AIMS: To rationalise confusion in the literature concerning the analysis of combined antimicrobials, specifically to see if the combination index (CI) method of analysis was as rigorous as claimed. METHODS & RESULTS: data from previous studies of the inhibition of Staphylococcus aureus by mixed antimicrobials were re-analysed using the CI method and a model which takes account of differences in the concentration exponents of individual antimicrobials. CONCLUSIONS: The Chou-Talalay combination index method for the analysis of combined antimicrobials was found to be valid only in the specific cases where concentration exponents were equal. In these cases the CI method was found to be a function of the residuals of fitting the additive model to the observed data. Where concentration exponents were not equal the CI method was invalid, whereas the additive model took these differences into account. SIGNIFICANCE AND IMPACT OF STUDY: The CI method can be replaced wholly by the additive model described. The model allows simple regression to be used to analyse whole data sets and provides simple graphical output. This article is protected by copyright. All rights reserved.