Fig. 2 | Nature Communications

Fig. 2

From: Inferring collective dynamical states from widely unobserved systems

Fig. 2

Disease propagation. In epidemic models, the reproductive number m can serve as an indicator for the infectiousness of a disease within a population, and predict the risk of large incidence bursts. We have estimated \(\hat m_{}^{}\) from incidence time series of measles infections for 124 countries worldwide (Supplementary Note 9); as well as noroviral infection, measles, and invasive meticillin-resistant Staphylococcus aureus (MRSA) infections in Germany. a MR estimation of \(\hat m\) is shown for measles infections in three different countries. Error bars here and in all following figures indicate 1SD or the corresponding 16 to 84% confidence intervals if asymmetric. The reproductive numbers \(\hat m\) decrease with the vaccination rate (Spearman rank correlation: r = −0.342, p < 10−4). b Weekly case report time series for norovirus, measles and MRSA in Germany. c Reproductive numbers \(\hat m\) for these infections. d When artificially subsampling the measles recording (under-ascertainment), conventional estimation underestimates \(\hat m_{\mathrm{C}}\), while MR estimation still returns the correct value. Both estimators return the same \(\hat m\) under full sampling