Covariates
Generally, adjustments beyond age, calendar year and length of education
rarely make substantial changes in risk estimates in epidemiological
studies, including studies on risks and benefits of oral contraceptives.
Nevertheless, a missing potential confounder is often claimed as
invalidating a study if the message or results is disliked. An example
is the missing control for family disposition of depression in study 1
and 2. That variable was included in study 3 and turned out not to be a
confounder at all, indicating that when HC is prescribed, family history
of psychiatric diseases is rarely taken into consideration.