The Volokh Conspiracy

Mostly law professors | Sometimes contrarian | Often libertarian | Always independent

Preprint COVID Vaccine Study Withdrawn: Key Number Was Off by Factor of 25


Mistakes happen, especially in non-peer-reviewed studies (though they also happen in peer-reviewed ones); but it's a good reminder to be skeptical. From Reuters:

A pre-print study which claimed that there is a 1 in 1,000 risk of contracting a heart inflammation condition known as myocarditis has since been retracted due to a calculation.

The study which was conducted by researchers at The University of Ottawa Heart Institute, was featured in numerous blogs and social media posts as proof that the COVID-19 vaccine is unsafe for use….

The study had calculated an incidence rate of myocarditis in the Ottawa region post-vaccination by dividing the number of occurrences of the heart inflammation condition over a two-month period (June and July 2021) in Ottawa (32) by the total number of vaccinations in the area (reported as 32,379 in the pre-print) [vaccinations here refers to vaccine doses -EV] ….

[But b]etween the week beginning May 30 and the week starting July 25, there had [actually] been 845,930 vaccines administered in the Ottawa region, according to data published by Ottawa Public Health ….

I should note that this error should have been apparent just by eyeballing things: The Ottawa area has nearly 1 million people, June and July were heavy vaccination periods in Canada (especially for the second doses), and 32,379 would thus be a visibly strange number of vaccinations for that period. Still, as I said, mistakes happen, even to the best of us.

The retraction is here; the original study is here. True to the rule that stories about corrections will have errors of their own, the Reuters article has been corrected, with a note:

Correction Oct. 4: corrects spelling of denominator in paragraph 13, and in the same paragraph corrects "25 orders of magnitude off" to "25 times smaller". Changes 'magnitude' to 'factor' in verdict.