Towards a common susceptibility testing

We agree with Gould’s comments about the shortcomings of the Stokes’ method and the need for the ‘comparability of data between different centres across the country so that resistance surveillance data can be reliably compared’. In fact we would go further and suggest that, since bacteria do not respect national borders, there should be worldwide comparability of data. Where we part company with Gould is in how we approach that goal. In his paper, Gould outlines the relative merits of agar disc diffusion and breakpoint MIC methods and, in doing so, indicates many of the factors that affect the reproducibility of the former method. These include inoculum, temperature, atmosphere and generation time of the bacterium. Surely, however, these affect both methods? Furthermore, we would not agree with the view that only the broth dilution method lends itself well to automation: automatic plate readers, with their data collation system, have a lot in their favour. Other than on these points we would not disagree with Gould’s precis of the pros and cons of the different approaches. The thrust of Gould’s paper is to argue for the worldwide use of the NCCLS method for susceptibility testing and here we disagree. We believe that the US-derived method, although having a number of creditable features, is too rigid in its concept to be recommended for use in the UK (or for many other EU countries). Our reasons can be listed: