Background: Clostridium difficile infection remains a major challenge for hospitals. Although targeted infection control initiatives have been shown to be effective in reducing the incidence of hospital-acquired C.difficile infection, there is little evidence available to assess the effectiveness of specific interventions. Aim: To use statistical modelling to detect substantial reductions in the incidence of C.difficile from time series data from two hospitals in England, and relate these time points to infection control interventions. Methods: A statistical breakpoints model was fitted to likely hospital-acquired C.difficile infection incidence data from a teaching hospital (2002-2009) and a district general hospital (2005-2009) in England. Models with increasing complexity (i.e. increasing the number of breakpoints) were tested for an improved fit to the data. Partitions estimated from breakpoint models were tested for individual stability using statistical process control charts. Major infection control interventions from both hospitals during this time were grouped according to their primary target (antibiotics, cleaning, isolation, other) and mapped to the model-suggested breakpoints. Findings: For both hospitals, breakpoints coincided with enhancements to cleaning protocols. Statistical models enabled formal assessment of the impact of different interventions, and showed that enhancements to deep cleaning programmes are the interventions that have most likely led to substantial reductions in hospital-acquired C.difficile infections at the two hospitals studied.
- Clostridium difficile
- Infection control Interventions
- Nosocomial infection
- Statistical model