Background: Mental disorders are highly prevalent and are associated with substantial disease burden, but their economic costs have been relatively less well researched. Moreover, few cost-of-illness studies used population-based psychiatric surveys for estimating direct medical, direct non-medical and indirect costs, and were able to do so for several well diagnosed mental disorders.
Aims: To calculate the cost of nine common mental disorders. The costs were calculated at individual level (per capita costs), and at population level per one million population for both prevalence (current cases) and incidence (new cases).
Method: Data were derived from the Netherlands Mental Health Survey and Incidence Study (Nemesis), a population-based psychiatric cohort study among 5,504 adults in the age bracket of 18-65 years. DSM-III-R disorders were assessed with help of the Composite International Diagnostic Interview (CIDI). The costs of health service uptake, patients' out-of-pocket costs, and production losses were calculated for the reference year 2003. Robust regression methods, with 1,000 bootstrap replications, were used to estimate the excess costs of the distinct mental disorders and their 95% confidence intervals, while adjusting for physical illnesses and concurrent mental disorders in the regression equation.
Results: The annual per capita excess costs of the mood disorders (5,009 euros) were higher than those of the anxiety disorders (3,587 euros) and alcohol-related disorders (1,431 euros). Being more prevalent, the excess costs of anxiety disorders are higher than those of mood disorders at population level. The annual influx of new cases (incidence) accounts for 39.2% of the costs at population level. It appeared that in the general population, in the productive age of 18-65 years, the bulk of the costs (85%) were related to production losses.
Discussion: The study has some strengths and limitations. The data were derived from a large and representative population-based sample. Disorders were assessed with a reliable instrument. The costs were comprehensive in that they included direct medical, direct non-medical and indirect costs. The costs attributable to mental disorders were obtained with robust regression models while adjusting for the presence of somatic illnesses. For several reasons the costs figures must be seen as conservative lower bounds of the true costs. (i) People who were hospitalised were likely to be underrepresented in the sample, and it is well known that hospitalisation is one of the major cost drivers. (ii) Resource use was based on self-report, and this is likely to have resulted in underreporting. (iii) Work loss days were included in the analysis, but work cutback data were unavailable, while it is known that the costs due to work cutback can be substantial.
Implications: (i) The costs of mental disorders are comparable to those of physical illnesses. This throws some light on the allocation of budgets for research and development in mental versus physical illnesses. (ii) At population level a substantial part of the costs are caused by new cases, and this is a strong argument for strengthening the role of preventive psychiatry in public health with the aim to reduce incidence and avoid the future costs. (iii) In particular, reducing the incidence of major depression, panic disorder, agoraphobia and dysthymia should be considered as public health priorities, because these disorders are associated with substantial disability, and have, in addition, important economic ramifications. (iv) The bulk of the costs are due to production losses; this makes employers pertinent stakeholders in mental health promotion, and thoughts should be given to the question how to involve them more actively in health promotion. (v) It is well to emphasise that adoption of the above mentioned policies will require that first more prevention trials and cost-effectiveness studies are conducted in the selected disorders.