Correspondence (letter to the editor): Large Numbers Do not Make for Great Science

The presented data on time trends are essentially summarized in Table 1. Every two-year period represents a different group of institutions, so that a comparison without detailed confirmation of homogeneity of the compared groups of intensive care units and surgical wards is not actually valid. From a statistical perspective, the given baseline population differs in each period. Their similarity could be made plausible on the basis of the KISS data, but the authors omitted to do this. I was perplexed by the small rise in confirmed cases of S. aureus infections which was below 10% from 2007/2008 to 2011/2012, while during the same time period the numbers of participating institutions rose by 38% and 57%. This is obviously due to the effect of relevant variables that were ignored but would have deserved attention; instead, discrete changes of the absolute and relative incidence of nosocomial MRSA infections were focussed on that cannot be interpreted in any meaningful way because of the lack of background information.