Australian evaluation of Autobac I with suggested interpretive and technical modifications.

Autobac I, a recently introduced semiautomated method for rapid antibiotic susceptibility testing, has been evaluated by comparison with the calibrated dichotomous sensitivity disk diffusion technique, which is routinely used in many Australian hospitals. Only the most common clinical isolates, Staphylococcus aureus, Escherichia coli, Klebsiella sp., and Proteus mirabilis, were included in this evaluation, and an overall interpretive agreement of 93% was obtained. However, an unusually high rate of discrepancy was noted in several organism-antibiotic combinations, in particular E. coli and P. mirabilis with ampicillin, S. aureus with penicillin, and methicillin-resistant S. aureus with methicillin, erythromycin, and clindamycin. The discrepancies associated with ampicillin have been reduced from 29 and 24% for E. coli and P. mirabilis, respectively, to less than 5% after the utilization of commercial 10-micrograms diffusion disks, in preference to the lower antibiotic content disks supplied by the Autobac manufacturer. Furthermore, modifications in the interpretive procedure have eliminated discrepancies associated with S. aureus and penicillin.