Rheumatic Disease Unit, Wellesley Hospital, Toronto, ON, Canada.
This paper provides an overview of the epidemiological techniques used in pharmacoepidemiology. Definitions, advantages and disadvantages of the various types of study design are presented and potential sources of bias are discussed. An important challenge and the source of major controversies is the evaluation of cause and effect relationships when exposure to a drug is associated with the occurrence of an adverse event; we present an established set criteria suggesting inferences about causation.