The demands for cost information in health services are many and various but the supply of such information is less than might be expected and is compounded by the expense of undertaking costs research. This paper examines a short cut to mental health research which, if certain rules and conventions are obeyed, can still produce valid costs data. By distilling evidence from previous research a reduced list of services is identified which accounts for the greater part of the total costs of care packages. By concentrating on these key services, 94 per cent of the total costs of care were predicted for discharged long-stay patients and no less than 91 per cent for people supported by community psychiatric nurses. The results suggest that this reduced list method can work well where the aim is to obtain broad orders of magnitude for the costs of care. By reducing the resources required for research, the approach could expand the current costs information base and move towards meeting the pressing practice demands for costs data.