Reference population
When women start using HC, some will within days, weeks, or months
experience mood deterioration or even depression due to that use, and
will, therefore, stop using these products. These former users are at an
increased risk of depression – not due to their former use of HC, but
that use was the first test to reveal their mental sensitivity. In
addition, during the study period some users of HC will develop mood
deterioration without getting a prescription of anti-depressants or a
depression diagnosis, e.g., those seeking psychological therapy, and
will stop using HC. Both of these groups of previous users will be at an
increased risk of depression and use of antidepressants. By including
those women in the reference group of non-users, will underestimate the
risk of depression development in current users.
In pharmacoepidemiology it is default practice to compare exposed women
with non-exposed women. For rare outcomes it doesn’t make any difference
whether non-exposed or never exposed are used as reference population.
But with a frequent outcome, such as depression development, it makes a
huge difference, which was illustrated in study 1, where the risk of
depression with use of oral contraceptives changed from 23% increased
risk to 70% increased risk with change in reference group from
non-users to never-users. In study 3 the risk of depression with HC use
changed from 1% increased risk to 29% increased risk with the same
change in reference group.