IEEE Computational Intelligence Magazine - November 2020 - 33

C. Resurgence Risks in the United States

We also used the proposed framework to estimate the epidemic
evolution in different states of the United States. We observed
that, as of the week ending 31st May, the averaged reproduction
number R t in 30 states exceeds 1 (Figure 10). These could be
related to the recent lift of government restrictions and alert us
to take a close monitoring on the epidemic evolution.
At the time of preparing this paper (18th June 2020), 29 out
of the 30 states we alerted on 9th June 2020 have experienced
an increased number of daily confirmed cases compared to that
of 31st May, and 14 states have recorded all-time high after 31st
May. When we prepared the final version in early August, this
alarming prediction of a second wave outbreak is unfortunately
proven true for all the states listed.
So far, the application of the framework to many countries
and the retrospective impact analysis of intervention measures in
European countries indicate the effectiveness of our approach
in monitoring R t . This can be further validated by predicting
the evolution of  and projected infections in the
future study. Our current study has several limitations. Firstly,
the reporting protocols and standards of confirmed cases, as
well as the detection rates, vary among countries. However, as
long as the reporting bias is consistent over time, the inference
results of p t, D t and R t should also be consistent under the
protocol. Since the impacts of interventions are assessed by
measuring the evolution of these parameters, the framework
can be generally applicable to assess the policy impacts among
different reporting protocols. We also note that the implementation of multiple intervention measures within a short interval makes it challenging to quantify the impact of a single
measure which needs further statistical analysis.
VII. Conclusions

In conclusion, we propose a comprehensive data assimilation
approach of using Bayesian updating to timely estimate parameters of COVID-19 epidemic models. The disease transmission
dynamics is modelled by renewal equations with time-varying
parameters. Instead of purely focusing on estimating instantaneous reproduction number R t , we introduce two complementary parameters, the mitigation factor ( p t ) and the suppression
factor (D t ), to quantify intervention impacts at a finer granularity. A Bayesian updating scheme is adopted to dynamically
infer model parameters. By monitoring and analyzing the evolution of the estimated parameters, the impacts of intervention
measures in different response levels can be quantitatively
assessed. We have applied our method to European countries,
the United States and Wuhan, and reveal the effects of interventions in these countries and the resurgence risk in the
United States. Our work opens a promising venue to inform
policy for better decision-making in response to a possible second-wave outbreak.
Acknowledgment

We express our sincere thanks to all members of the joint analysis team between Imperial College London, University of

Cambridge, University of Kent, and Hong Kong Baptist University. We thank Yuting Xing for helping collect epidemic data
in Wuhan and the United States. We thank Siyao Wang and
Liqun Wu for their efforts on developing a digital tracing app
for validation and visualization.
References

