Local and marginal control charts applied to methicillin resistant "Staphylococcus aureus bacteraemia" reports in UK acute National Health Service trusts

We consider the general problem of simultaneously monitoring multiple series of counts, applied in this case to methicillin resistant "Staphylococcus aureus" (MRSA) reports in 173 UK National Health Service acute trusts. Both within-trust changes from baseline ('local monitors') and overall divergence from the bulk of trusts ('relative monitors') are considered. After standardizing for type of trust and overall trend, a transformation to approximate normality is adopted and empirical Bayes shrinkage methods are used for estimating an appropriate baseline for each trust. Shewhart, exponentially weighted moving average and cumulative sum charts are then set up for both local and relative monitors: the current state of each is summarized by a "p"-value, which is processed by a signalling procedure that controls the false discovery rate. The performance of these methods is illustrated by using 4.5 years of MRSA data, and the appropriate use of such methods in practice is discussed. Copyright (c) 2009 Royal Statistical Society.