Designing a study for evaluating a protocol for the selective performance of preoperative tests

Some basic concepts concerning classification of studies evaluating diagnostic procedures are given. The specific problem of assessment of routine preoperative tests is discussed: actual knowledge and methodological criticisms of previous studies, hypothesis for further evaluations, difficulties in the choice of a design and realization of a clinical study to assess a protocol for selective ordering of preoperative tests during one year in a teaching hospital. 3866 consecutive patients were included. The protocol led to perform only 33 per cent of all potential tests and did not apparently produce major adverse effects on the quality of care as appreciated by a clinical decision analysis.