[1] E. Dong, H. Du, and L. Gardner, "An interactive web-based dashboard to track COVID-19 in real time," Lancet Infect. Dis., vol. 20, no. 5, pp. 533-534, May 2020. doi:
10.1016/S1473-3099(20)30120-1.
[2] T. Hale, A. Petherick, T. Phillips, and S. Webster, "Variation in government responses
to COVID-19," 2020.
[3] A. Pan et al., "Association of public health interventions with the epidemiology of the
COVID-19 outbreak in Wuhan, China," JAMA, vol. 323, no. 19, p. 1915, May 2020. doi:
10.1001/jama.2020.6130.
[4] R. Li et al., "Substantial undocumented infection facilitates the rapid dissemination of
novel coronavirus (SARS-CoV2)," Science, vol. 3221, p. eabb3221, Mar. 2020.
[5] L. Ferretti et al., "Quantifying SARS-CoV-2 transmission suggests epidemic control
with digital contact tracing.," Science, vol. 6936, pp. 1-13, Mar. 2020.
[6] E. Vynnycky and R. White, An Introduction to Infectious Disease Modelling. New York:
Oxford Univ. Press, 2010.
[7] N. C. Grassly and C. Fraser, "Mathematical models of infectious disease transmission,"
Nat. Rev. Microbiol., vol. 6, no. 6, pp. 477-487, 2008. doi: 10.1038/nrmicro1845.
[8] C. Fraser, "Estimating individual and household reproduction numbers in an emerging epidemic," PLoS One, vol. 2, no. 8, 2007. doi: 10.1371/journal.pone.0000758.
[9] J. Ma and D. J. D. Earn, "Generality of the final size formula for an epidemic of a
newly invading infectious disease," Bull. Math. Biol., vol. 68, no. 3, pp. 679-702, 2006.
doi: 10.1007/s11538-005-9047-7.
[10] K. Leung, J. T. Wu, D. Liu, and G. M. Leung, "First-wave COVID-19 transmissibility and severity in China outside Hubei after control measures, and second-wave scenario
planning: a modelling impact assessment," Lancet, vol. 395, no. 10233, pp. 1382-1393,
Apr. 2020. doi: 10.1016/S0140-6736(20)30746-7.
[11] R. N. Thompson et al., "Improved inference of time-varying reproduction numbers during infectious disease outbreaks," Epidemics, vol. 29, Aug. 2019. doi: 10.1016/
j.epidem.2019.100356.
[12] A. Cori, N. M. Ferguson, C. Fraser, and S. Cauchemez, "A new framework and
software to estimate time-varying reproduction numbers during epidemics," Am. J. Epidemiol., vol. 178, no. 9, pp. 1505-1512, 2013. doi: 10.1093/aje/kwt133.
[13] J. Wallinga and P. Teunis, "Different epidemic curves for severe acute respiratory
syndrome reveal similar impacts of control measures," Am. J. Epidemiol., vol. 160, no. 6,
pp. 509-516, 2004. doi: 10.1093/aje/kwh255.
[14] D. Adam, "A guide to R-the pandemic's misunderstood metric," Nature, vol. 583, no.
7816, pp. 346-348, 2020. doi: 10.1038/d41586-020-02009-w.
[15] N. Imai, I. Dorigatti, A. Cori, C. Donnelly, S. Riley, and N. Ferguson, "Report 2:
Estimating the potential total number of novel Coronavirus cases in Wuhan City, China,"
2020.
[16] Q. Li et al., "Early transmission dynamics in Wuhan, China, of novel coronavirusinfected pneumonia," N. Engl. J. Med., vol. 382, no. 13, pp. 1199-1207, 2020.
[17] J. T. Wu, K. Leung, and G. M. Leung, "Nowcasting and forecasting the potential
domestic and international spread of the 2019-nCoV outbreak originating in Wuhan,
China: a modelling study," Lancet, vol. 395, no. 10225, pp. 689-697, 2020. doi: 10.1016/
S0140-6736(20)30260-9.
[18] S. Flaxman et al., "Estimating the effects of non-pharmaceutical interventions on
COVID-19 in Europe," Nature, pp. 1-5, 2020.
[19] P. Nouvellet et al., "Report 26: Reduction in mobility and COVID-19 transmission."
[20] M. Asch, M. Bocquet, and M. Nodet, Data Assimilation: Methods, Algorithms, and
Applications. 2016.
[21] Z. Chen, "Bayesian filtering: From Kalman filters to particle filters, and beyond,"
Statistics, vol. 182, no. 1, pp. 1-69, 2003.
[22] C. J. Rhodes and T. D. Hollingsworth, "Variational data assimilation with epidemic models," J. Theor. Biol., vol. 258, no. 4, pp. 591-602, 2009. doi: 10.1016/j.
jtbi.2009.02.017.
[23] L. M. A. Bettencourt and R. M. Ribeiro, "Real time bayesian estimation of the epidemic potential of emerging infectious diseases," PLoS One, vol. 3, no. 5, p. e2185, 2008.
doi: 10.1371/journal.pone.0002185.
[24] L. Cobb, A. Krishnamurthy, J. Mandel, and J. D. Beezley, "Bayesian tracking of
emerging epidemics using ensemble optimal statistical interpolation," Spat. Spatiotemporal.
Epidemiol., vol. 10, pp. 39-48, 2014. doi: 10.1016/j.sste.2014.06.004.
[25] E. Goldstein, J. Dushoff, M. Junling, J. B. Plotkin, D. J. D. Earn, and M. Lipsitch,
"Reconstructing inf luenza incidence by deconvolution of daily mortality time series,"
Proc. Nat. Acad. Sci. U. S. A., vol. 106, no. 51, pp. 21,825-21,829, 2009. doi: 10.1073/
pnas.0902958106.
[26] C. Mark, C. Metzner, L. Lautscham, P. L. Strissel, R. Strick, and B. Fabry,
"Bayesian model selection for complex dynamic systems," Nat. Commun., vol. 9,
no. 1, 2018.
[27] A. Agresti, An Introduction to Categorical Data Analysis. Hoboken, NJ: Wiley, 2018.



NOVEMBER 2020 | IEEE COMPUTATIONAL INTELLIGENCE MAGAZINE

33
IEEE Computational Intelligence Magazine - November 2020

Table of Contents for the Digital Edition of IEEE Computational Intelligence Magazine - November 2020
