Data manipulation in Alberta's vaccination outcomes page

The devil is in the details

“Facts are stubborn things, but statistics are pliable.”

— Mark Twain

Governments take it for granted that they can easily hide flaws in their data collection. After all, private citizens have no access to individual health records. If the data collection is biased, then anything that is done with the data after does not matter. Even if the data is collected without bias, it should only be used to compare non-random populations with the greatest hesitation. That is why we needed long-term double-blinded, randomized control trials for the vaccines. No one should have trusted the original data collection of Pfizer or other vaccine manufacturers for a variety of reasons, not the least of which is that they have financial motives to manipulate the data and have done so in the past. But once they unblinded the vaccine trials, it was clear the fix was in ipso facto.

At this point, so much of the population has been vaccinated in Canada that it is nigh impossible to intelligently compare those who are vaccinated to those who are not. It is unclear who, exactly, is getting vaccinated. What if, say, people with natural immunity to the virus are actually more likely to be vaccinated? In what proportions might this be occuring? How does that skew the vaccine effectiveness calculations that all of these lowly health analysts are calculating without a second thought? I doubt these benefit-cost ratios are even being calculated and if they are I am certain they are not including any factors beyond age in their calculations. No one with natural immunity would have taken the vaccine in the first place. Especially if it could be shown that those with natural immunity are more likely to have adverse events. No. The people making our health decisions for us are completely devoid of nuance.

Because the control trials were unblinded, those who are still uncertain about taking the vaccine, then, cannot make a decision based on good data and must throw the full force of their trust into government to be as honest as possible with them about the flawed data that is available. Most provincial governments in Canada have taken the opposite approach and only released the cliff notes — perhaps because they believe the general public is too stupid to interpret the data presented to them. Or perhaps, less likely, because presenting the data factually, however flawed and biased it already is, would interfere with what they believe to be a noble lie.

The provincial government of Alberta went a completely different route. To their credit, they have provided a fair amount of summary statistics and access to a data file (that does not match their summary statistics). They went further than any other province in doing so. Too bad the only thing more refreshing then their transparency is the reminder that the government thinks we are stupid.

For now, ignore the fact that they claim a full 23.8% of people that have died from COVID had cancer. Rather, let’s look at their vaccine outcomes page, which is the most interesting data in my opinion.

Upon entering the page, the above summary of how they see the last year is the first thing that pops up. Notice the first problem, “Since Jan 1, 2021”. The problem is, of course, that this is a moving target. We could safely ignore the entire page because of this misrepresentation of the facts if it were not so compelling to look through the rest. For example, Alberta had their 500th death with the virus on March 30th — at the time, 12.8% of the population had one dose of the vaccine, and only 2.5% were fully vaccinated. To pretend it is a fair representation of the facts to say that 77.7% of covid-19 deaths or 85.6% of cases were in the unvaccinated or diagnosed with the first dose less 14 days (more on that later) is strange.

But you’ll also notice their numbers do not add up. 10,422 were diagnosed after a full 14 days with the vaccine, and 18,301 with two doses after a full 14 days with the vaccine. Those unvaccinated or vaccinated less 14 days account for 178,322 of 208,369 cases. So vaccinated cases are 28,723, and the difference between unvaccinated and all cases is 30,047. Where are the missing cases? Well, those people are vaccinated with two doses less 14.

Strange that they group those with one dose less 14 with the unvaccinated but not those with two doses less 14 with the partially vaccinated? Another key point is they represent the cases with the vaccinated as a rather small percentage, 0.3% and 0.6% respectively, but because they are counting since January 1, they are grouping people that got vaccinated 15 days ago with those that got vaccinated at the beginning of the year. This says nothing about the vaccinated probability to get the virus, but it is represented as such. Or at least, it would be if we were stupid as they seem to believe.

The hospitalization rate is equally nonsensical — not only because of what has come before, but because most Canadian provinces are knowingly not counting those in long-term care homes with serious cases of the virus as in the hospital since, technically, they are not. Yet it is an important fact to neglect considering the vast majority of deaths in Canada have been in long term care homes when they died.

More, or perhaps less, interestingly, the next part of Alberta’s dashboard does not align with the statistics above. You might say that the image above does not say anything about new cases, but if you watch carefully, it does.

I cannot fully account for the holiday weekend and do not know if these new cases match up with their reported numbers, but last Tuesday, as an example, they claimed there were 473 new unvaccinated cases. Yet the case numbers of unvaccinated or one dose less 14 only went up by 418. That’s 55 cases overreported. The differences over the last week added 224 extra cases to the unvaccinated, 27 new cases for those with one dose, and counted 110 less for the fully vaccinated. And these numbers, obviously, do not add up. Where did the extra cases come from? How long has this been occuring? In fairness, they reported 3214 new cases over that period. 224 cases is only about 7% of that and their numbers are pretty skewed towards more unvaccinated cases. But, aside from the fact that new restrictions could mean it is more likely for unvaccinated to get tested, how can you trust their numbers when they are presented as they have been? How can you trust any inaccuracy? What kind of programming error accounts for this?

Well, the programming error is not a programming error at all. Those 224 cases are those who received the first dose less 14 days. Lumping them in with the unvaccinated cases is not misleading. It is unethical. Those overcounted in the 1st dose are 2nd dose less 14. But where are the 110 less fully vaccinated cases? Are they people on the third dose with the virus? This completely skews any rational accounting of the numbers.

But the most compelling part of their entire dashboard are their graphs because these allow us to see into how many of these dose less 14 are being miscounted. They also make one wonder what the chart above would look like if they were included in vaccine effectiveness.

Obviously, you can see how many cases (9839 in the 14 days after one dose…almost 6% of the “unvaccinated” from the first chart, 1324 after two) they are excluding from their calculations by not including those 14 days and, while they claim that first doze immunisation includes those who had the virus before the first 14 days of their second immunisation, I would be surprised if that’s actually the case considering they completely ignored it in earlier calculations. One also has to wonder, what is the probability that many cases occured naturally in the first 14 days of the vaccine? Are you more likely to get the virus in those days? Why, if so, are these people not told to isolate at home for 14 days after getting the vaccine?

Even more alarming is the amount of people that died in those days, a full 133 people died within the first 14 days of getting the first dose (nearly 13% of those lumped in with the unvaccinated in the first chart), and 19 within the first 14 days of getting the second dose (were these 19 even counted anywhere?). What was the probability that many would die in those days? Are they vaccine deaths? If the vaccine makes people more susceptable to the virus, many of them might as well be. Coupled with the aforementioned point about 500 of the deaths in Alberta coming when only 2.5% of people were fully vaccinated, it is obviously they are trying to make it look like this is a “pandemic of the unvaccinated”. We are also missing data on all cause deaths. Are vaccinated people less likely to die of

According to their numbers, more people have died in the unvaccinated population in the last two weeks. We would actually expect more to die in the vaccinated population as those who are older are more vaccinated if all things were equal. Though, it is unclear what characteristics those in the unvaccinated population had. For example, 5-7 cancer patients die every day “of” the virus in Alberta (are they vaccinating cancer patients in the province? Unclear). But things will get worse for the vaccinated in the coming months as even the vaccine companies acknowledge. Do the benefit-cost ratios account for boosters every few months? Is the data as good as it looks? It does not seem so. There is no reason to lie about good data, and this is not good data. At the end of the day, it leaves us with more questions than answers.