Background: The care of the severe mentally ill who need intensive support for their daily living (dependent persons), accounts for an increasingly large proportion of public expenditure in many European countries., Aims of the Study: The main aim of this study was the design and implementation of solid methodology to calculate unit costs of different types of care. To date, methodologies used in Spain have produced inaccurate figures, suggesting few variations in patient consumption of the same service., Methods: An adaptation of the Activity-Based-Costing methodology was applied in Navarre, a region in the North of Spain, as a pilot project for the public mental health services. A unit cost per care process was obtained for all levels of care considered in each service during 2005. The European Service Mapping Schedule (ESMS) codes were used to classify the services for later comparisons. Finally, in order to avoid problems of asymmetric cost distribution, a simple Bayesian model was used., Results: As an illustration, we report the results obtained for long-term residential care and note that there are important variations between unit costs when considering different levels of care. Considering three levels of care (Level 1-low, Level 2-medium and Level 3-intensive), the cost per bed in Level 3 was 10% higher than that of Level 2., Discussion: The results obtained using the cost methodology described provide more useful information than those using conventional methods, although its implementation requires much time to compile the necessary information during the initial stages and the collaboration of staff and managers working in the services. However, in some services, if no important variations exist in patient care, another method would be advisable, although our system provides very useful information about patterns of care from a clinical point of view., Conclusions: Detailed work is required at the beginning of the implementation in order to avoid the calculation of distorted figures and to improve the levels of decision making within the Health Care Service. IMPLICATIONS FOR HEALTH CARE POLICY AND FORMULATIONS: As other European countries, Spain has adopted a new care system for the dependent population. To finance this new system, reliable figures must be calculated for each type of user in order to establish tariffs or public prices. This study provides a useful management tool to assist in decision making., Implications for Future Research: The methodology should be implemented in other regions of Spain and even in other countries in order to compare our results and validate the cost system designed.