"Effects of species diversity on disease risk".
Reviewed 10/28/11
How does the presence of disease within an organism affect their ability to tolerate and survive? More broadly how does the presence of disease in a population or community of hosts affect the dynamics of that system, and how does the host community affect the pathogen? The study of the movement of disease in a community has long been an area of interest in ecology. The authors of this paper looked specifically at many recent experiments designed to tease out the role of host species diversity on disease prevalence and risk for a community.
Simple mathematical models were used at each stage in their discussion to illustrate how proposed mechanisms could be projectable onto a real ecosystem. For example an increase in the number of non-host species in a community could affect the encounter rate of a host and its pathogen. There are a number of ways it is possible to imagine how this would occur, if a predator induced its prey to move less and hide more, the prey would have fewer interaction events with its disease enemy and so disease risk would be driven down.
This is just one example of 'Encounter Reduction' that the authors propose as a mechanism for how diversity increases could affect the role of disease in a community. However, a rising level of diversity does also have the potential to increase disease risk. This is possible in systems where there is only one host of study, but it is maybe easier to imagine in a system with multiple hosts. If a community already has an abundance of low quality host species, and rising levels of diversity introduce a higher quality host, then the net effect would be one of an amplification of disease within the system.
The concept of a 'dilution host' was introduced by the authors in the later stages of the paper. As I see it this type of host introduction could be thought of as something of a keystone species. Just as we have included the concept of a keystone predator into our ecological vocabulary it is intriguing to think abou the possibility of a keystone disease regulator. A host that, when present in the system, regulates the prevalence of disease across the entire community. This could occur if the species is a poor reservoir for a pathogen, but a high-quality host for a predatory vector. It is unclear how common this type of species is in natural communities, but it is nevertheless an interesting venue to consider for conservation techniques.
The models described relied heavily on the ability to connect a mechanism for diversity effects with the net change in the density of infected hosts. There are other means of tracking disease risk in a system, and these include prevalence of the pathogen and density of infected vectors, among others. One area of future research would be to look at how using any of the various metrics as the measure of disease risk alters the outcome of the net effect of diversity on disease. Though the authors claimed to use the rate of change of infected hosts they often referred to disease risk in the paper simply as disease prevalence. I am wondering if their is a scenario that could arise within a natural system where the prevalence of disease remains high while the future risk of disease begins to decline. This scenario seems possible especially when the dynamics of a disease are primarily caused by non-population level drivers, such as seasonal or environmental. If this is indeed the case, than it seems as though using disease prevalence and disease risk interchangeably may not be the most appropriate. Future empirical or theoretical studies could illuminate this problem.
Friday, October 28, 2011
Thursday, October 27, 2011
Seabloom, E.W., Borer, E.T., Mitchell, C.E., & Power, A.G. Ecology. (2010).
"Viral diversity and prevalence gradients in North American Pacific Coast grasslands".
Reviewed 10/27/11
Host-pathogen interactions do not exist inside a bubble, they must be considered within a larger community and geographical context. It is possible to find multiple pathogens coexisting inside a single host species and/or to observe pathogens across a wide range of host species. The authors of this paper collected observational data at 26 separate field sites, analyzing the role of four Barley and Cereal Yellow Dwarf Viruses (B/CYDVs) on three grass host species. These sites spanned a breadth of close to fifteen degrees latitude and represent a wide range of environmental variables and community composition.
The key questions the researchers hoped to answer revolved around how host identity, environmental attributes, and host community of the site affected pathogen prevalence and diversity. Lastly researchers were interested in the coinfection of pathogens and whether or not that was regulated by the total available pathogen pool at a site. This coinfection can be thought of as synonymous with alpha diversity, a traditional measure of local scale diversity in ecology. Beta diversity in this study was the total possible pool of pathogens at a site, and as these viruses are aphid vectored, the turnover among co-inhabiting host species is of particular importance.
For this study system, prevalence was not found to correlate with coinfection. Statistical analyses of the model variance showed that the two metric of pathogen dynamics could have some minor drivers in common, but that overall the majority of the variation seen likely arises from different primary drivers, ie what primarily causes patterns of coinfection does not cause patterns of prevalence.Prevalence was found to be largely a result of site precipitation and soil nitrate patterns, and coinfection was most correlated with latitude.
