Closer Look: How did they test the Covid-19 vaccines?
The results of the covid-19 vaccine clinical trials are very complex and can’t be easily interpreted. A lot of news stories are putting too much emphasis on percentages that are meant to be estimates and aren’t hugely accurate. So this article is about how those numbers are calculated, and what has been found for the different vaccines. If you want to know about what vaccine efficacy means for you as a person, and are wondering which is best, then read this article instead.
The Pfizer Clinical Trial
The percentages you’ve seen in the news come from clinical trials. If we take the Pfizer vaccine trial as an example , 43,661 people between 12 and 85 years old were enrolled. Initially a group were given two doses of the vaccine and another group were given the placebo, but eventually the placebo group was also given the actual vaccine. About 2000 people dropped out of the trial. They chose a range of people from around the world, almost equal split of males and females, and 42% were non-white. The trial had several objectives, most importantly to find out if the vaccine was safe, how severe any side effects are, and how many vaccinated people would be protected from covid-19 disease.
Here are the results:
Firstly, no one with the vaccine died from covid-19 during the study, so the vaccine has an effective immunity of 100%. Only 10 people in the whole trial got severe cases of covid-19, and just one of those was in the vaccinated group and the rest had received a placebo. So immunity against severe disease is potentially 90% (1/10 is 10%)… but they have since said the single severe case was not actually severe, so it could actually be 100% again. Medical trials are complex!
170 participants tested positive for covid-19 during the trial, where 162 cases were in the placebo group and 8 were from the vaccinated group. So we can say that the immunity to any disease is about 95% (because 8/170 is about 5%). They did not try to find out if any vaccinated people had caught covid-19 but were asymptomatic, so we still don’t know if the vaccine provides sterilising immunity.
One important note here, the numbers of ill people are all very small. We can see not many of the placebo group caught covid-19 during the trial, so we can guess that overall, not many vaccinated people in the trial were likely exposed to the virus. That means we can’t take these percentages at face value, they just give us a rough idea. The main point of the trial was about safety and seeing if there is a difference between placebo and vaccine, and we can see there’s a big difference.
What are the results from Israel about the Pfizer vaccine?
However, we now have a much bigger data set from Israel about the Pfizer vaccine . Over a million people that have been vaccinated have been followed closely and compared with unvaccinated people to understand about how the Pfizer vaccine is working in a real life population:
Amazingly, in this data set over 80,000 participants were over 80. We’d expect vaccines to be less effective in the very elderly, as their immune response is usually reduced as a result of aging. No one in Isreal has died from covid-19 following two doses, so they still can’t give an accurate answer about the efficacy because there is “a lack of data”!
One thing that this does tell researchers is that there is a big improvement in outcomes after the second dose. Understanding this is really important for governments and health officials in their vaccination strategy. For example, one big question is whether it is best to give full vaccination to everyone vulnerable or to prioritise getting as many people a single dose as quickly as possible. Researchers can use these numbers in mathematical models to make predictions about the best strategy.
What about other covid-19 vaccines?
It’s more difficult to summarise the other clinical trials in detail the same way as the Pfizer trial, as most of them were testing multiple things at once. Instead here is a summary table adapted from https://twitter.com/MonicaGandhi9
Big caveat here! It’s Against Science Law to compare the efficacy percentages from one trial to another. No trying to work out if one vaccine is better than another. Not allowed. These trials used different age groups, different countries, different doses, different delays between doses, different health conditions from participants, different numbers of people, and had different levels of covid-19 exposure. You’re not comparing like for like. The only way to do that is a trial which puts them head to head… which basically never happens because no one has enough money to fund this kind of thing.
So here’s a summary…but don’t use this table as a Top Trumps for Covid-19 vaccines.
The main conclusion is in the “protection from death” and “protection from hospitalisation” rows. It’s all brilliant news.
Preliminary data from Scotland and England for the population vaccinated with Pfizer and the AstraZeneca vaccines has strong evidence that the efficacy is very high. A single dose of Oxford AstraZeneca vaccine reduced a person’s risk of hospital admission by 94% after four weeks. A single dose of Pfizer vaccine reduced risk of hospitalisation by 85% between 28 and 34 days after the first dose, which is a longer delay than tested in Israel. 
We also can’t say for certain about asymptomatic spread of covid-19 between vaccinated people, which is why that final row is currently unknown. If you want to read more about how this could affect your behaviour if you are vaccinated, then see this article: