Haogao Gu1, Wangjian Zhang1, Hao Xu1, Pengyuan Li1, Luolin Wu2, Pi Guo1, Yuantao Hao1, Jiahai Lu1, Dingmei Zhang3. 1. Department of Medical Statistic and Epidemiology, School of Public Health. 2. Department of Atmospheric Science, School of Environmental Science and Engineering, Sun Yat-Sen University, Guangzhou 510080, China. 3. Department of Medical Statistic and Epidemiology, School of Public Health; Email: zhdingm@mail.sysu.edu.cn.
Abstract
OBJECTIVE: To establish a risk early warning model of human infection with avian influenza A (H7N9) virus and predict the area with high risk of the outbreak of H7N9 virus infection. METHODS: The incidence data of human infection with H7N9 virus at prefecture level in China from February 2013 to June 2014 were collected, and the geographic and meteorological data during the same period in these areas were collected too. Spatial auto regression (SAR) model and generalized additive model (GAM) were used to estimate different risk factors. Afterwards, the risk area map was created based on the predicted value of both models. RESULTS: All the human infections with H7N9 virus occurred in the predicted areas by the early warning model in February 2014. The early warning model successfully predicted the spatial moving trend of the disease, and this trend was verified by two outbreaks in northern China in April and May 2014. CONCLUSION: The established early warning model showed accuracy and precision in short-term prediction, which might be applied in the active surveillance, early warning and prevention/control of the outbreak of human infection with H7N9 virus.
OBJECTIVE: To establish a risk early warning model of humaninfection with avian influenza A (H7N9) virus and predict the area with high risk of the outbreak of H7N9 virus infection. METHODS: The incidence data of humaninfection with H7N9 virus at prefecture level in China from February 2013 to June 2014 were collected, and the geographic and meteorological data during the same period in these areas were collected too. Spatial auto regression (SAR) model and generalized additive model (GAM) were used to estimate different risk factors. Afterwards, the risk area map was created based on the predicted value of both models. RESULTS: All the human infections with H7N9 virus occurred in the predicted areas by the early warning model in February 2014. The early warning model successfully predicted the spatial moving trend of the disease, and this trend was verified by two outbreaks in northern China in April and May 2014. CONCLUSION: The established early warning model showed accuracy and precision in short-term prediction, which might be applied in the active surveillance, early warning and prevention/control of the outbreak of humaninfection with H7N9 virus.