Fig. 4

Effect of Go To Travel campaign prefectures above threshold values for incidence across different models. We counted the prefectures where the number of newly reported cases before and after the campaign (intervention) exceeded five cases per 100,000 population per week. The vertical lines indicate when the second campaign (intervention) started. The yellow dots represent observed data; the blue lines show the estimated results of the interrupted time-series analysis. The blue shaded areas indicate 95% confidence intervals. The blue dashed lines during the second campaign period represent the estimated counterfactual scenario, i.e., if the campaign had not been launched. Model 1 assumed changes in the baseline level and rate of increase following the second campaign. Model 2 accounted for temperature in addition to the causal effect of the campaign. Compared with model 1, model 3 assumed that only the change in the rate of increase occurred following the campaign