Limits...
The Effect of Disease-Induced Mortality on Structural Network Properties.

Gallos LK, Fefferman NH - PLoS ONE (2015)

Bottom Line: While many studies have explored the influence of individual epidemiological parameters and/or underlying network topologies on the resulting disease dynamics, we here provide a systematic overview of the interactions between these two influences on population-level disease outcomes.Lastly, we discuss how the expected individual-level disease burden is influenced by the complete suite of epidemiological characteristics for the circulating disease and the ongoing process of network compromise.Our results have broad implications for prediction and mitigation of outbreaks in both natural and human populations.

View Article: PubMed Central - PubMed

Affiliation: Department of Ecology, Evolution, and Natural Resources, Rutgers University - New Brunswick, NJ 08901, United States of America; DIMACS, Rutgers University - Piscataway, NJ 08854, United States of America.

ABSTRACT
As the understanding of the importance of social contact networks in the spread of infectious diseases has increased, so has the interest in understanding the feedback process of the disease altering the social network. While many studies have explored the influence of individual epidemiological parameters and/or underlying network topologies on the resulting disease dynamics, we here provide a systematic overview of the interactions between these two influences on population-level disease outcomes. We show that the sensitivity of the population-level disease outcomes to the combination of epidemiological parameters that describe the disease are critically dependent on the topological structure of the population's contact network. We introduce a new metric for assessing disease-driven structural damage to a network as a population-level outcome. Lastly, we discuss how the expected individual-level disease burden is influenced by the complete suite of epidemiological characteristics for the circulating disease and the ongoing process of network compromise. Our results have broad implications for prediction and mitigation of outbreaks in both natural and human populations.

No MeSH data available.


Related in: MedlinePlus

Duration of epidemic: time till stochastic die-out.From top to bottom: two-dimensional lattices, random scale-free networks (λ = 2.5), and self-organized networks. The x-axis corresponds to the infection probability, β, and the y-axis to the death probability, f, of an infected node. The columns correspond to the protection loss rate, r, of a node (left to right): r = 0.05, r = 0.20, and r = 0.50.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4552173&req=5

pone.0136704.g003: Duration of epidemic: time till stochastic die-out.From top to bottom: two-dimensional lattices, random scale-free networks (λ = 2.5), and self-organized networks. The x-axis corresponds to the infection probability, β, and the y-axis to the death probability, f, of an infected node. The columns correspond to the protection loss rate, r, of a node (left to right): r = 0.05, r = 0.20, and r = 0.50.

Mentions: An important feature of the spreading process is the duration of the epidemics. A longer duration leaves a much larger time window for possible intervention, while a shorter duration may complete the maximum spreading cycle before any action can be taken. In lattices, the duration is dictated by the value of β and is almost independent of f, except for large β and small f values where we observe somewhat longer durations (Fig 3, top row). This is in contrast to scale-free networks (Fig 3, middle row), where three regimes are found: a) low-β regime: the disease lasts only for a couple of steps and dies rapidly without causing any damage, b) intermediate-to-large β regime and large f values: the duration of the epidemics in the N = 200 network is of the order of 10 steps. Even though the duration is small, the damage is considerable, and c) intermediate-to-large β regime and small f values: the epidemics now can last for more than 100 steps, even though it is not as lethal as the previous case. In self-organized networks (Fig 3, bottom row) the picture is basically the same as in scale-free networks, but now the epidemics may spread extremely fast even at large values of β. Interestingly, even though the duration may change by an order of magnitude as we vary the mortality probability, the end result is always a large fraction of the population dying (from 75–100%).


The Effect of Disease-Induced Mortality on Structural Network Properties.

Gallos LK, Fefferman NH - PLoS ONE (2015)

Duration of epidemic: time till stochastic die-out.From top to bottom: two-dimensional lattices, random scale-free networks (λ = 2.5), and self-organized networks. The x-axis corresponds to the infection probability, β, and the y-axis to the death probability, f, of an infected node. The columns correspond to the protection loss rate, r, of a node (left to right): r = 0.05, r = 0.20, and r = 0.50.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4552173&req=5

pone.0136704.g003: Duration of epidemic: time till stochastic die-out.From top to bottom: two-dimensional lattices, random scale-free networks (λ = 2.5), and self-organized networks. The x-axis corresponds to the infection probability, β, and the y-axis to the death probability, f, of an infected node. The columns correspond to the protection loss rate, r, of a node (left to right): r = 0.05, r = 0.20, and r = 0.50.
Mentions: An important feature of the spreading process is the duration of the epidemics. A longer duration leaves a much larger time window for possible intervention, while a shorter duration may complete the maximum spreading cycle before any action can be taken. In lattices, the duration is dictated by the value of β and is almost independent of f, except for large β and small f values where we observe somewhat longer durations (Fig 3, top row). This is in contrast to scale-free networks (Fig 3, middle row), where three regimes are found: a) low-β regime: the disease lasts only for a couple of steps and dies rapidly without causing any damage, b) intermediate-to-large β regime and large f values: the duration of the epidemics in the N = 200 network is of the order of 10 steps. Even though the duration is small, the damage is considerable, and c) intermediate-to-large β regime and small f values: the epidemics now can last for more than 100 steps, even though it is not as lethal as the previous case. In self-organized networks (Fig 3, bottom row) the picture is basically the same as in scale-free networks, but now the epidemics may spread extremely fast even at large values of β. Interestingly, even though the duration may change by an order of magnitude as we vary the mortality probability, the end result is always a large fraction of the population dying (from 75–100%).

Bottom Line: While many studies have explored the influence of individual epidemiological parameters and/or underlying network topologies on the resulting disease dynamics, we here provide a systematic overview of the interactions between these two influences on population-level disease outcomes.Lastly, we discuss how the expected individual-level disease burden is influenced by the complete suite of epidemiological characteristics for the circulating disease and the ongoing process of network compromise.Our results have broad implications for prediction and mitigation of outbreaks in both natural and human populations.

View Article: PubMed Central - PubMed

Affiliation: Department of Ecology, Evolution, and Natural Resources, Rutgers University - New Brunswick, NJ 08901, United States of America; DIMACS, Rutgers University - Piscataway, NJ 08854, United States of America.

ABSTRACT
As the understanding of the importance of social contact networks in the spread of infectious diseases has increased, so has the interest in understanding the feedback process of the disease altering the social network. While many studies have explored the influence of individual epidemiological parameters and/or underlying network topologies on the resulting disease dynamics, we here provide a systematic overview of the interactions between these two influences on population-level disease outcomes. We show that the sensitivity of the population-level disease outcomes to the combination of epidemiological parameters that describe the disease are critically dependent on the topological structure of the population's contact network. We introduce a new metric for assessing disease-driven structural damage to a network as a population-level outcome. Lastly, we discuss how the expected individual-level disease burden is influenced by the complete suite of epidemiological characteristics for the circulating disease and the ongoing process of network compromise. Our results have broad implications for prediction and mitigation of outbreaks in both natural and human populations.

No MeSH data available.


Related in: MedlinePlus