Expert Elicitation Provides a Rapid Alternative to Formal Case-Control Study of an H7N9 Avian Influenza Outbreak in the United States

摘要

An expert elicitation was staged to rapidly decipher plausible routes and risks of pathogen transmission in the 2017 H7N9 avian influenza (AI) outbreak in the four-state region of Tennessee, Alabama, Georgia, and Kentucky. The process included the identification of risk factors found in a preponderance of commercial broiler breeder case farms over matched controls and an opinion-based weighting of risks and mitigations perceived influential to this outbreak. Although the two highly pathogenic AI case farms had general location and company ownership in common, obvious connections were lacking for the remainder of H7N9-infected (all low pathogenicity) commercial farms. Expert elicitation of differences between known cases and controls suggested a key role for environmental rather than lateral (business network) pathways in the distribution of low pathogenicity AI across commercial broiler breeder operations. Factors with greatest strength as predictors of disease, whether or not they were causal, included mesopredator or rodent incursions, enclosure defects, and habitat disturbance that might attract wildlife to the farm (e.g., feed spills and vacating of neighboring properties). Business affiliations that may have facilitated farm-to-farm transfer, in contrast, were limited. Biosecurity standards varied across this study group but were no more or less stringent among cases over controls. However, results from a parallel hypothetical scenario staged to address field data gaps suggest that uniformity and consistency in the implementation of biosecurity practices may impact risk of disease introduction. Importantly, this study was conducted within a few weeks and with little disruption to emergency response activities. As such, the approach offers an alternative model for interim field investigation of new or emerging high-consequence diseases with immediate decision support needs.