Abstract Mental and substance use disorders are the leading cause of disease burden (disability-adjusted life years, DALYs) in Belgium. Strong mental health policies are needed to reduce the burden, taking into account effectiveness and cost-effectiveness of possible interventions. Next to providing an overview of the economic burden of mental health conditions (what is), routine cost-of-illness studies also serve as a source of information for modelling the cost-effectiveness of policies and interventions to improve the mental health of populations (what could be). In the context of an extension of the Belgian Health Status Report and the Belgian National Burden of Disease Study, we calculated the direct medical costs of a range of health conditions, including mental health conditions such as depression. We used self-reported diagnostic data from the Belgian health Interview Survey linked with claims data to calculate the incremental costs attributable to mental health disorders with g-computation. We correct for confounders such as age and sex, and also look at uncertainty brought about by the data source and modelling. This presentation aims to share the experience of calculating the economic burden of mental ill-health in the context of a routine cost-of-illness study. We will present the results of the study and discuss alternative ways in which the costs of common mental health conditions such as depression could be estimated, for example based entirely on claims data using pharmaceutical cost groups.