Reflecting on the Walgreens data

Dr. Paul Alexander pointed his subscribers to an incredibly interesting data set featuring data from over 5000 of 7000 Walgreens testing locations. I wanted to take some time to share my take on it because the data shows why I was wrong about test negative designs being used to justify vaccinations even long after the vaccines became useless. At the time, I figured that hysterics among vaccine enthusiasts would cause them to test far more frequently than the unvaccinated even without symptoms or for unrelated illnesses. The idea was, simply put, that frequent testing would outpace positive tests and the unvaccinated would test less frequently, but with more understanding of their symptoms.

I see no reason why my assumption about hysterics is incorrect — in fact, 45.3% of the tests taken at these Walgreens locations were in the boosted populations, which is a higher rate than the boosted population in the U.S. population by a sizeable margin. It seems implausible that these 5000 or so Walgreens locations are located in areas where people have taken the booster at rates far above the general population, so hysterics seems to satisfy Occam’s razor.

While it is possible that the unvaccinated are testing more often for work purposes (ie., exemptions to mandates), the trends in the data are largely unidirectional regardless of age. The unvaccinated below and above working age are testing positive slightly more frequently than their working age counterparts, but that effect, if it exists, would not be strong enough for the difference in outcomes witnessed here.

I also believe the fact that the unvaccinated are doing better than the vaccinated in the youngest age groups, ie., 5-11, means that understanding the symptoms is probably not the main factor here. Unvaccinated children of a young age can still have psychotic, coronavirus obsessed parents.

Now, I want to contextualize this before I say anything else: test negative control designs are still improper in this context. I will remind you that these designs are only, only, useful when the same type of people are getting tested for the same reasons in the same setting using the same type of test. But we are seeing similar results in the raw Canadian data, say, without using test-negative designs, which no one collecting the data wants to see, so the Walgreens data set does not exist in a vacuum.

In short, this data is troubling.

In total, vaccinated individuals, regardless of “booster” status, have a test positivity rate of 18.16%, compared to just 10.49% in the unvaccinated.

If we were to use a test negative design, that means vaccine effectiveness is -73.2%, and most of that negative effectiveness is concentrated in boosted individuals. In other words, the boosted are doing worse than the unvaccinated, those with one dose, and those with two doses.

In terms of raw cases, 82.6% of all cases (8835 of 10697) are in the vaccinated, despite just 73.3% of all tests being in this group. Again, the majority of cases are in boosted (6077) despite far from the majority of individuals in America having received the booster.

Here is Walgreens explanation:

“All results, including the positivity rates by vaccination status graph, are unadjusted. The team has observed that the positivity rates among unvaccinated individuals seen on the bar graph appear to be lower in comparison to vaccinated individuals. Furthermore, repeat testing among those who were previously positive in the last 90 days appear to confound the results. The team conducted additional analyses examining characteristics of the patient population by vaccination status and the impact of excluding recent COVID-19 cases (5.0% of total tests). Findings show that the unvaccinated group are typically younger and healthier, less symptomatic and less likely to report direct COVID-19 exposure or recent travel compared to vaccinated groups. Controlling for recent COVID-19 cases, results show that the unvaccinated group has a 17.1% higher positivity rate compared to the 3-dose group. Controlling for additional factors leads to a larger difference between groups”

This explanation was provided on March 11 and remains a headscratcher as far as explanations go. First off, it is true that unvaccinated groups are typically younger, but I find it laughable that a pharmacy claims to have detailed information on the health status of those they are testing. More likely, they are extrapolating, ie., because unvaccinated individuals are younger, they must be healthier. But this is a ridiculous argument, in my opinion, as Walgreens breaks down test positivity by age and, regardless of age group, the unvaccinated still do better. Only in those aged 65 and over do individuals with a single dose have a lower positivity rate — which appears to be a statistical anomaly due to small sample size (there were less than 100 tests in this age group for those with one dose).

The more interesting points are that Walgreens claims the unvaccinated are:

1. Less symptomatic; and,

2. Less likely to report direct COVID-19 exposure or recent travel.

Of note, magnitude is not given here. How much “less”?

We are also left to guess what symptomatic means in this context, but I will assume that they *mean* less symptomatic from self-reports, which is decidedly not the same as being less symptomatic. Furthermore, it is unclear whether they mean individuals that test positive are reporting less symptoms or all individuals testing are reporting less symptoms.

For starters, if unvaccinated individuals testing positive are reporting less symptoms, that could mean they have less trust in Walgreens pharmacy (seems rational) or it may mean the vaccinated are more likely to associate any symptom as being from the virus. Or it could be suggestive of antibody dependent enhancement in the vaccinated. Who knows?

If, on the other hand, they mean all individuals testing are reporting less symptoms, it may mean that my theory on unvaccinated individuals being forced to test for work holds some weight. I would like to see “less symptomatic” stratified by age as it seems unlikely that any parent is going to let someone stick a Q-tip covered in chemicals far up their child’s nose just for the hell of it.

Furthermore, one would expect unvaccinated individuals to be less symptomatic if they are testing negative more frequently. I cannot emphasize this enough, so I will capitalize it to make it seem like I am yelling: THAT IS THE POINT OF THE TEST. The fact they controlled for that should make people extremely suspicious of their ability to control for anything.

Finally, being less likely to report direct COVID exposure or recent travel (interesting that these are combined) does not say much. Once again, I would like to see it stratified by age to remove the “test to work” effect, but even then, it might just be that unvaccinated individuals are more likely to hang around unvaccinated individuals who are less likely to be sick.

For a company that claims to “champion the health and well-being of every community across America”, it seems like they are more interested in championing a failed vaccine due to the profit incentives associated with sticking needles in people’s arms. Otherwise, I do not understand the need for the manipulative disclaimer attached to their own data.