[Quality control of multicenter evaluation studies of the in vitro activity of antibiotics]

The estimation of the in vitro activity of an antibiotic, by different laboratories, requires the measurement, in standardized conditions, of the inhibition zone diameters versus a series of strains distributed to these laboratories (kits). Using the studies realized with two antibiotics, denoted by A and B, we observed an important variability between the different laboratories for a given strain and a less but still significant variability within the laboratories. Variance analysis revealed a significant interaction between strains and experimental laboratories. These results prevent the establishment of a standard diameter for a given strain but permit the quality control of the measurements and the improvement of the experimental design for future studies.