Coinfection levels increased significantly the farther north the site occurred, and as global gamma diversity remained largely the same across all sites, this resulted in a net decrease in beta-diversity, lower pathogen turnover. The authors posit that this is the result of a higher density of generalist vectors for the pathogens found in the north. A higher rate of vector herbivory on the grass hosts in general would lead to an increase in the overall transmission of the separate pathogen species and therefore a higher level of coexisting pathogens within a single host. More experimental work would be needed to prove this mechanism however.
It is possible to think of latitude as being a collation of many possible environmental variables (such as sunlight, temperature, etc), however the fact that coinfection was strongly correlated with this metric and not with host species type lends evidence to the idea that maybe the viral coexistence within a single host species is a function of the underlying site-level environmental conditions. Viral coexistence and within host niche allocations is not a traditional example used to think about niche-theory, but if pathogen coexistence is somehow determined by resource availability, mediated by their mutual host of course, this raises many interesting questions for future work, including: what categories of pathogen might be able to coexist with one another (rust, viral, bacterial, etc) based on their differential niches? This sort of question can still apply to viral species such as B/CYDVs.
Reviewed 10/27/11
Host-pathogen interactions do not exist inside a bubble, they must be considered within a larger community and geographical context. It is possible to find multiple pathogens coexisting inside a single host species and/or to observe pathogens across a wide range of host species. The authors of this paper collected observational data at 26 separate field sites, analyzing the role of four Barley and Cereal Yellow Dwarf Viruses (B/CYDVs) on three grass host species. These sites spanned a breadth of close to fifteen degrees latitude and represent a wide range of environmental variables and community composition.
The key questions the researchers hoped to answer revolved around how host identity, environmental attributes, and host community of the site affected pathogen prevalence and diversity. Lastly researchers were interested in the coinfection of pathogens and whether or not that was regulated by the total available pathogen pool at a site. This coinfection can be thought of as synonymous with alpha diversity, a traditional measure of local scale diversity in ecology. Beta diversity in this study was the total possible pool of pathogens at a site, and as these viruses are aphid vectored, the turnover among co-inhabiting host species is of particular importance.
For this study system, prevalence was not found to correlate with coinfection. Statistical analyses of the model variance showed that the two metric of pathogen dynamics could have some minor drivers in common, but that overall the majority of the variation seen likely arises from different primary drivers, ie what primarily causes patterns of coinfection does not cause patterns of prevalence.Prevalence was found to be largely a result of site precipitation and soil nitrate patterns, and coinfection was most correlated with latitude.
Coinfection levels increased significantly the farther north the site occurred, and as global gamma diversity remained largely the same across all sites, this resulted in a net decrease in beta-diversity, lower pathogen turnover. The authors posit that this is the result of a higher density of generalist vectors for the pathogens found in the north. A higher rate of vector herbivory on the grass hosts in general would lead to an increase in the overall transmission of the separate pathogen species and therefore a higher level of coexisting pathogens within a single host. More experimental work would be needed to prove this mechanism however.
It is possible to think of latitude as being a collation of many possible environmental variables (such as sunlight, temperature, etc), however the fact that coinfection was strongly correlated with this metric and not with host species type lends evidence to the idea that maybe the viral coexistence within a single host species is a function of the underlying site-level environmental conditions. Viral coexistence and within host niche allocations is not a traditional example used to think about niche-theory, but if pathogen coexistence is somehow determined by resource availability, mediated by their mutual host of course, this raises many interesting questions for future work, including: what categories of pathogen might be able to coexist with one another (rust, viral, bacterial, etc) based on their differential niches? This sort of question can still apply to viral species such as B/CYDVs.
Tuesday, October 18, 2011
Cottingham, K.L., Lennon, J.T., & Brown, B.L. Front. in Ecol. and the Envr. (2005).
"Knowing when to draw the line: designing more informative ecological experiments."
Reviewed 10/18/11
Data analysis has at its root the experimental design of a project. Every scientist who hopes to one day analyze their data must begin with the most appropriate arrangement of treatments and replications across space and time. Two alternative methods for analysis are linear regression and analysis of variance (ANOVA), both of which have their pros and cons when it comes to information produced and statistical power of that information.
