Influence of control selection in nested case‐control studies: the example of exposure to antiretroviral treatment (ART) and the risk of myocardial infarction (MI) in the French Hospital Database on HIV (FHDH‐ANRS CO4)