Epidemiological evidence in forensic pharmacovigilance.

Until recently epidemiological evidence was not regarded as helpful in determining cause and effect. It generated associations that then had to be explained in terms of bio-mechanisms and applied to individual patients. A series of legal cases surrounding possible birth defects triggered by doxylamine (Bendectin) and connective tissue disorders linked to breast implants made it clear that in some instances epidemiological evidence might have a more important role, but the pendulum swung too far so that epidemiological evidence has in recent decades been given an unwarranted primacy, partly perhaps because it suits the interests of certain stakeholders. Older and more recent epidemiological studies on doxylamine and other antihistamines are reviewed to bring out the ambiguities and pitfalls of an undue reliance on epidemiological studies.