The Lancet didn’t say Covid-19 vaccines are only 1% effective

A post on Facebook, apparently discussing the Covid-19 vaccines, claims: “Imagine a product expected to be taken multiple times by billions of people…that was presented in a way that led those people to believe its effectiveness was 98% good… The reality was that the product’s effectiveness dropped from 98% to just 1% on a much wider younger and fitter population that didn’t actually need it. This fact was then displayed in a specialist publication (The Lancet) explaining these shortcomings but the mainstream media and government did not share that vital information”.

Though it’s not totally clear, the Facebook post’s claim seems to be that the Covid-19 vaccines are not 98% effective but just 1% effective, supposedly according to the medical journal the Lancet. 

This is not true. Confusion over the vaccines’ effectiveness seems to have stemmed from an April 2021 comment piece in the Lancet which suggested that the trials publish an additional measure of how effective vaccines are, called the absolute risk reduction.

How effective are Covid-19 vaccines?

Covid-19 vaccines approved for use in the UK do effectively prevent symptomatic Covid-19, as well as serious disease, hospitalisation and death. 

Vaccine effectiveness is measured by comparing rates of disease (or other endpoints like hospitalisation or death) in vaccinated vs unvaccinated individuals. It can be estimated by conducting clinical trials or looking at real world data (observational studies).

In the initial trials of the vaccines, which were conducted largely before the widespread emergence of variants of concern, the Pfizer and Moderna vaccines were both found to be over 90% effective at preventing Covid-19, and AstraZeneca was 76%.

However vaccine effectiveness against symptomatic Covid-19 does appear to wane over time, and new variants have also been a factor. The UK Health Security Agency says: “Vaccine effectiveness against symptomatic disease with the Omicron variant is substantially lower than against the Delta variant, with rapid waning. However, protection against hospitalisation remains high, particularly after 3 doses.”

The latest data from the UK Health Security Agency looking at the effectiveness of vaccines against symptomatic Covid-19 infection found that: “After 2 doses of the AstraZeneca vaccine, vaccine effectiveness against the Omicron variant starts at 45 to 50% then drops to almost no effect from 25 weeks after the second dose. 

“With 2 doses of Pfizer or Moderna effectiveness dropped from around 65 to 70% down to around 15% by 25 weeks after the second dose. 

“Two to 4 weeks after a booster dose of either the Pfizer or Moderna vaccine following an AstraZeneca or Pfizer primary course, effectiveness ranges from around 60 to 75%, dropping to almost no effect from 20+ weeks after the booster. Vaccine effectiveness estimates for the booster dose are very similar, irrespective of the primary course received.”

Where does the 1% effectiveness claim come from?

The Facebook post claims the effectiveness of the vaccines “dropped from 98% to just 1%” and that “this fact was then displayed in a specialist publication (The Lancet)”. This appears to refer to a comment piece (that was not peer-reviewed) that we have written about in detail before.

The Lancet piece compared the conventional way of measuring the effectiveness of a vaccine, which is called ‘relative risk reduction’, with a different measure called ‘absolute risk reduction’. Relative risk reduction describes how much less your risk, in this case of getting a symptomatic Covid-19 infection, is if you’re vaccinated relative to someone who isn’t vaccinated, measured in percent.

So, hypothetically (these are not the actual figures), if the risk for an unvaccinated person was 1% and the risk for a vaccinated person was 0.2%, the relative risk reduction would be 80% (as 0.2% is 80% less than 1%).

But absolute risk reduction compares the difference between the proportion of people who got sick in vaccinated and unvaccinated groups. 

So in the same imagined example, the absolute risk reduction of vaccination would be 0.8 percentage points, as 0.2% (the risk for a vaccinated person) is 0.8 percentage points lower than 1% (the risk for an unvaccinated person).

The Lancet article compared the relative risk reductions for the vaccines in the trials (the figures given above of 90%+ and 76%) with absolute risk reductions, which it says were “1.3% for the AstraZeneca–Oxford, 1.2% for the Moderna–NIH, 1.2% for the J&J, 0.93% for the Gamaleya, and 0.84% for the Pfizer–BioNTech vaccines”. The article was published in April 2021, so predates the detection of several variants of concern.

These figures don’t undermine the fact that the vaccines are effective, particularly in the short term. Instead they show that people in those trials had a low baseline risk of getting ill with Covid, but the vaccines still took away most of that risk.

The article wasn’t arguing that the relative risk reduction wasn’t, say, 95% for the Pfizer vaccines in trials, but instead that the lower figure should be published alongside the higher one. 

In fact, one of the article’s authors, Professor Piero Olliaro, told Reuters when asked about misleading claims using the article as evidence that “bottom line: these vaccines are good public health interventions”, adding “we do not say vaccines do not work”.

Image via Daniel Schludi