Individual-Based Modelling on the Spread of Infectious Diseases in Singapore

Participants: Cheng Tee Hiang, Xiuju Fu, Kee Khoon Gary Lee, Harold Soh, Stefan Ma, Limsoon Wong, Gaoxi Xiao, Shi Xiao, Zhang Tianyou, Zhou Jie.

Background

The project will not only handle a special case of infectious diseases like avian influenza, but also contribute to enhance fundamental knowledge of epidemiology sciences. Essential information obtained from the simulation model will help health agency develop effective containment strategies in advance of a pandemic infectious disease. The basic simulation model embedding contexts of social connectivity of Singapore can be extended to measure scale and make predictions on the spread of other infectious diseases.

Project Activity Records / Project Progress Reports / Questionaires

Achievements

• Modeled the local society into six components including household, school, shopping mall, hospital, transportation, and workplace, where most social contacts occur. Basic social structure data has been acquired from Singapore Statistic Department, Ministry of Education, Ministry of Manpower, Land Transport Authority, a few hospitals, as well as nearly 2,000 online and offline surveys.

• Developed an efficient contact network generator, HPCgen. It is developed with a generic structure that allows convenient incorporation of new social contact components as needed. It uses an efficient adjacency matrix structure to represent social contacts, so that millions of pair-wise operations can be flexibly encapsulated into a single matrix operation during contact (linkage) generation. On an 8-CPU parallel computer, a 4.8-million node network model can be built in 100 seconds. The fast generation of ultra-large networks based on input real-life data makes extensive and repetitive simulations of various scenarios technically feasible.

• Worked on two simulation and analysis platforms. The primary simulation platform is based on recently developed percolation theory. This platform can simulate disease spread dynamically and allow geographic information to be integrated and visualized. It can also accommodate multiple types of infectious diseases. The second platform is based on classical network-based simulation program. While the second platform is more restrictive, it is more traditional and allows us to validate results of the percolation theory platform on simpler cases before we apply it to difficult cases.

• Based on the generated contact network models, extensive individual-based simulations were carried out for evaluating disease spread in different scenarios. E.g.,
  • We found that a typical flu with a reproductive number of 1.5-2.0 may spread throughout the Singapore in about five to nine months if nothing is done to control it. Simulation results also revealed some interesting results on the effectiveness of different control strategies and their combinations: School closure may be very useful if (and probably only if) parents keep their children at home. Workforce shift, unless combined with proper control in public transportation systems, may be actually less effective. Partial vaccination, unless the vaccinated population is of a large enough portion (at least 50%), may not be able to prevent the disease from having an outbreak. However, even a low percentage vaccination may buy valuable time. Enforced quarantine policy, though with a relatively high social cost, may not prevent an outbreak from happening if the disease has a nontrivial infectious period before symptoms generate; the situation can easily get worse where there is a steady flow of imported cases. It is also found that the timing of implementing various control strategies is of crucial importance.

  • We tested the effects of carrying out extensive control at "hub" nodes such as MRT stations and clinics, or vaccinating the most vulnerable individuals, etc. Our studies also showed that existing analytical methods tend to over-estimate the infected population size when people are heavily clustered. The human society in Singapore however is indeed heavily clustered, where close colleagues/friends of the same person are generally close colleagues/friends to each other.

  • We made studies on the effects of fear factors. (i) For the simple model where people homogeneously reduce their social connections in panic, it is found that such reductions do not significantly lower the chance of having an outbreak of disease, yet even a moderate reduction in social connections help to significantly reduce the infected population size and buy valuable time before a full-size outbreak occurs. The conclusion holds even if most people take action only after at least one of their close contacts is infected. (ii) The above mentioned effects in reducing the infected population size however becomes much less significant when taking into account that people may choose, at any cost, to keep their connections with their family members and a few closest friends. (iii) If people only weaken their social connections rather than temporarily or permanently cut some of them, very complex dynamics can be observed in certain scenarios. (iv) It is found for the first time that, without any other disturbances to the social systems, reemergence of infectious diseases can be caused singly by the effect that people reduce social connections in pandemic situation and start to reconstruct lost connections when the disease appears to be dying out. (v) It is observed in our surveys that people who tend to take more seriously the information from the Internet, on average, spend less time outdoor or staying with their friends. The conclusion holds even when we restrict the sampling to only young people between 15-24 years old.

