & DEATHS) HAVE ON COVID-19 MORTALITY RISK?
The Pfizer vaccine appears to be associated with increased rates of infection. If we assume for a moment that this vaccine does reduce covid mortality, then this benefit will be lost once a certain level of increased infection is reached. I borrowed and simplified an idea from substacker El Gato Malo - this section is speculative, but interesting to play around with.
Vaccination efficacy wears off rapidly, and becomes negative - leaving vaccinated individuals at an increased risk of COVID-19.
In England, this increased risk is between 2-4 x (compared to unvaccinated) depending on age category (older people have more increased risk).
The argument in favour of vaccination is that, despite the increased risk of becoming a case, it reduces the risk of death. However - even if vaccination was reducing covid mortality risk for an individual, if it also drives up case numbers there comes a point at which the vaccine starts to cause an increase in population covid mortality overall.
We can explore that through a very simple model. It is fairly crude, but it illustrates how a neutral or detrimental outcome can easily be sold as a beneficial one.
We will utilise the following data and assumptions:
For simplicity, we will assume vaccination is the only variable affecting covid outcomes (obviously not true in reality)
85% of people are vaccinated (approx 85% of the total NZ pop is vaxxed).
Vaccination increases the risk of catching COVID-19 by 2-4 times according to UK data. On July 2nd the vaccinated were just over 2x more likely to be a case. NZ’s vax rollout is a little behind the UK temporally.
Assume 10% of unvaccinated people catch covid (and therefore 20, 30 and 40% respectively of the vaccinated). For the sake of this model, it doesn’t matter what this figure is as it will not change the final change in percentage of mortality.
IFR of 0.5% (this is on the high end, but as above, it won't affect our outcomes).
Vaccination reduces mortality by 50% (this is very generous, given that the boosted are experiencing disproportionately high mortality relative to the unvaccinated, and given that an all-cause mortality benefit of -0.3% has been demonstrated with the mRNA vaccines).
Vaccination increases mortality by 2.4%: This is the face value data for NZ as of July 2nd (see Deaths: figure 11) - bear in mind that this data does not include adjustment for any variables, so the true figure will likely differ.
Interpretation
It is now apparent that vaccination is followed by an increase in covid case numbers.
The argument in favour of vaccination is that, despite the increased risk of becoming a case, it reduces your risk of death.
However - even if vaccination was reducing covid mortality risk for an individual, if it also drives up case numbers there comes a point at which the vaccine starts to cause an increase in population covid mortality overall.
I have made an optimistic assumption: the vaccine is 50% efficacious in reducing covid deaths…
…and a pessimistic assumption: the vaccine has a -2.4% efficacy against death - this is what is currently seen in the unadjusted (adjustment would give us more accurate results) NZ data.
Assume vaccination leads to a 2 x increase in cases (vax is associated with just over a 2 x increase in NZ at present):
Optimistic: Vaccination ⇒ 0% incr in covid mortality.
Pessimistic: Vaccination ⇒ 47% incr in covid mortality. The assumptions that lead to a 47% incr best reflect NZ’s current situation.
Assume vaccination leads to a 3 x increase in cases :
Optimistic: Vaccination ⇒ 30% incr in covid mortality.
Pessimistic: Vaccination ⇒ 64% incr in covid mortality.
Assume vaccination leads to a 4 x increase in cases (this is the case in older age groups in the UK at present):
Optimistic: Vaccination ⇒ 46% incr in covid mortality.
Pessimistic: Vaccination ⇒ 72% incr in covid mortality.
'If a vaccinated 5-11 year old (of which there are over a quarter of a million) becomes a covid case, hospitalisation or death the MoH counts them all to the unvaccinated category. ' do you have a reference for this Splendid? I did a post on the high number of kids being hospitaised with covid. This would explain that.