Limits...
Epidemic Wave Dynamics Attributable to Urban Community Structure: A Theoretical Characterization of Disease Transmission in a Large Network.

Hoen AG, Hladish TJ, Eggo RM, Lenczner M, Brownstein JS, Meyers LA - J. Med. Internet Res. (2015)

Bottom Line: However, this explanation has been largely overlooked in favor of temporal shifts in environmental conditions and human behavior and because of the difficulties associated with estimating large-scale contact patterns.Our results suggest that ordinary contact patterns can produce multiwave epidemics at the scale of a single urban area without the temporal shifts that are usually assumed to be responsible.Understanding the role of community structure in epidemic dynamics allows officials to anticipate epidemic resurgence without having to forecast future changes in hosts, pathogens, or the environment.

View Article: PubMed Central - HTML - PubMed

Affiliation: Computational Epidemiology Group, Children's Hospital Informatics Program, Boston Children's Hospital, Boston, MA, United States.

ABSTRACT

Background: Multiple waves of transmission during infectious disease epidemics represent a major public health challenge, but the ecological and behavioral drivers of epidemic resurgence are poorly understood. In theory, community structure—aggregation into highly intraconnected and loosely interconnected social groups—within human populations may lead to punctuated outbreaks as diseases progress from one community to the next. However, this explanation has been largely overlooked in favor of temporal shifts in environmental conditions and human behavior and because of the difficulties associated with estimating large-scale contact patterns.

Objective: The aim was to characterize naturally arising patterns of human contact that are capable of producing simulated epidemics with multiple wave structures.

Methods: We used an extensive dataset of proximal physical contacts between users of a public Wi-Fi Internet system to evaluate the epidemiological implications of an empirical urban contact network. We characterized the modularity (community structure) of the network and then estimated epidemic dynamics under a percolation-based model of infectious disease spread on the network. We classified simulated epidemics as multiwave using a novel metric and we identified network structures that were critical to the network's ability to produce multiwave epidemics.

Results: We identified robust community structure in a large, empirical urban contact network from which multiwave epidemics may emerge naturally. This pattern was fueled by a special kind of insularity in which locally popular individuals were not the ones forging contacts with more distant social groups.

Conclusions: Our results suggest that ordinary contact patterns can produce multiwave epidemics at the scale of a single urban area without the temporal shifts that are usually assumed to be responsible. Understanding the role of community structure in epidemic dynamics allows officials to anticipate epidemic resurgence without having to forecast future changes in hosts, pathogens, or the environment.

No MeSH data available.


Related in: MedlinePlus

A typical epidemic curve with 2 waves (R0=3.7). Community III drove the first wave; communities I and II drove the second wave.
© Copyright Policy - open-access
Related In: Results  -  Collection

License 1 - License 2
getmorefigures.php?uid=PMC4526984&req=5

figure5: A typical epidemic curve with 2 waves (R0=3.7). Community III drove the first wave; communities I and II drove the second wave.

Mentions: We simulated epidemics through the Montreal network with a stochastic susceptible-infected-recovered model [37] across a range of R0 values. At low R0, only 2 of 3 communities (I and III) sustained transmission, whereas at high R0, epidemic spread was relatively synchronized between communities. Under both of these scenarios, multiwave epidemics were possible but relatively infrequent (Figure 3). At an intermediate value, in the range estimated for the 1918 influenza pandemic in the United States (R0=2.4) [34,35], 44.60% (446/1000) of all epidemics exhibited multiple waves and 87.9% (392/446) of these had an initial epidemic wave in community III with a subsequent wave dominated by the 2 larger communities (Figures 4 and 5). When the first wave was dominated by community III, its peak occurred a mean of 27 time steps (SD 7; n=4074) before that of the second wave. When the second wave was dominated by community III, its peak lagged behind the peak of the first wave by a mean of 22 time steps (SD 8; n=572). The relative size of the second wave increased with R0 because community III became epidemiologically saturated more quickly than the other 2 communities (Figures 3 and 4). Sensitivity analysis suggested that these results are robust to the length of the infectious period (see Figure S4 in Multimedia Appendix 2).