ANOVA has traditionally been used for discrete independent variables of the presence/absence or type-based, etc. It has also been used for continuous variables that can be grouped into a gradient of categories, such as levels of nutrients in a gradient, or classes of densities in a population. Regression on the other hand is strictly meant for fully continuous independent variables that have a linear relationship with the response. The relationship between the response and independent variables can be transformed, but the basic linear pattern must be met.
ANOVA and linear regression both have at their base the same mathematical model, the general linear model. The difference is that while regression operates to find the parameter estimate for the relationship of the independent and response variable, ANOVA creates dummy variable terms for each level of the discrete independent variable. Then can look at each level and determine whether it differs significantly from the other terms in the model. It becomes readily apparent then that the benefits of using ANOVA come from its power to look at each term separately, without tying to force any sort of pattern onto the relationship between the response and independent variables. The consequences, however, could be a model with an overabundance of terms, leading to a lack of statistical power and making it harder to tie relationships between terms to one another.
Choosing one method over the other for data analysis and experimental design can be easy in some cases: where there is no limit to the number of experimental units for example, or where an independent variable has no underlying continuity to it and so analysis must be done using discrete dummy terms. However, as mentioned before, it is more complicated a decision process when the continuous variable can be grouped.
After reading this review I have decided that there are several main questions to focus on. Limitations on the number of experimental units may require analysis using ANOVA, as more replication ability would be possible. If regression really is the desired method of analysis in this case, than the experiment should be designed so that a fall back of ANOVA is possible. If the relationship comes back as non-linear, than it will also be necessary to use ANOVA. Regression analysis requires that the response and the residuals be normally distributed; this is not as much of a requirement for ANOVA, but accuracy of measurement for the independent variable is critical.
Lastly, the benefits of regression are very simply described but can have huge effects on the quality of research produced. Estimates derived from regression analysis have a much higher statistical power for a lower R-squared value than does ANOVA. Regression is also much better equipped to denote the relationship between a response and independent variable. Those parameter estimates can then be fed into ecological models and used in future research, a great incentive for those running simulation based analyses.
Reviewed 10/18/11
Data analysis has at its root the experimental design of a project. Every scientist who hopes to one day analyze their data must begin with the most appropriate arrangement of treatments and replications across space and time. Two alternative methods for analysis are linear regression and analysis of variance (ANOVA), both of which have their pros and cons when it comes to information produced and statistical power of that information.
ANOVA has traditionally been used for discrete independent variables of the presence/absence or type-based, etc. It has also been used for continuous variables that can be grouped into a gradient of categories, such as levels of nutrients in a gradient, or classes of densities in a population. Regression on the other hand is strictly meant for fully continuous independent variables that have a linear relationship with the response. The relationship between the response and independent variables can be transformed, but the basic linear pattern must be met.
ANOVA and linear regression both have at their base the same mathematical model, the general linear model. The difference is that while regression operates to find the parameter estimate for the relationship of the independent and response variable, ANOVA creates dummy variable terms for each level of the discrete independent variable. Then can look at each level and determine whether it differs significantly from the other terms in the model. It becomes readily apparent then that the benefits of using ANOVA come from its power to look at each term separately, without tying to force any sort of pattern onto the relationship between the response and independent variables. The consequences, however, could be a model with an overabundance of terms, leading to a lack of statistical power and making it harder to tie relationships between terms to one another.
Choosing one method over the other for data analysis and experimental design can be easy in some cases: where there is no limit to the number of experimental units for example, or where an independent variable has no underlying continuity to it and so analysis must be done using discrete dummy terms. However, as mentioned before, it is more complicated a decision process when the continuous variable can be grouped.
After reading this review I have decided that there are several main questions to focus on. Limitations on the number of experimental units may require analysis using ANOVA, as more replication ability would be possible. If regression really is the desired method of analysis in this case, than the experiment should be designed so that a fall back of ANOVA is possible. If the relationship comes back as non-linear, than it will also be necessary to use ANOVA. Regression analysis requires that the response and the residuals be normally distributed; this is not as much of a requirement for ANOVA, but accuracy of measurement for the independent variable is critical.
