Epidemiological assessment of drug-induced disease.

Adverse reactions are a potential concern for physicians when they prescribe or recommend drugs. Epidemiological principles, when combined with clinical judgment, can be of help in this situation, starting with an appreciation of the strengths and weaknesses of different sources of information on adverse reactions--clinical trials, case reports, and formal epidemiological studies. The latter studies generally provide the most comprehensive information on the risks of serious adverse drug reactions. An understanding of the different types of risk estimates, relative and absolute, is also needed--we stress the value of the absolute risk as the best measure of the impact of an adverse reaction. Rare serious reactions, although striking, have little impact on individual risk, whereas more common reactions, even with much lower fatality rates, are more likely to lead to adverse outcomes for patients. The importance of balancing risks and benefits, taking into account all the information about an individual patient's risk profile, is also highlighted.