We matched urbanization level between cases and controls
in effort to better match cases and controls in terms
of unmeasured neighbourhood socioeconomic characteristics.
We assigned the date of their first use of medical
services occurring during the index year as the index date
for controls.