Lastly, the benefits of regression are very simply described but can have huge effects on the quality of research produced. Estimates derived from regression analysis have a much higher statistical power for a lower R-squared value than does ANOVA. Regression is also much better equipped to denote the relationship between a response and independent variable. Those parameter estimates can then be fed into ecological models and used in future research, a great incentive for those running simulation based analyses.
Friday, October 14, 2011
Swinton, J., & Gilligan, C.A. Phil. Trnas. R. SOc. Lond. B (1996).
"Dutch elm disease and the future of the elm in the U.K.: a quantitative analysis".
Reviewed: 10/14/11
Dutch elm disease and its effects on the native populations within Great Britain have been a problem for longer than a century. Separate epidemics in the last century have been caused by two related strains of fungal pathogen, O. ulmi and O. novo-ulmi. The fungus is shuttled between recently dead trees, that function as a breeding grounds, by a beetle vector. The authors of this article have developed a model to look at the long-term predictions possible from a simple density based characterization of elm trees and their transition to infection and otherwise.
I have already mentioned that recently deceased trees function as a breeding ground for the vector that helps propagate the fungus in the system. However, long-dead trees are no longer able to fulfill this role and this time dependency was accounted for in the model parameters. A couple of assumptions were made by the authors when developing the basics of their model equations. One was that the vector density was proportional to the number of dead trees total based on previous research in the system of study. This allowed for simplification in the equations for overall force of infection.
The model was run against data collected by the National Forestry Commission of Britain in the 1970's, however only non-woodland systems were looked at in order to minimize sampling uncertainty. The most interesting results from their model showed how endemic levels of pathogen could be maintained in the tree populations depending on their lethality and the transmission levels. High lethality, it seems, actually causes a pulse in the system that could allow for recovery of the host tree populations and a lower overall level of endemic infection. This is also related to the reproductive rate of the fungal pathogen, where a low R-naught for the pathogen requires a very high degree of lethality to even reach a point where persistence is possible.
There are however several possible ways in which the model presented was confusing and could be improved. The authors posit that competition between the two fungal species plays a role in the long-term disease dynamics, yet they run their model without the inclusion of the less-aggressive species, O. ulmi, even when it became clear that their was a serious underestimation of infected and dead trees occurring in their model when compared to the data collected by the Forestry Commission. This study would also benefit from a greater amount of information on the life-history traits of both the tree genus and the fungi. The parameter estimates used in their models were chosen using a best-guess method in the absence of data pertaining to the life-history of the trees at juvenile and adult stages and transmission estimates for the pathogens. The fungi themselves have two roles in natural systems, saprophytic and parasitic; a better understanding of the transition between these two states would be of use in future studies.
Reviewed: 10/14/11
Dutch elm disease and its effects on the native populations within Great Britain have been a problem for longer than a century. Separate epidemics in the last century have been caused by two related strains of fungal pathogen, O. ulmi and O. novo-ulmi. The fungus is shuttled between recently dead trees, that function as a breeding grounds, by a beetle vector. The authors of this article have developed a model to look at the long-term predictions possible from a simple density based characterization of elm trees and their transition to infection and otherwise.
I have already mentioned that recently deceased trees function as a breeding ground for the vector that helps propagate the fungus in the system. However, long-dead trees are no longer able to fulfill this role and this time dependency was accounted for in the model parameters. A couple of assumptions were made by the authors when developing the basics of their model equations. One was that the vector density was proportional to the number of dead trees total based on previous research in the system of study. This allowed for simplification in the equations for overall force of infection.
The model was run against data collected by the National Forestry Commission of Britain in the 1970's, however only non-woodland systems were looked at in order to minimize sampling uncertainty. The most interesting results from their model showed how endemic levels of pathogen could be maintained in the tree populations depending on their lethality and the transmission levels. High lethality, it seems, actually causes a pulse in the system that could allow for recovery of the host tree populations and a lower overall level of endemic infection. This is also related to the reproductive rate of the fungal pathogen, where a low R-naught for the pathogen requires a very high degree of lethality to even reach a point where persistence is possible.
