In India, majority outbreaks of highly pathogenic avian influenza (HPAI) H5N1 have occurred in eastern states of West Bengal, Assam and Tripura. This study aimed to identify disease clusters and risk factors of HPAI H5N1 in these states, for targeted surveillance and disease control. A spatial scan statistic identified two significant disease clusters in West Bengal and Assam, occurring during January and November-December 2008, respectively. Key risk factors were identified at sub-district level using bootstrapped logistic regression and boosted regression trees model. With both methods, HPAI H5N1 outbreaks in backyard poultry were associated with accessibility in terms of time taken to access a city with >50,000 persons, human population density and duck density (P<0.005). In addition, areas at lower elevation were also identified as high risk by BRT model. It is recommended that risk-based surveillance should be implemented in high duck density areas and all live-bird markets in high-throughput locations.
【저자키워드】 Risk factors, Disease surveillance, Eastern India, HPAI H5N1 in poultry, Predictive risk modeling, Spatial and temporal clusters.,