Three vaccine quirks that aren’t what they seem

Why dynamic health problems can be so counterintuitive

Last month, a new study estimated that COVID vaccines had directly saved around 1.6 million lives between late 2020 and March 2023. In reality, this number is likely to be even higher, because the study only accounted for the direct effect of the vaccine (i.e. protecting from death if infected) rather than any indirect effects (i.e. reducing transmission in the population and hence risk of infection in the first place).

And yet, despite such evidence, some prominent politicians, journalists and podcasters continue to misinterpret the role vaccines played during the COVID pandemic. In some cases, this had led to bold but false claims, such as the idea that COVID vaccines had no effect, or even that they caused more deaths than they prevented.

Unfortunately, vaccine rollouts – like any intervention during a dynamic epidemic – can sometimes lead to counterintuitive patterns. To illustrate the challenge, I’ve outlined three common quirks below. See if you can spot what might be happening, then read on for the explanations.

1.     How can a vaccine be effective if most people who die have been vaccinated?

2.     If most disease deaths occur after a vaccine rollout, how can the vaccine have had an impact?

3.     If 90% of the highest-risk groups have been vaccinated, why won’t future epidemic waves be 90% smaller than earlier ones?

Now you’ve had time to think, let’s go through each one.

1. How can a vaccine be effective if most people who die have been vaccinated?

We saw this question asked a lot in early 2021, after the rollout of the COVID vaccine. Intuitively, we might expect most people who die of a disease to be unvaccinated if a vaccine is protective against that disease. But we need to remember that two things influence the proportion of deaths (or cases) that are vaccinated: the effectiveness of the vaccine, and the proportion of the population at risk who are vaccinated.

During the early COVID vaccine rollout, older groups – who were at higher risk of severe outcomes – were prioritised in many countries. And in countries like the UK, vaccine coverage was very high. That meant the small proportion of unprotected vaccinated individuals who got severely ill ended up being larger than the group of unprotected unvaccinated individuals in this category.

As an extreme example, suppose vaccine coverage was 100%. The COVID vaccine was very good, but not perfect, at preventing death. So by definition, in this scenario there would be some deaths and all of them would have been vaccinated. It doesn’t mean the vaccine doesn’t work at all; it just means that coverage is very high among those most at risk.

2. If most disease deaths occur after a vaccine rollout begins, how can the vaccine be effective?

Earlier this summer, a paper on excess COVID deaths during 2020–22 fuelled speculative headlines such as ‘Covid vaccines may have helped fuel rise in excess deaths’.

The central claim in the paper was as follows: ‘excess mortality has remained high in the Western World for three consecutive years, despite the implementation of COVID-19 containment measures and COVID-19 vaccines. This is unprecedented and raises serious concerns’. Although the paper included no new analysis of vaccine side effects, it was heavy on speculation, and the journal that published it has since issued an expression of concern.

Unfortunately, there was a much simpler explanation for why there was so much mortality in the post-vaccine era: relatively few people got infected before the introduction of vaccines.

Take the UK as an example. In February and early March 2020, the COVID was spreading against a background of pretty much all the population being susceptible. Because of massive changes in behaviour that spring, transmission came down, but it meant lots of people remained unexposed, and hence lots of susceptibility remained.

The same basic dynamic happened in late 2020. There was lots of transmission, which meant rising infections. Then there was massive behaviour change and the introduction of strict control measures, which meant by early 2021 there were still lots of people who hadn’t been exposed yet, and therefore remained susceptible.

Then vaccination rolled out widely, reducing susceptibility to severe disease in particular. However, because so few people had been exposed early on, the number of people at risk who were exposed during the 2021-22 waves in European countries was generally larger than the number of people at risk exposed in early 2020, even accounting for the protective effects of vaccines.

The underlying problem was that people got anchored to the idea that the spring 2020 waves were the ‘big ones’, rather than being waves in which a relatively small proportion of people were exposed. In other words, the ‘big ones’ were still to come. In the end, almost the entirely of the UK would be exposed across multiple variant waves, albeit with vaccines protecting most people against the most severe outcomes.

3. If 90% of the highest-risk groups have been vaccinated, why won’t future epidemic waves be 90% smaller than earlier ones?

This was a question I got asked a lot in spring 2021, during the COVID vaccine rollout. Why not go back to normal earlier, even if it means an earlier epidemic wave? The answer follows a similar reasoning to the previous section. Although 90% of the highest-risk group has been vaccinated in this scenario, the earlier epidemic waves weren’t necessarily reflective of the maximum number of people who could have been exposed (because behaviour change and control measures had prevented widespread infection in many countries).

To get some intuition into the dynamics, which of the following scenarios would you expect to cause more disease?

1. Everyone in a population is susceptible to disease and 5% get infected

2. 10% of the population is susceptible to disease, and 50% get infected

In the first scenario, we’d expect 5% of the population to get ill; in the second we’d expect 5% (= 50% of 10%). In other words, a large post-vaccine wave could be just as bad as the earlier wave in this situation.

‘Obvious’ isn’t always correct

Despite COVID dominating the media and political landscape for years, there are still public figures who misrepresent and muddle pandemic data and dynamics. As the above illustrates, we can get counterintuitive patterns when a dynamic epidemic interacts with a vaccine rollout.

There’s a lot still to learn about COVID and how best to deploy pandemic vaccines in future, but we have to start by getting the basics right.