California state mosquito-borne virus surveillance and response plan: A retrospective evaluation using conditional simulations

The California Mosquito-Borne Virus Surveillance and Response Plan recently was developed to provide a semi-quantitative means for assessing risk for western equine encephalomyelitis (WEE) or St. Louis encephalitis (SLE) viruses and to provide intervention guidelines for mosquito control and public health agencies during periods of heightened risk for human infection. West Nile virus recently has arrived in California, and the response plan also will provide a baseline for assessing the risk for human and equine infection with this virus. In the response plan, overall risk is calculated by averaging risk due to 1) environmental conditions, 2) adult mosquito vector abundance, 3) vector infection rates, 4) sentinel chicken seroconversion rates, 5) equine cases (for WEE), 6) human cases, and 7) the proximity of virus activity to populated areas. Overall risk is categorized into three levels: normal season, emergency planning, or epidemic conditions. We evaluated this response plan using historical data from years with no, enzootic, and epidemic activity of WEE and SLE in several areas of California to determine whether calculated risk levels approximated actual conditions. Multiple methods of risk calculation were considered for both viruses. Assessed risk based on cumulative temperature, rainfall, and runoff levels over the entire season provided more or equally accurate assessments than biweekly assessments based solely on the previous half-month. For WEE, during years with enzootic activity or earlyseason periods of years with WEE epidemic activity, combining horse and human cases as a single risk factor improved the model's ability to forecast pending WEE activity, but separating the two factors allowed a better indication of WEE activity during epidemics and periods with no activity. For SLE, assignment of higher risk to drier conditions as measured by rainfall and runoff yielded the most accurate representation of actual virus activity during all recent study periods.