Vaccines Aren't Safe (or Unsafe)
How the binary mental model ‘safe’ vs ‘unsafe’ influences our perception of vaccine safety.
Binary classification
One and a half years into the Covid-19 pandemic there is still considerable vaccination hesitancy. A popular and reasonable argument goes like this
“We don’t know if the Covid vaccine ABC is safe.”
In this post I want to draw your attention to the fact that this simple question looks like a yes/no type question when in fact it really isn’t.
The human brain excels at classification, the task of ingesting complex information and turning it into simple categories. Or as we say in German, to ‘put things into drawers’.
The simplest classification is binary: Do this, but don’t do that. In fact, our brain is so good at simplification that it can go into overdrive, and a persistence of this pattern of thinking can become malign, cf. 0-1 thinking.
When we ask
“Is this vaccine safe?”
we are expecting a binary classification. We expect the vaccine to be safe, or unsafe. In reality however, safety is a continuous variable ranging from not safe at all to very safe and all shades in between. A better mental model is to think of risk instead:
Low risk (safe) - medium risk - high risk (unsafe).
Estimating risk
Still, the true risk of a vaccine is not known.
Circling back to the example of vaccine safety. When the
Vaccines don’t became safe from one day to the next [1], in fact they don’t change at all from phase 1 trial to marketing approval. What changes over time is the estimation of the safety and the associated uncertainty. With more information from phase 1-4 trials coming in uncertainty decreases.
Let’s make a simple example. Assume the safety of vaccine ABCVaX is in fact 9/10 (with 10 being super safe). However ABCVaX is a completely new vaccine so nobody knows yet that it is in fact a relatively safe vaccine.
In order to receive marketing approval the vaccine has to be at least 8/10 safe. Now first data from phase 1-3 trials comes in. The authorities review the data and crunch the numbers and find out its roughly 9.1 safe, but there is considerable uncertainty left with a 95% confidence interval of 8.5-9.7.
Because there is still uncertainty around the true risk, the body will likely not issue marketing approval. Not because the vaccine isn’t safe, or better, safe enough, but because we do not have enough data to prove the vaccine is safe enough.
In the real world, the body might still allow the vaccine for emergency use, because relative to the disease the vaccine is safe enough.
Where can binary thinking lead us astray
The novel Covid vaccines aren’t 100% safe and therefore vaccinations are a human experiment.
Let’s break this statement down. First, as we have discussed before 100% safety does not exist. Even after phase 3 trials have been completed the vaccines will still not bet 100% safe. In fact any medication today comes bundled with the risk of adverse effects. Interestingly enough, so do homeopathic treatments [2]. The second statement is just a variation of the first - if we set the bar high enough any medication is a ‘human experiment’, whatever that means. So what sounds reasonable at first turns about to be a truism.
The continuous heuristic is still a heuristic
This new continuous mental model is better than the binary one, yet it still remains a heuristic.
Vaccines aren’t safe (or unsafe)
black and white grey safety changes
Known cognitive bias here 0-1 thinking.
[1] in real terms, in legal terms they might [2] for a massive rabbit hole google Nocebo