Surgical decision analysis. Methodological aspects of statistical evaluations in therapeutic trials

This report aims at discussing some of the sources of error and pitfalls involved in evaluating the efficiency of alternative modes of therapy in clinical materials. One of the main obstacles in such analysis is that the therapeutic groups are basically different due to the original selection of patients. In a previous paper (Pellettieri, 1980) a special method was worked out to circumvent this problem. The method requires that the variables of main prognostic importance can be identified. Patients in the different therapeutic groups who have identical patterns of combination of variables (risk profiles) can then be compared and the efficiency of different treatment programs can be evaluated. We believe that this method can be used in other fields of medicine.