Vaccine skepticism is not anti-science

There is a widespread idea that skepticism about claims regarding covid vaccine efficacy is anti-science. But this is not true — it is pro-science.

Science is about demonstrating things through objective studies. Science is not going by claims of authorities.

I am thinking about claims such as that:

• Covid vaccine efficacy against severe illness and death is high (90%+)

• The vaccine is net beneficial for people under the age of 21.

• The vaccine has high efficacy for people who have been previously infected.

These claims may be true, but they have not been shown by scientific studies. 

What we have instead are many people with expertise making an assessment and asserting that these claims are true. But science is not assessment from experts.

The main scientific issue is that we do not have that much data from randomized control trials (RCTs). Some limitations of the RCT data we have about the covid vaccines:

• The studies were conducted with a different covid variant than the one currently in circulation.

• None of the participants were previously infected with covid.

• There was no significant difference in the number of deaths between the case and control groups (possibly due to low sample size).

• The efficacy was not demonstrated for young people.

What we have instead is a lot of observational data. But observational data is fraught with issues, as everyone who has worked with it knows. This is a reason that nutritional science is hard, and that we have had so many conflicting messages over time.

One omnipresent problem with observational studies is selection effects. The people who choose to get vaccinated are not the same as those who choose not to, even before anyone gets covid. This is a crucial point. This figure shows how large the difference can be:

There is no solid way to get around this issue. You can try to control for various variables. But this is exactly what nutritional science has attempted over the decades, leading to misleading conclusions such as the advice to avoid eating eggs. More details in this post: Why controlling for variables is insufficient.

Once we are considering questions like the efficacy of people below 21 years of age, or the efficacy for previously infected people, the foundation from observational data becomes even more shaky.

A thing we can then do, is have experts make informed guesses. Eg, they can come with a statement like this: "Giving the totality of the data, our estimate is that it is likely that the efficacy of the covid vaccines in 50+ years old is reasonably high, and very worth taking." 

And this is fine. The informed guesses of experts can be very helpful in a situation where people have to make a decision. It is a type of useful information. But it is not science.

Here are some lessons we should learn for the future:

• Observational data is hard to use. (We should know this already, but even so.)

• We should not claim that things are backed by science and force people to comply without solid scientific evidence.
  (Here I am thinking especially about the more doubtful claims, such as those of strong cost-benefit balance in 18 year olds, or large benefit to previously infected.)

And most importantly:

• We should take more seriously the challenge of acquiring solid scientific data and knowledge.

In this case that would mean performing more and stronger random control trials, able to answers the various scientific questions.

Such as:

• A longer running RCT, to investigate longer term cost-benefits.

• The initial RCT could include more elderly and frail people, to have the numbers to definitively show that the vaccine can protect against death.

• A large RCT of previously infected individuals to estimate efficacy for this group.

And so on.

We know how to gain real scientific knowledge. We just need the willingness to do so. And to stop shaming people who criticize the scientific basis of our current dogma.