Fig. 4.

Counterfactual analysis demonstrates the effect of lockdown timing on epidemiological outcomes. We examined counterfactual introductions of three lockdowns in Northern Ireland and the Republic of Ireland, assuming that they would have started seven days and 14 days earlier. The percentage difference in cumulative hospital admissions between the counter-factual and factual scenarios is shown. The black line and grey band indicates the median and 95% predictive interval, respectively. The predictive interval signifies 95% of simulated outcomes generated based on samples from the posterior distribution of model parameters and measurement errors.