There are however several possible ways in which the model presented was confusing and could be improved. The authors posit that competition between the two fungal species plays a role in the long-term disease dynamics, yet they run their model without the inclusion of the less-aggressive species, O. ulmi, even when it became clear that their was a serious underestimation of infected and dead trees occurring in their model when compared to the data collected by the Forestry Commission. This study would also benefit from a greater amount of information on the life-history traits of both the tree genus and the fungi. The parameter estimates used in their models were chosen using a best-guess method in the absence of data pertaining to the life-history of the trees at juvenile and adult stages and transmission estimates for the pathogens. The fungi themselves have two roles in natural systems, saprophytic and parasitic; a better understanding of the transition between these two states would be of use in future studies.
Morozov, A.Y., Robin, C., & Franc, A. Jour. of Theor. Biol. (2007)
"A simple model for the dynamics of a host-parasite-hyperparasite interaction".
Reviewed 20/14/11
The study of parasites and their importance in ecological systems is growing. But what about hyperparasites? There does in fact exist parasites of parasites, and it is possible that their role in systems could be just as important, or more so, than any other tri-trophic interaction. The authors of this paper developed a theoretical model for the study of a host-parasite-hyperparasite interaction using a modification of the classic SIR models for epidemiology. The system they applied their model to includes the former chestnut tree populations in America, the ascomycete fungus C. parasitica, and its parasite Cryphonectrica Hypovirus (CHV). There are several strains of the hyperparasite all which affect the fungus in differing ways. The authors also included a section on the vegetative compatibility of viruses with the fungus. Depending on this interaction the number of infected spores produced by the infected fungus could alter.
The basics of the model centered around 4 different character states for the tree: Susceptible (S), Infected with hyperparasite-free fungus (I), Infected with a hyperparasite-bearing fungus (H), and Dead (R). Infected state individuals can die or be turned into the H state. It is assumed that trees infected with hyperparasite-bearing fungi are able to recover and return to the S state, but a transition to dead (R) was not included, assuming that the reduction of fungal pressure on the tree was so largely reduced by its own parasite that it could not cause death in its tree host.
Two conditions were set for hyperparasite establishment to occur. If both are met, but not one alone, than the establishment success of the hyperparasite, and their corresponding role as a method of biological control, depended next upon the initial conditions. One of these metrics includes a term stating that there needs to be less full recovery of the tree host than their is horizontal transmission of hyperparasite-bearing fungus. A certain number of H state trees need to remain in the system in order to provide new sources of hyperparasite. This also relates to the result that hyperparasite establishment is easier if parasite (the fungus) transmission is high. Evolutionarily I think this result is quite interesting. The high level of reproduction that would typically be considered advantageous for the parasitic fungus becomes less so if it also helps its enemy succeed.
Virulence levels for the hyperparasite also play a role in the model. It would seem that a more mid-line virulence would be more successful for endemic virus levels to persist; too high and fungal pressure on the tree becomes so reduced that the tree is able to recover at a faster rate than the hyperparasite can propagate itself in the system.
I found this paper to be very intriguing. It is interesting to explore the role of hyperparasites in natural systems, especially as their presence in the chestnut example may be a key reason for why stability of the tree populations in Europe was obtained after an introduction of the fungus while American populations were decimated to the point of extinction. The authors propose that their models could be altered into a framework useable for invasion into a community by transforming the character states of the hosts into states of patch or niche such as; susceptible to invasion (S), invaded (I), invaded by a enemy-controlled invader (H), etc. This is an interesting mental-exercise, but I am not sure of its application ability. The unit of study at the host level is easier as it tends to function more on a live or die scenario. Patches, or niches, are somewhat more complicated as it becomes more necessary to measure the extent of invasion in a patch, and then there is always the question of defining patch boundaries. Either way, it would seem as if the model presented has some merit for study in hyperparasitic systems in the future.
