Academic journal article Canadian Journal of Public Health

Predicting Geographical Human Risk of West Nile Virus - Saskatchewan, 2003 and 2007

Academic journal article Canadian Journal of Public Health

Predicting Geographical Human Risk of West Nile Virus - Saskatchewan, 2003 and 2007

Article excerpt

ABSTRACT

Objectives: To detail the use of a model to predict areas of low, medium, and high risk of West Nile virus (WNV) in humans in both 2003 and 2007 in the province of Saskatchewan. To identify consistent high-risk areas from year to year as well as important environmental variables within those high-risk areas.

Methods: The number of laboratory-confirmed WNV individuals was obtained from Saskatchewan Health by rural municipality. The population at risk was obtained from Statistics Canada by rural municipality. Climate and habitat variables were incorporated into a discriminant analysis model with the production of risk maps as an end product.

Results: The discriminant analysis models had testing classification accuracies of 67% in 2003 and 44% in 2007. Climate and habitat variables remained important in all models while some habitat variables were less important in 2007. Risk maps from historically trained 2007 model revealed a southwest to northeast decreasing trend of risk.

Conclusion: The models could be useful for indicating areas of high risk on a year-to-year basis or based on historical data. High-risk regions are characterized by less rainfall in June and July followed by higher temperatures in July and August with less vegetation and water coverage than low-risk regions.

Key words: Zoonoses; arboviruses; Saskatchewan; public health

La traduction du résumé se trouve à la fin de l'article. Can J Public Health 2009;100(5):344-48.

The introduction of West Nile virus (WNV) into North America sparked an interest in predicting where and when the virus would appear.1 Predictive risk mapping is a process by which components of the disease cycle are used to create models and subsequent risk maps.2,3 The methods have become more practical for a broader range of diseases and study locations because remote sensing can now provide environmental information at required spatial and temporal resolution.4,5

Vector-borne diseases, such as WNV, are particularly amenable to spatial and temporal analysis as they are highly influenced by regular, seasonal climate, and environmental changes.3,6-8 During the season, mosquitoes become infected with the West Nile virus primarily through bird-blood meals and then retransmit the virus to any one of multiple bird species, a cycle which amplifies the virus. Governed by environmental conditions and host behaviours, infected mosquitoes can spread WNV to other incidental hosts, such as humans and horses.

Defining the risk of WNV infection is a key component to public health intervention strategies.9 Prioritization of vector-borne disease programs in the overall public health budget is a juggling act, affected by limited funding availability. In Saskatchewan, interventions are prioritized largely based on environmental conditions conducive to mosquito development and surveillance for clinical disease in humans. Health officials could increase the cost effectiveness of control and surveillance programs with a method of predicting differences in regional risk of infection.

The primary objective of this study was to describe the application of a previously established model to predict areas of low, medium, and high risk of WNV in humans in both 2003 and 2007 in the province of Saskatchewan.10 The second objective was to use historical surveillance data from 2003-2005 to make predictions of areas of risk of WNV in humans in 2007.

MATERIALS AND METHODS

WNV infection risk - 2003 and 2007

Human surveillance data were obtained from Saskatchewan Health as the number of laboratory-confirmed WNV individuals (which included WN fever, WN neurological syndrome and asymptomatic individuals, http://www.health.gov.sk.ca/wnv-surveillance-resultsarchive (accessed July 14, 2009)) per rural municipality (RM). In 2003, each RM with WNV individuals (sampled RM) was classified by category of WNV infection risk using the 25th and 75th percentiles: low-risk (0. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.