Selected Publications

• X. Fu, W. J. I. Teo, H. Soh, G. Xiao, G. Lee. Simulating Avian Influenza Spread over Singapore Clinic Networks. Proceedings of X International Symposium on Respiratory Viral Infections, pages ???--???, Singapore, 28 February - 3 March 2008.

• X. Fu, S. Lim, L. Wang, G. Lee, S. Ma, L. Wong, G. Xiao. Key Node Selection for Containing Infectious Disease Spread Using Particle Swarm Optimization. Proceedings of IEEE Swarm Intelligence Symposium 2009, pages 109--113, Nashville, USA, 29 March - 2 April 2009.

• Y. Wang, G. Xiao, T. H. Cheng, S. Xiao, X. Fu. Robustness of Complex Communication Networks under Link Attacks. Proceedings of ACM International Conference on Advanced Infocomm Technology, pages ???---???, Shenzhen, China, 30 July 2008.

• Yubo Wang, Jie Hu, Gaoxi Xiao, Limsoon Wong, Stefan Ma, Tee Hiang Cheng. A Preliminary Study on the Effects of Fear Factors in Disease Propagation. Proceedings of 1st International Conference on Complex Sciences: Theory and Applications (Complex'2009), pages 1387-1397, Shanghai, China, 23-25 February 2009.

• Yubo Wang, Gaoxi Xiao, Jie Hu, Tee Hiang Cheng, Limsoon Wong. Imperfect Targeted Immunization in Scale-Free Networks. Physica A: Statistical Mechanics and its Applications, 388(12):2535--2546, June 2009.

• Tianyou Zhang, Soon Hong Soh, Xiuju Fu, Kee Khoon Lee, Limsoon Wong, Stefan Ma, Gaoxi Xiao, Chee Keong Kwoh. HPCgen---A Fast Generator of Contact Networks of Large Urban Cities for Epidemiological Studies. Proceedings of International Conference on Computational Intelligence, Computer Modeling and Simulation, pages 198--203, Brno, Czech Republic, September 2009.

• Jie Zhou, Gaoxi Xiao, Limsoon Wong, Xiuju Fu, Stefan Ma, Tee Hiang Cheng. Generation of arbitrary two-point-correlated directed networks with given modularity. Physics Letters A, 374(31-32):3129--3135, July 2010.

• Harold Soh, Sonja Lim, Tianyou Zhang, Xiuju Fu, Gary Kee Khoon Lee, Terence Gih Guang Hung, Limsoon Wong, Pan Di, Silvester Prakasam. Weighted Complex Network Analysis of Travel Routes on the Singapore Public Transportation System. Physica A: Statistical Mechanics and its Applications, 389(24):5852--5863, December 2010.

• Tianyou Zhang, Xiuju Fu, Chee Keong Kwoh, Gaoxi Xiao, Limsoon Wong, Stefan Ma, Harold Soh, Gary Kee Khoon Lee, Terence Hung, Michael Lees. Temporal Factors in School Closure Policy for Mitigating the Spread of Influenza. Journal of Public Health Policy, 32(2):180--197, May 2011.

• Yubo Wang, Goaxi Xiao, Limsoon Wong, Xiuju Fu, Stefan Ma, Tee Hiang Cheng. Effects of Fear Factors in Disease Propagation. Journal of Physics A: Mathematical and Theoretical, 44(35):355101, September 2011.

• Tianyou Zhang, Xiuju Fu, Stefan Ma, Gaoxi Xiao, Limsoon Wong, Chee Keong Kwoh, Michael Lees, Gary Kee Khoon Lee, Terence Hung. Evaluating Temporal Factors in Combined Interventions of Workforce Shift and School Closure for Mitigating the Spread of Influenza. PLoS ONE, 7(3):e32203, March 2012.

• J. Zhou, G. Xiao, S. A. Cheong, X. Fu, L. Wong, S. Ma, T. H. Cheng. Epidemic Reemergence in Adaptive Complex Networks. Physical Review E, 85(3):036107, March 2012.

Selected Presentations

• None yet ...

Acknowledgements