Epidemic Wave Dynamics Attributable to Urban Community Structure: A Theoretical Characterization of Disease Transmission in a Large Network.

Hoen AG, Hladish TJ, Eggo RM, Lenczner M, Brownstein JS, Meyers LA - J. Med. Internet Res. (2015)

A typical epidemic curve with 2 waves (R0=3.7). Community III drove the first wave; communities I and II drove the second wave.
© Copyright Policy - open-access
Related In: Results  -  Collection

License 1 - License 2
Show All Figures
getmorefigures.php?uid=PMC4526984&req=5

figure5: A typical epidemic curve with 2 waves (R0=3.7). Community III drove the first wave; communities I and II drove the second wave.
Mentions: We simulated epidemics through the Montreal network with a stochastic susceptible-infected-recovered model [37] across a range of R0 values. At low R0, only 2 of 3 communities (I and III) sustained transmission, whereas at high R0, epidemic spread was relatively synchronized between communities. Under both of these scenarios, multiwave epidemics were possible but relatively infrequent (Figure 3). At an intermediate value, in the range estimated for the 1918 influenza pandemic in the United States (R0=2.4) [34,35], 44.60% (446/1000) of all epidemics exhibited multiple waves and 87.9% (392/446) of these had an initial epidemic wave in community III with a subsequent wave dominated by the 2 larger communities (Figures 4 and 5). When the first wave was dominated by community III, its peak occurred a mean of 27 time steps (SD 7; n=4074) before that of the second wave. When the second wave was dominated by community III, its peak lagged behind the peak of the first wave by a mean of 22 time steps (SD 8; n=572). The relative size of the second wave increased with R0 because community III became epidemiologically saturated more quickly than the other 2 communities (Figures 3 and 4). Sensitivity analysis suggested that these results are robust to the length of the infectious period (see Figure S4 in Multimedia Appendix 2).

Bottom Line: However, this explanation has been largely overlooked in favor of temporal shifts in environmental conditions and human behavior and because of the difficulties associated with estimating large-scale contact patterns.Our results suggest that ordinary contact patterns can produce multiwave epidemics at the scale of a single urban area without the temporal shifts that are usually assumed to be responsible.Understanding the role of community structure in epidemic dynamics allows officials to anticipate epidemic resurgence without having to forecast future changes in hosts, pathogens, or the environment.

View Article: PubMed Central - HTML - PubMed

Affiliation: Computational Epidemiology Group, Children's Hospital Informatics Program, Boston Children's Hospital, Boston, MA, United States.

ABSTRACT

Background: Multiple waves of transmission during infectious disease epidemics represent a major public health challenge, but the ecological and behavioral drivers of epidemic resurgence are poorly understood. In theory, community structure—aggregation into highly intraconnected and loosely interconnected social groups—within human populations may lead to punctuated outbreaks as diseases progress from one community to the next. However, this explanation has been largely overlooked in favor of temporal shifts in environmental conditions and human behavior and because of the difficulties associated with estimating large-scale contact patterns.

Objective: The aim was to characterize naturally arising patterns of human contact that are capable of producing simulated epidemics with multiple wave structures.

Methods: We used an extensive dataset of proximal physical contacts between users of a public Wi-Fi Internet system to evaluate the epidemiological implications of an empirical urban contact network. We characterized the modularity (community structure) of the network and then estimated epidemic dynamics under a percolation-based model of infectious disease spread on the network. We classified simulated epidemics as multiwave using a novel metric and we identified network structures that were critical to the network's ability to produce multiwave epidemics.

Results: We identified robust community structure in a large, empirical urban contact network from which multiwave epidemics may emerge naturally. This pattern was fueled by a special kind of insularity in which locally popular individuals were not the ones forging contacts with more distant social groups.

Conclusions: Our results suggest that ordinary contact patterns can produce multiwave epidemics at the scale of a single urban area without the temporal shifts that are usually assumed to be responsible. Understanding the role of community structure in epidemic dynamics allows officials to anticipate epidemic resurgence without having to forecast future changes in hosts, pathogens, or the environment.

No MeSH data available.


Related in: MedlinePlus