tetano
Editor, Senior Moderator
PLoS One. 2018 Jan 31;13(1):e0190824. doi: 10.1371/journal.pone.0190824. eCollection 2018.
[h=1]Identification of high risk areas for avian influenza outbreaks in California using disease distribution models.[/h] Belkhiria J[SUP]1[/SUP], Hijmans RJ[SUP]2[/SUP], Boyce W[SUP]3[/SUP], Crossley BM[SUP]4[/SUP], Mart?nez-L?pez B[SUP]1[/SUP].
[h=3]Author information[/h]
[h=3]Abstract[/h] The coexistence of different types of poultry operations such as free range and backyard flocks, large commercial indoor farms and live bird markets, as well as the presence of many areas where wild and domestic birds co-exist, make California susceptible to avian influenza outbreaks. The 2014-2015 highly pathogenic Avian Influenza (HPAI) outbreaks affecting California and other states in the United States have underscored the need for solutions to protect the US poultry industry against this devastating disease. We applied disease distribution models to predict where Avian influenza is likely to occur and the risk for HPAI outbreaks is highest. We used observations on the presence of Low Pathogenic Avian influenza virus (LPAI) in waterfowl or water samples at 355 locations throughout the state and environmental variables relevant to the disease epidemiology. We used two algorithms, Random Forest and MaxEnt, and two data-sets Presence-Background and Presence-Absence data. The models performed well (AUCc > 0.7 for testing data), particularly those using Presence-Background data (AUCc > 0.85). Spatial predictions were similar between algorithms, but there were large differences between the predictions with Presence-Absence and Presence-Background data. Overall, predictors that contributed most to the models included land cover, distance to coast, and broiler farm density. Models successfully identified several counties as high-to-intermediate risk out of the 8 counties with observed outbreaks during the 2014-2015 HPAI epizootics. This study provides further insights into the spatial epidemiology of AI in California, and the high spatial resolution maps may be useful to guide risk-based surveillance and outreach efforts.
PMID: 29385158 DOI: 10.1371/journal.pone.0190824
Free full text
[h=1]Identification of high risk areas for avian influenza outbreaks in California using disease distribution models.[/h] Belkhiria J[SUP]1[/SUP], Hijmans RJ[SUP]2[/SUP], Boyce W[SUP]3[/SUP], Crossley BM[SUP]4[/SUP], Mart?nez-L?pez B[SUP]1[/SUP].
[h=3]Author information[/h]
[h=3]Abstract[/h] The coexistence of different types of poultry operations such as free range and backyard flocks, large commercial indoor farms and live bird markets, as well as the presence of many areas where wild and domestic birds co-exist, make California susceptible to avian influenza outbreaks. The 2014-2015 highly pathogenic Avian Influenza (HPAI) outbreaks affecting California and other states in the United States have underscored the need for solutions to protect the US poultry industry against this devastating disease. We applied disease distribution models to predict where Avian influenza is likely to occur and the risk for HPAI outbreaks is highest. We used observations on the presence of Low Pathogenic Avian influenza virus (LPAI) in waterfowl or water samples at 355 locations throughout the state and environmental variables relevant to the disease epidemiology. We used two algorithms, Random Forest and MaxEnt, and two data-sets Presence-Background and Presence-Absence data. The models performed well (AUCc > 0.7 for testing data), particularly those using Presence-Background data (AUCc > 0.85). Spatial predictions were similar between algorithms, but there were large differences between the predictions with Presence-Absence and Presence-Background data. Overall, predictors that contributed most to the models included land cover, distance to coast, and broiler farm density. Models successfully identified several counties as high-to-intermediate risk out of the 8 counties with observed outbreaks during the 2014-2015 HPAI epizootics. This study provides further insights into the spatial epidemiology of AI in California, and the high spatial resolution maps may be useful to guide risk-based surveillance and outreach efforts.
PMID: 29385158 DOI: 10.1371/journal.pone.0190824
Free full text