A major reason for the dramatic increase in laboratory costs is the increased number of laboratory investigations, many of which are perceived to be a waste of time and money. In consequence, several strategies, reviewed below, for curtailing clinicians' laboratory usage have been proposed but none has been universally successful. It is clear, however, that a combination of several strategies is likely to be more successful than any single approach. There may be several reasons why no single strategy has been widely adopted in routine practice. We suggest that modern computers have the ability to overcome these obstacles. We propose, therefore, that expert systems at the clinical-laboratory interface utilising a multifactorial approach within a framework of commitment from clinicians and laboratorians offer the best prospect for reducing inappropriate laboratory usage.