ARS | Publication request: Operationalizing climate-based epidemic prediction models: Rift Valley fever prediction system experience

Submitted to: Meeting Abstract
Publication Type: Abstract Only
Publication Acceptance Date: March 18, 2008
Publication Date: March 18, 2008
Citation: Chretien, J., Anyamba, A., Small, J., Tucker, C.J., Britch, S.C., Linthicum, K. 2008. Operationalizing climate-based epidemic prediction models: Rift Valley fever prediction system experience. Meeting Abstract.

Technical Abstract: Background There is considerable optimism that climate data and predictions will facilitate early warning of infectious disease epidemics. Interest in climate-based epidemic forecasting stems from climate-disease associations and global climate change (rising temperatures may extend arthropod vector habitats and enhance vectorial capacity, bringing infections to susceptible populations, and increase the frequency or severity of El Nino/Southern Oscillation, which has precipitated epidemics). Currently, though, there are few operational climate-based epidemic early warning systems. Methods We qualitatively assess 10 years of operation of our climate-based epidemic early warning system to identify key considerations in system development. Our system was designed originally using satellite observations, epidemiological data, and field experiments to forecast Rift Valley fever (RVF) epidemics in East Africa, and operationalized in a global epidemiological surveillance-response network. Results The system enabled model development and epidemic forecasts for various diseases, including East Africa RVF alerts beginning in September 2006, 3 months before a regional epidemic. Guided by geographical risk assessments, field partners in Kenya identified infected vectors and likely human cases in December, facilitating rapid international response. Other applications included warnings preceding yellow fever and RVF epidemics in Sudan, and retrospective identification of precursors to a chikungunya fever epidemic in Kenya which spread widely. Conclusion In our experience, key factors in successful system implementation are organizations bridging model developers and end-users, and epidemiological surveillance programs for model validation and public health response.