Weather variability and influenza A (H7N9) transmission in Shanghai, China: A Bayesian spatial analysis

Abstract

Background: A novel avian influenza A (H7N9) virus was first found in humans in Shanghai, and infected over 433 patients in China. To date, very little is known about the spatiotemporal variability or environmental drivers of the risk of H7N9 infection. This study explored the spatial and temporal variation of H7N9 infection and assessed the effects of temperature and rainfall on H7N9 incidence. Methods: A Bayesian spatial conditional autoregressive (CAR) model was used to assess the spatiotemporal distribution of the risk of H7N9 infection in Shanghai, by district and fortnight for the period 19th February-14th April 2013. Data on daily laboratory-confirmed H7N9 cases, and weather variability including temperature (°C) and rainfall (mm) were obtained from the Chinese Information System for Diseases Control and Prevention and Chinese Meteorological Data Sharing Service System, respectively, and aggregated by fortnight. Results: High spatial variations in the H7N9 risk were mainly observed in the east and centre of Shanghai municipality. H7N9 incidence rate was significantly associated with fortnightly mean temperature (Relative Risk (RR): 1.54; 95% credible interval (CI): 1.22-1.94) and fortnightly mean rainfall (RR: 2.86; 95% CI: 1.47-5.56). Conclusion: There was a substantial variation in the spatiotemporal distribution of H7N9 infection across different districts in Shanghai. Optimal temperature and rainfall may be one of the driving forces for H7N9.