RESUMEN
In 2015, an outbreak of H5N2 highly pathogenic avian influenza (HPAI) occurred in the United States, severely impacting the turkey industry in the upper midwestern United States. Industry, government, and academic partners worked together to conduct a case-control investigation of the outbreak on turkey farms in the Upper Midwest. Case farms were confirmed to have HPAI-infected flocks, and control farms were farms with noninfected turkey flocks at a similar stage of production. Both case and control farms were affiliated with a large integrated turkey company. A questionnaire administered to farm managers and supervisors assessed farm biosecurity, litter handling, dead bird disposal, farm visitor and worker practices, and presence of wild birds on operations during the 2 wk prior to HPAI confirmation on case premises and the corresponding time frame for control premises. Sixty-three farms, including 37 case farms and 26 control farms were included in the analysis. We identified several factors significantly associated with the odds of H5N2 case farm status and that may have contributed to H5N2 transmission to and from operations. Factors associated with increased risk included close proximity to other turkey operations, soil disruption (e.g., tilling) in a nearby field within 14 days prior to the outbreak, and rendering of dead birds. Observation of wild mammals near turkey barns was associated with reduced risk. When analyses focused on farms identified with H5N2 infection before April 22 (Period 1), associations with H5N2-positive farm status included soil disruption in a nearby field within 14 days prior to the outbreak and a high level of visitor biosecurity. High level of worker biosecurity had a protective effect. During the study period after April 22 (Period 2), factors associated with HPAI-positive farm status included nonasphalt roads leading to the farm and use of a vehicle wash station or spray area. Presence of wild birds near dead bird disposal areas was associated with reduced risk. Study results indicated that the initial introduction and spread of H5N2 virus likely occurred by both environmental and between-farm pathways. Transmission dynamics appeared to change with progression of the outbreak. Despite enhanced biosecurity protocols, H5N2 transmission continued, highlighting the need to review geographic/topologic factors such as farm proximity and potential dust or air transmission associated with soil disruption. It is likely that biosecurity improvements will reduce the extent and speed of spread of future outbreaks, but our results suggest that environmental factors may also play a significant role in farms becoming infected with HPAI.