Fig. 4

Impact of stop-and-go lockdown scenarios on hospitalizations, effective days under restrictions, and number of lockdowns. a-c Heatmaps showing the relative variation in cumulative hospital admissions (a), the relative variation in effective days under restrictions (b), and the number of lockdowns (c), as functions of the thresholds for trigger (y-axis) and release (x-axis) of nationwide lockdowns. Relative changes are computed with respect to observations. The red squares indicate specific values of trigger and release threshold that are discussed in the main text and presented in detail in Fig. 5. Numerical values are reported only in the area where both hospitalizations and effective days under restrictions are reduced by the lockdowns compared to observations. d, e Probability distributions of the relative variations in hospitalizations (d) and in effective days spent under restrictions (e) compared to observations, in the region of the trigger-release parameter space where both quantities are reduced by the lockdowns. The vertical dashed lines represent the median values of the distribution. f Histogram of the percentage of scenarios with a given number of lockdowns, among the scenarios that reduce both hospitalizations and effective days