In antimicrobic susceptibility testing, minimum inhibitory concentration (MIC) susceptibility break points are defined by correlation of bacteriologic-clinical outcome data with MIC data for the infecting organisms. Disk diffusion [that is, zone-diameter (Z)] correlates are then established that provide for the prediction of organism susceptibility, while misclassification errors are kept to a minimum. The determination of Z break points through an error-rate-bounded classification scheme was first proposed by Metzler and DeHaan (1972). This method involves one MIC break point that separates susceptible and resistant strains. More recently, researchers have preferred to use two MIC break points (susceptible and resistant) that separate susceptible, moderately susceptible, and resistant strains. There is no known methodology for determining the Z break points for this latter situation, other than enumerating solutions for all feasible Z break-point pairs and choosing among the results. Our interest lay in presenting a methodology for determining the Z break points once the MIC break points are established. By deriving an index as a function of Z break points, a search method for finding the optimal Z break points is given. For the data set examined, our index interval solution required only a small percentage of solutions to be examined.