Reviewed 20/14/11
The study of parasites and their importance in ecological systems is growing. But what about hyperparasites? There does in fact exist parasites of parasites, and it is possible that their role in systems could be just as important, or more so, than any other tri-trophic interaction. The authors of this paper developed a theoretical model for the study of a host-parasite-hyperparasite interaction using a modification of the classic SIR models for epidemiology. The system they applied their model to includes the former chestnut tree populations in America, the ascomycete fungus C. parasitica, and its parasite Cryphonectrica Hypovirus (CHV). There are several strains of the hyperparasite all which affect the fungus in differing ways. The authors also included a section on the vegetative compatibility of viruses with the fungus. Depending on this interaction the number of infected spores produced by the infected fungus could alter.
The basics of the model centered around 4 different character states for the tree: Susceptible (S), Infected with hyperparasite-free fungus (I), Infected with a hyperparasite-bearing fungus (H), and Dead (R). Infected state individuals can die or be turned into the H state. It is assumed that trees infected with hyperparasite-bearing fungi are able to recover and return to the S state, but a transition to dead (R) was not included, assuming that the reduction of fungal pressure on the tree was so largely reduced by its own parasite that it could not cause death in its tree host.
Two conditions were set for hyperparasite establishment to occur. If both are met, but not one alone, than the establishment success of the hyperparasite, and their corresponding role as a method of biological control, depended next upon the initial conditions. One of these metrics includes a term stating that there needs to be less full recovery of the tree host than their is horizontal transmission of hyperparasite-bearing fungus. A certain number of H state trees need to remain in the system in order to provide new sources of hyperparasite. This also relates to the result that hyperparasite establishment is easier if parasite (the fungus) transmission is high. Evolutionarily I think this result is quite interesting. The high level of reproduction that would typically be considered advantageous for the parasitic fungus becomes less so if it also helps its enemy succeed.
Virulence levels for the hyperparasite also play a role in the model. It would seem that a more mid-line virulence would be more successful for endemic virus levels to persist; too high and fungal pressure on the tree becomes so reduced that the tree is able to recover at a faster rate than the hyperparasite can propagate itself in the system.
I found this paper to be very intriguing. It is interesting to explore the role of hyperparasites in natural systems, especially as their presence in the chestnut example may be a key reason for why stability of the tree populations in Europe was obtained after an introduction of the fungus while American populations were decimated to the point of extinction. The authors propose that their models could be altered into a framework useable for invasion into a community by transforming the character states of the hosts into states of patch or niche such as; susceptible to invasion (S), invaded (I), invaded by a enemy-controlled invader (H), etc. This is an interesting mental-exercise, but I am not sure of its application ability. The unit of study at the host level is easier as it tends to function more on a live or die scenario. Patches, or niches, are somewhat more complicated as it becomes more necessary to measure the extent of invasion in a patch, and then there is always the question of defining patch boundaries. Either way, it would seem as if the model presented has some merit for study in hyperparasitic systems in the future.
Friday, October 7, 2011
Viboud, C. et al. (2006). Science.
Synchrony, waves, and spatial hierarchies in the spread of influenza.
Reviewed: 10/07/11
There are 3 common subtypes of the influenza virus that have been affecting national health in the USA for 40 years or more; A/H3N2, A/H1N1, and B. The authors of this paper used mortality data for Pheumonia and Influenza over this time range to analyze the inter-state travel and timing of flue epidemics and pandemics. Ranking states by hte size of their populations and noting more intense pandemic years, they found that their is more synchrony in the epidemic as the size of the state population increases and the degree of intensity for the epidemic increases. They also found more synchrony between states that were in total a shorter distance from one another. This synchrony was measured in the correlation of peek epidemic weeks, a metric of timing, and through correlations of death measurements, a metric of intensity.
Mortality was used as a metric rather than cases reported because of the high probability when using number of reported cases that some person was either misdiagnosed with the disease or that the case was never reported at all due to low degree of symptoms. Using mortality does tend to make the predictions for the more severe subtype, A/H3N2, more accurate, but the researchers attempted to correct for sampling error in their measurements, and their results still showed that more intense mortality impacts were associated with the severer subtype and that seasons dominated by this subtype over the past 40yrs had more synchrony over all the states (as measured by standard deviation from the national average) in the timing of the disease.
