I counted people as vaccinated from the day of the first dose. But there's a low number of deaths in the first few weeks after vaccination, so if I would've counted people as unvaccinated until 3 weeks from the first dose, it would've worked in the favor of unvaccinated people because it would've reduced the mortality of unvaccinated people and increased the mortality of vaccinated people: https://sars2.net/czech4.html#ASMR_when_people_are_classified_as_unvaccinated_until_three_weeks_from_vaccination.
You will have immortal time bias favoring the vaccinated when your definition of "vaccinated" depends on having survived the 3-4 weeks between the first and second doses. So counting people as vaccinated starting at the first dose is correct.
How are the vaccinated determined? Is there a time window when deaths are deemed to be "unvaccinated", thus skewing the results?
I counted people as vaccinated from the day of the first dose. But there's a low number of deaths in the first few weeks after vaccination, so if I would've counted people as unvaccinated until 3 weeks from the first dose, it would've worked in the favor of unvaccinated people because it would've reduced the mortality of unvaccinated people and increased the mortality of vaccinated people: https://sars2.net/czech4.html#ASMR_when_people_are_classified_as_unvaccinated_until_three_weeks_from_vaccination.
You will have immortal time bias favoring the vaccinated when your definition of "vaccinated" depends on having survived the 3-4 weeks between the first and second doses. So counting people as vaccinated starting at the first dose is correct.