A number of papers relating to COVID-19 Vaccination refer to the occurrence of thrombotic events.
Rates of occurrence are usually expressed in the form of "cases of thrombotic events per million person vaccinated-days."

I am familiar with the concept of events per million but not events per million person vaccinated days.

What does the term "person vaccinated days" mean?

Example here -
"Thrombotic complications of vaccination against SARS-CoV-2: what pharmacovigilance reports tell us – and what they don't"

  • The study by Smadja et al. confirms the rarity of possible thrombotic complications in association with COVID-19 vaccination, reporting only 0.21 [95% CI 0.19–0.22] cases of thrombotic events per million person vaccinated-days.

"Person vaccinated-days" is just expressing units of people*days; if one person was vaccinated 7 days ago and another 3 days ago that would be a total of "10 person vaccinated-days".

  • 1
    While that would be what I would take from such a construct in cases where there was a clear and logical link - eg worker accident_free-days, this would seem to be a very misleading measure here. Eg if 1000 people had no thrombotic events after 10 or 100 days then the p a-d measure would climb from 10000 to 100000 while still only telling us that none of the 1000 had been affected. One may intuitively expect any events to occur if at all within say 10 days (or 1) but this measure simply keeps increasing and swamps occasional occurences. Aug 17 at 23:29
  • 2
    @RussellMcMahon If you look at the original study, they've based it on 90 days of reporting, but not everyone is vaccinated at the same time. It seems they've chosen this denominator because they are limited by the data source they have; they probably would not have designed a prospective trial this way. I think these numbers would be best used in comparison to overall rates of thrombotic events in unvaccinated people. If you prefer, they have 2161 events among 361734967 individuals. Aug 17 at 23:47
  • 1
    @RussellMcMahon: it looks like the author tried to apply the concept of "exposure days" (which is used in vaccine trials with respect to the pathogen) to a somewhat inappropriate context of side-effects. For the latter, exposure events, i.e. just vaccine administration, is typically more relevant.
    – Fizz
    Aug 18 at 8:01
  • 1
    @RussellMcMahon: on the other hand, better thrombosis studies usually try to quantify the risk as dependent on the time since the exposure to the risky/initial event. So taking time after the exposure into account isn't unreasonable, but assuming equal odds per such day probably is wrong.
    – Fizz
    Aug 18 at 8:21

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.