Spatio-temporal stochastic modelling of Clostridium difficile

Clostridium difficile-associated diarrhoea (CDAD) occurs sporadically or in small discrete outbreaks. Stochastic models may help to inform hospital infection control strategies. Bayesian framework using data augmentation and Markov chain Monte Carlo methods were applied to a spatio-temporal model of CDAD. Model simulations were validated against 17 months of observed data from two 30-bedded medical wards for the elderly. Simulating the halving of transmission rates of C. difficile from other patients and the environment reduced CDAD cases by 15%. Doubling the rate at which patients become susceptible increased predicted CDAD incidence by 63%. By contrast, doubling environmental load made hardly any difference, increasing CDAD incidence by only 3%. Simulation of different interventions indicates that for the same effect size, reducing patient susceptibility to infection is more effective in reducing the number of CDAD cases than lowering transmission rates.