A new study from India has been reported claiming to show that the AstraZeneca vaccine offers 97% protection against Covid infection. The observational study involved healthcare workers who reported to Delhi’s Indraprastha Apollo Hospital with symptomatic COVID-19 after having been vaccinated using the Covishield (AstraZeneca) vaccine.
According to the newspaper report, there were a total of 3,235 healthcare workers in the study, all of whom had received at least one dose of the vaccine. Eighty five of them experienced symptomatic Covid, of which 65 were fully vaccinated and 20 were partially vaccinated. 0.06% were hospitalised, which appears to be two people. No one was admitted to intensive care or died.
This seems good. However, by itself it tells us almost nothing about how effective the vaccines are. Most crucially, there is no control group of unvaccinated people for comparison. We also don’t know how many of those vaccinated had a previous infection or exposure which would have provided some immunity – these are healthcare workers after all. Also, although this is India we don’t know which variant anyone was infected with.
Unfortunately, we can’t look into this in more detail as the study does not appear to have been published anywhere. Instead all we have is a newspaper report based on a press release.
The figures provided in the newspaper report also don’t make sense. It says:
The study covered 3,235 healthcare workers (HCWs). A total of 85 of the 3,235 HCWs acquired the SARS-COV-2 infection during the study period. Out of these, 65 (2.62%) were fully vaccinated, and 20 (2.65%) were partially vaccinated.
What does the 2.62% refer to? 65 is 2.62% of 2,480.9, but where does 2,480.9 come from? Again, 20 is 2.65% of 754.7, but what is that number?
Eighty five is 2.63% of the 3,235 healthcare workers, which is close to 2.62% and would (almost) correspond to the headline of 97.38% protection. But the percentages in the brackets have no obvious relationship to the figures they accompany.
This is yet another example of science-by-press-release and shows again why it is such a poor way to present findings. Yet the results have appeared in newspaper headlines around the world, despite no one being able to read the actual study and the shortcomings of the uncontrolled design and the confused and incomplete reporting of the data.
Nonetheless, it does seem likely that immunity from infection or vaccination will also work against the Indian variant. But science by newspaper report is no way to show it or reassure doubters.