The proper disk mass for diffusion susceptibility tests with cefoperazone-sulbactam was determined by using a predictor panel of clinical isolates that included staphylococci and gram-negative bacteria intrinsically susceptible, intrinsically resistant, and of various susceptibilities because of the production of different types and amounts of beta-lactamase. A primary panel of 24 isolates was used to screen various disk masses of cefoperazone and sulbactam in disk diffusion susceptibility tests. Regression analyses were performed for each combination by comparing MICs to zone diameters. Analysis of each component demonstrated that decreasing the disk mass of cefoperazone shifted the regression line to the left while decreasing the disk mass of sulbactam diminished the slope of the line. Ten candidate disks that adequately separated susceptible and resistant strains among the primary panel were identified, and these 10 disks, along with the previously proposed 75/30-micrograms disk, were then tested against an expanded panel of 265 isolates. Results indicated that a 30/20-micrograms cefoperazone-sulbactam disk provided the best separation between susceptible and resistant strains when interpretive criteria of less than or equal to 15 mm for resistance, 16 to 19 mm for moderate susceptibility, and greater than or equal to 20 mm for susceptibility were used. They also identified discrepancies between agar and broth microdilution MICs of sufficient size to warrant separate interpretive criteria for the two methods. Overall, the use of a predictor panel to develop interpretive criteria for susceptibility tests appeared to be a very useful approach, especially when antibiotics designed to be used against drug-resistant organisms are involved.