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.
Dr. med. Andreas Heer-Sonderhoff
|1.||Meyer E, Schröder C, Gastmeier P, Geffers C: The reduction of nosocomial MRSA infection in Germany—an analysis of data from the Hospital Infection Surveillance System (KISS) between 2007 and 2012. Dtsch Arztebl Int 2014; 111: 331–6.|