OBJECTIVETo describe the effect of universal methicillin-resistant Staphylococcus aureus (MRSA) decolonization therapy in a large intensive care unit (ICU) on the rates of MRSA cases and acquisitions in a UK hospital.DESIGNDescriptive study.SETTINGUniversity Hospitals Birmingham (UHB) NHS Foundation Trust is a tertiary referral teaching hospital in Birmingham, United Kingdom, that provides clinical services to nearly 1 million patients every year.METHODSA break-point time series analysis and kernel regression models were used to detect significant changes in the cumulative monthly numbers of MRSA bacteremia cases and acquisitions from April 2013 to August 2016 across the UHB system.RESULTSPrior to 2014, all ICU patients at UHB received universal MRSA decolonization therapy. In August 2014, UHB discontinued the use of universal decolonization due to published reports in the United Kingdom detailing the limited usefulness and cost-effectiveness of such an intervention. Break-point time series analysis of MRSA acquisition and bacteremia data indicated that break points were associated with the discontinuation and subsequent reintroduction of universal decolonization. Kernel regression models indicated a significant increase (P<.001) in MRSA acquisitions and bacteremia cases across UHB during the period without universal decolonization.CONCLUSIONWe suggest that routine decolonization for MRSA in a large ICU setting is an effective strategy to reduce the spread and incidence of MRSA across the whole hospital.Infect Control Hosp Epidemiol 2017;38:430–435
Strong uniform convergence rates in robust nonparametric time series analysis and prediction: Kernel regression estimation from dependent observations
