[PepsiFree]

Originally Posted by Darkknight

https://science.thewire.in/health/na...-israel-study/

Is an excerpt from the study that's been referenced. I don't believe this study has been refuted, but if it has I would be interested in seeing that. I've seen it referenced in the Washington Post as well for whatever that's worth.

That study hasn't been peer reviewed, which may or may not make a difference, but there are also some flaws in the study. You can read this comment from the study page:

Dear colleagues,

With interest did we read this manuscript which fueled a lively discussion during our journal club of the department of infectious diseases epidemiology at the University Medical Center Utrecht. The authors address a relevant research question. If there is a substantial difference in the risk of SARS-CoV-2 infections between previously infected and vaccinated individuals – as suggested - this may have consequences for social distancing, testing recommendations, and for projections of the impact of vaccination on future COVID-19 trends. However, we have several concerns regarding generalizability, selection bias, information bias, and confounding that we would like to address. We focus our discussion on model 1: the comparison of the fully vaccinated non-infected group (group 1) to the infected non-vaccinated group (group 2).

In regard to generalizability:
- Due to the matching process, only 4% of the available data is used (i.e. for model 1 only 32430/736559) and as a consequence the study population is fairly younger (with expectedly less comorbidity) than the source population (i.e. vaccinated individuals, infected individuals). Therefore, the study population may not be representative of this source population which severely limits the external validity of results for all vaccinated/infected people.
- Naturally, subjects who died due to previous SARS-CoV-2 infection were not included in the study. Yet, without information on morbidity and mortality and contribution to the spread of SARS-CoV-2 from the primary infection, the results of the study are not informative for the question whether people without previous SARS-CoV-2 infection should be vaccinated or await natural infection.
- All three study groups – vaccinated or infected at baseline (28th of February) – were established upon future information (no infection, no additional vaccination after June 1, 2021), which severely limits the use of the results for today’s decision making.

In regard to selection bias:
- People with a SARS-CoV-2 infection between February 28, 2021 and June 1, 2021, or those who received a first (infected group) or third vaccine (vaccinated group) between February 28, 2021 and August 14, 2021 were excluded from this study. Thus the study population of group 2 consists of previously infected people that do not take the opportunity to receive a booster vaccine, which may well be the less vulnerable people with a lower baseline risk of getting infected/hospitalized. This would bias the estimate in favor of the infected group.
- Similarly, though at a smaller scale, people who died from COVID were not included in the analysis. This decreases the vulnerability of the infected group for secondary infections and/or hospitalization. This too would bias the estimate in favor of the infected group.

In regard to information bias:
- A difference in willingness to test between the vaccinated and previously infected group can result in biased estimates. Vaccinated people may be more on guard in regard to COVID-19 symptoms (especially if they adhere less to regulations because they are vaccinated) and will be tested more frequently. This can bias the estimate, again in favor of the infected group. However, this form of bias should not have affected the outcome hospitalization due to COVID-19, for which differences had the same direction. Yet, the number of those endpoints was low, limiting statistical power.

In regard to confounding:
- The authors acknowledge absence of information about health behavior, such as social distancing and masking. If the vaccinated group would adhere less to these preventive measures due to a sense of safety, this would also bias the estimates in favor of the infected group.
- A potential important aspect is the young average age (36 years) of the study population. As they were all fully vaccinated before February 28th, we thought that a large proportion may have been health care workers, who have a higher chance of exposure to SARS-CoV-2, and thus infection after vaccination. This would also bias the estimate in favor of the infected group.

We have scrutinized the paper in search of the fatal flaw; the one major methodological limitation that could explain the extreme effect in favor of the infected group, as reported. We conclude that it is not there, as we don’t think that any of the above biases can explain all of the effect. However, we did found several weaknesses that each have the potential to yield a modest bias, all in the same direction. Five modest biases may yield a large effect estimate. We, therefore, consider the question whether natural immunity provides better protection than full vaccination with Pfizer/BioNTech’s COVID vaccine remains unanswered.

The authors (Annemarijn de Boer, Valentijn Schweitzer, Marc Bonten and Henri van Werkhoven, all at University Medical Center Utrecht) acknowledge all other journal club participants for their time dedicated to discussing the paper.

Basically, common sense suggests that a layman does not make decisions about their own health against the advice of medical professionals based on one study which has not yet been peer-reviewed.

You're free to do what you want, but I suppose you'll spend the next month paying $100 for a test each time you and your family want to participate with the rest of society in non-essential activities, which is fine. I don't necessarily see the value or logic in any of your decision making, but you're confident, so enjoy.