The cost to the National Health Service of treatment for clinical depression for England and Wales has been estimated as being in the area of 416 million pounds (1990 price level), and the social burden in terms of increased morbidity and mortality due to depression is known to be considerable. Prescription of antidepressants is the most common treatment for people with clinical depression. The majority are diagnosed by general practitioners who issue 95% of all prescriptions for antidepressants. In 1992 the English National Health Service spent 81.1 million pounds on antidepressant drugs. However, the understanding of the disease process, the health, and economic impact of various treatment options are surrounded by much uncertainty. Few cost-effectiveness studies of antidepressive treatments can be found in the literature. They are often based on small sample sizes, a short time horizon, and a narrow focus on subjective measures of process or intermediate outcome and are therefore less than robust when generalized to a wider population of patients with clinical depression. We have developed a stochastic simulation model aimed at analysing cost-effectiveness aspects of treatment for depression and have used it to test the consequences of a range of treatment policies. Features of the model are discussed in this paper. The model is described with a flow chart that shows patients' pathways through the health-care system. Based on the incidence approach, the model simulates a cohort of patients, using decision and chance nodes which occur during treatment. A range of critical decision variables can be changed to assess the consequences for cost-effectiveness.