Some of the most interesting results to come from this study however, are the results linking adult workflows to national and interstate trends in disease spread. Several types of movement were modeled, including air travel, rare long-distance events, and simple Euclidean movements, however when all of these movement types were adjusted for one another, workflow surfaces as the only significant correlation for disease phase and intensity. This is a pretty interesting result as it leads to the notion that adults and workflow dynamics are more important for regional spread of disease, not schoolchildren. Schoolchildren may be responsible for more local aspects of spread, but they do not play as large a role in the national dynamics. Incorporating age structure into other disease systems, including animal, plant, and marine could have similarly fascinating results, and should be investigated in the future. As long as their is some sort of inherent age structure in movements of natural populations, this dynamic should at least be considered as a mechanism for movement, whether or not it turns out to be significant to disease spread models.
This study and the variables used to mark disease trends were generally coarse; they were unable to get down to intra-state disease flows. However, the study was able to efficiently use simple predictors to model synchrony and timing at the national scale. This included results showing how initial disease foci in large states with a high degree of international capitalism, like California, tend to spread disease faster and lead to a higher degree of synchrony nationwide when compared to smaller less connected states such as Wyoming. It can be noted however, that their is a certain capping out effect for large states in the synchrony of events, indicating that higher transmission rates that are an inherent property of the virus itself may be more important than the population size of the initially infected disease focus.
Reviewed: 10/07/11
There are 3 common subtypes of the influenza virus that have been affecting national health in the USA for 40 years or more; A/H3N2, A/H1N1, and B. The authors of this paper used mortality data for Pheumonia and Influenza over this time range to analyze the inter-state travel and timing of flue epidemics and pandemics. Ranking states by hte size of their populations and noting more intense pandemic years, they found that their is more synchrony in the epidemic as the size of the state population increases and the degree of intensity for the epidemic increases. They also found more synchrony between states that were in total a shorter distance from one another. This synchrony was measured in the correlation of peek epidemic weeks, a metric of timing, and through correlations of death measurements, a metric of intensity.
Mortality was used as a metric rather than cases reported because of the high probability when using number of reported cases that some person was either misdiagnosed with the disease or that the case was never reported at all due to low degree of symptoms. Using mortality does tend to make the predictions for the more severe subtype, A/H3N2, more accurate, but the researchers attempted to correct for sampling error in their measurements, and their results still showed that more intense mortality impacts were associated with the severer subtype and that seasons dominated by this subtype over the past 40yrs had more synchrony over all the states (as measured by standard deviation from the national average) in the timing of the disease.
Some of the most interesting results to come from this study however, are the results linking adult workflows to national and interstate trends in disease spread. Several types of movement were modeled, including air travel, rare long-distance events, and simple Euclidean movements, however when all of these movement types were adjusted for one another, workflow surfaces as the only significant correlation for disease phase and intensity. This is a pretty interesting result as it leads to the notion that adults and workflow dynamics are more important for regional spread of disease, not schoolchildren. Schoolchildren may be responsible for more local aspects of spread, but they do not play as large a role in the national dynamics. Incorporating age structure into other disease systems, including animal, plant, and marine could have similarly fascinating results, and should be investigated in the future. As long as their is some sort of inherent age structure in movements of natural populations, this dynamic should at least be considered as a mechanism for movement, whether or not it turns out to be significant to disease spread models.
This study and the variables used to mark disease trends were generally coarse; they were unable to get down to intra-state disease flows. However, the study was able to efficiently use simple predictors to model synchrony and timing at the national scale. This included results showing how initial disease foci in large states with a high degree of international capitalism, like California, tend to spread disease faster and lead to a higher degree of synchrony nationwide when compared to smaller less connected states such as Wyoming. It can be noted however, that their is a certain capping out effect for large states in the synchrony of events, indicating that higher transmission rates that are an inherent property of the virus itself may be more important than the population size of the initially infected disease focus.
Hess, G.R. et al. (2001).
"Spatial aspects of disease dynamics"
Ch 6 from The Ecology of Wildlife Diseases (ed. Hudson P.J.)
