Fig. 3

Hidden Markov model. C = the estimated number of detected and undetected VRE colonised/infected patients in the underlying hidden states at each time points (t = 1,2,3,….); D = the number of patients detected as VRE colonised/infected. The transition model linking the hidden states are represented by the horizontal lines, whereas the observation model linking the hidden states and the corresponding observed data are represented by the vertical lines. The probability of observing d new infections on day t, d(t) is assumed to be a Poisson distribution given by d(t)~ Poisson (λ). Each day, the hidden states were updated based on observed new acquisitions, admissions and discharges from the dataset, and also via the transition probabilities in Eqs. (1) to (3)