January 18, 2024 - 10:05am

What are we to make of research published in The Lancet this week looking at the impact of vaccination on severe Covid-related outcomes?

The paper is based on real data, so cannot be compared to discredited Neil Ferguson-style modelling of imaginary infection rates. The authors use a unique population level dataset and are careful to control for a number of confounding factors that might affect outcome, such as age, ethnicity and underlying health risks. As with any empirical paper, questions can be asked about the methods, underlying assumptions and robustness of the results. But the real story seems to be the policy implication the authors draw from their results, which have since been reported in the British national press.

The paper concludes that had the whole population of England and Scotland been fully vaccinated (defined as receiving the recommended dosage for each age), then severe Covid-related outcomes (hospitalisation or deaths) in the summer of 2022 would have been reduced by about 7,000. The authors then argue that this justifies public health interventions to increase coverage and to “tackle vaccine misinformation in a more direct fashion”.

To see if such a conclusion is justified, let’s look at the group with the highest rate of “undervaccinated” people: children aged 5-15. By June 2022, out of the 8.1 million children in the dataset, just 15% (1.2 million) had received the recommended number of doses (one for those aged 5-11, two for 12-15s). Given the very low mortality risk from Covid for children, virtually all the serious outcomes in this group will have been hospitalisations.

The paper estimates that having two fewer doses than recommended more than doubled the risk of Covid-related hospitalisation. Interestingly, having just one fewer dose than recommended had no significant effect in the main results. That means that for 5 to 11-year-olds (for whom only one dose was recommended), the study finds no significant impact of vaccination on hospitalisation.

If all children aged 5-15 in England and Scotland had been fully vaccinated, the authors estimate, there would have been 210 fewer Covid-19 hospitalisations in this group between June and September 2022. But to achieve that would have meant giving at least one more dose to 6.9 million children. In other words: for every hospitalisation avoided, an extra 33,000 children would have required vaccination. This estimate of the “number needed to vaccinate” (NNV) to avoid a single hospitalisation is consistent with the figures published by the ONS last October.

The Lancet study confirms that the risk of hospitalisation (and hence any potential benefit from being fully vaccinated) is dramatically lower for children who do not face a particular clinical risk. While the report doesn’t estimate the impact of vaccination specifically for those not at risk, the earlier ONS study estimates the NNV to avoid a single hospitalisation is over 10 times higher for adolescents in a clinical risk group, compared to those who are otherwise healthy.

All this suggests that fully vaccinating 1.2 million children may have resulted in about 36 fewer hospitalisations over a four-month period, but that most of this reduction is likely to have been limited to those in a clinical risk group.

On this basis, does vaccinating children pass any kind of cost-benefit test? This depends not only on the monetary cost of vaccination but also the risk of side effects, including hospitalisation.

Firm data on hospitalisation due to Covid-19 vaccination is hard to come by, but in the US the CDC V-Safe monitoring system reports that between 1 in 5,000 and 1 in 2,500 children (depending on age group and dose) were hospitalised in the seven days after being vaccinated. Applying this to the 1.2 million fully vaccinated children in the Lancet study would suggest between 240 and 480 additional non-Covid hospitalisations, dwarfing the estimated reduction in hospitalisation for Covid-19. 

By focusing only on hospitalisation, and ignoring other short and long-term side effects, the risk-benefit ratio does not come close to justifying the rollout of Covid-19 vaccination to healthy 5 to 15-year-olds. Even for those in a clinical risk group, the benefit-risk ratio is at best marginal. 

By its own logic, the data in the Lancet paper suggests that rolling out the vaccine to all children over the age of 5 was a policy error. If anything it led to more, not less hospitalisation.


David Paton is a Professor of Industrial Economics at Nottingham University Business School.

cricketwyvern