Comparison of epidemical features of seasonal influenza across different climatic zones in Australia

Background: Seasonal influenza epidemic patterns have been highly influenced by weather and usually vary by tropical, subtropical and temperate climates. Few studies investigate the features of seasonal influenza in different age groups among geo-climatic regions according to a specific climatic condition. Objective: This study aimed to assess the differences in the epidemical features of influenza A and B among six climatic zones in three age groups (<15, 15-64 and 65+ years) in Australia. Methods: National Notifiable Diseases Surveillance System (NNDSS) data on weekly laboratory-confirmed cases of influenza A and B at a postcode level were collected from the Australian Government Department of Health between 1st January 2011 and 31st December 2013. Spatial and temporal descriptive analyses and Dunnett-Tukey-Kramer (DTK) pairwise multiple comparison tests were used to investigate the differences in seasonal patterns, durations, peak timings and epidemic magnitude for influenza A and B, stratified by the six climatic zones and age group. Bayesian space-time models based on a spatial conditional autoregressive (CAR) model combined with a susceptible, infectious and removed (SIR) model was used to estimate transmission rates to explore differences in evolution of influenza A and B epidemics. Results: There were significant differences in mean weekly notification rates of influenza A and B among the six climatic zones in the 0-14 and 15-64 age groups. Mean weekly notification rates were more likely to be higher in the areas with a warm winter or a mild winter than in the area with relatively colder winter. The ≥65 age group showed less spatial variation in mean weekly notification rates of influenza A and B among the six climatic zones. Mean duration, peak timing and transmission rates of influenza A and B epidemics did not display synchronicity between either the three age groups or the six climatic zones. The magnitude of the linear growth and decay rates of mean weekly transmission rates varied by different climatic zones and age groups. Conclusion: This study suggests that the epidemic features of influenza A and B vary between geo-climatic regions and age groups. Our findings provide valuable insight for public health authorities to adjust prevention and control strategies of seasonal influenza for specific age groups in specific climatic regions in Australia.