Depression is a common, recurrent, disabling and potentially fatal disorder. Its effect on quality of life is more severe than that of some other chronic medical conditions. Its cost burden (direct and indirect) has been estimated at $US26 billion to $US43.7 billion in the US (in 1990), and 3.4 billion pounds sterling in the UK (in 1992). With contemporary levels of diagnosis and treatment, and of treatment failure, the indirect economic losses to society from depression-associated morbidity and mortality may be up to 7 times the direct costs, with 69 to 98% of these costs related to morbidity. Impaired capacity while at work may equal absenteeism in terms of costs. Depression carries an increased risk of suicide and suicide attempts, both of which are costly. It is ironic, then, that the drugs used to treat depression account for 9% of poisoning deaths in England and Wales. The newer antidepressants are less toxic than the tricyclic antidepressants (TCAs). The physician's choice of antidepressant should be motivated primarily by clinical considerations, but cost implications are of increasing importance. In some healthcare systems, the expenditure associated with prescribing by doctors has been successfully modified, at least temporarily, by drug formularies, audit and feedback. This may be to the disadvantage of the selective serotonin (5-hydroxytryptamine: 5-HT) reuptake inhibitors; they have higher acquisition costs than the TCAs, since they are newer and lack cheaper generic forms, and definitive evidence of benefits may be inadequate or equivocal. However, market price is only one of many factors that should affect the prescriber, and the increasing trend towards cost containment policies is liable to lead to false economies.