Reviewed Oct 07, 2011
Disease risk at the local or global scale can be modeled in a variety of different ways. Analyses range from historical looking approaches that attempt to recreate using models an observed dynamic in space and time to forward looking approaches that hope to identify key areas for disease prevention action. Metapopulation theory has been of interest to ecologists for a variety of different reasons for several decades now. The basic model describing a series of patches in space that are all equidistant from one another and between which their is some level of immigration and recolonization by members of one patch into another. This dynamical process is believed to prevent large scale, or regional, extinction even if a local population dies out due to stochastic effects.
There are some problems with the classic metapopulation theory that have been addressed by other researchers in recent years, attempting to add in a large dominant patch, where extinction is not possible, or adding environmental variation into the patch quality structure. Work has also been done looking at the application of this theory to epidemiological systems, either modeling hosts themselves as a patch, or modeling a population of hosts as a patch. many of the variables used as inputs into metapopulation models have direct analogues with those for disease systems e.g. an empty patch equates to a susceptible host, extinction of a colony is similar to recovered host or one that died.
Some modifications to the metapopulation model have shown that a metapopulation designed like a wheel and spoke (a central, hub colony connected to outer, spoke colonies - which themselves are not interconnected) have a more consistent reduction in metapopulation extinction than some other spatial orientations including island and stepping-stone. This could have important ramifications for conservationists designing corridors hoping to save host populations without allowing the pathogen levels to also rise.
Many scientists also hope to link this particular spatial framework with landscape epidemiology. Landscape epidemiology has the potential to map disease risk using GIS layers that connect abiotic and biotic inputs to observed presence of pathogen, vector, or hosts. The method for connection of the variables to the disease system can either use statistics to rank the importance of collected factors on disease prevalence or utilize a more mechanistic approach that could link the biology of a pattern with changing conditions. Landscape epidemiology has a lot of important applications but there are some major issues with resolution of the spatial and temporal data due to pixel size from remotely sensed sources or limits to data archiving on those same sources. A balance between spatial accuracy and long-term analyses must be achieved if landscape epidemiology is to have any hope of being relevant for study at more regional or local scales.
Ch 6 from The Ecology of Wildlife Diseases (ed. Hudson P.J.)
Reviewed Oct 07, 2011
Disease risk at the local or global scale can be modeled in a variety of different ways. Analyses range from historical looking approaches that attempt to recreate using models an observed dynamic in space and time to forward looking approaches that hope to identify key areas for disease prevention action. Metapopulation theory has been of interest to ecologists for a variety of different reasons for several decades now. The basic model describing a series of patches in space that are all equidistant from one another and between which their is some level of immigration and recolonization by members of one patch into another. This dynamical process is believed to prevent large scale, or regional, extinction even if a local population dies out due to stochastic effects.
There are some problems with the classic metapopulation theory that have been addressed by other researchers in recent years, attempting to add in a large dominant patch, where extinction is not possible, or adding environmental variation into the patch quality structure. Work has also been done looking at the application of this theory to epidemiological systems, either modeling hosts themselves as a patch, or modeling a population of hosts as a patch. many of the variables used as inputs into metapopulation models have direct analogues with those for disease systems e.g. an empty patch equates to a susceptible host, extinction of a colony is similar to recovered host or one that died.
Some modifications to the metapopulation model have shown that a metapopulation designed like a wheel and spoke (a central, hub colony connected to outer, spoke colonies - which themselves are not interconnected) have a more consistent reduction in metapopulation extinction than some other spatial orientations including island and stepping-stone. This could have important ramifications for conservationists designing corridors hoping to save host populations without allowing the pathogen levels to also rise.
Many scientists also hope to link this particular spatial framework with landscape epidemiology. Landscape epidemiology has the potential to map disease risk using GIS layers that connect abiotic and biotic inputs to observed presence of pathogen, vector, or hosts. The method for connection of the variables to the disease system can either use statistics to rank the importance of collected factors on disease prevalence or utilize a more mechanistic approach that could link the biology of a pattern with changing conditions. Landscape epidemiology has a lot of important applications but there are some major issues with resolution of the spatial and temporal data due to pixel size from remotely sensed sources or limits to data archiving on those same sources. A balance between spatial accuracy and long-term analyses must be achieved if landscape epidemiology is to have any hope of being relevant for study at more regional or local scales.
Subscribe to:
Posts (Atom)