Predictors of influenza seasonal characteristics by virus type (influenza A, B, and total) in 30 Chinese provinces, 2005–2011.

Putative predictors include climate, geographic, population, and sampling factors. Seasonal characteristics were based on linear seasonal regression models fitted to province-specific weekly laboratory surveillance data (Figure 4). Point estimates and standard errors provided in the table are based on hierarchical Bayesian regression of seasonal characteristics against province-specific covariates, with observation variances estimated by bootstrap resampling of seasonal parameters.aAmplitude of semi-annual cycle divided by the sum of amplitudes of semi-annual and annual cycles.b5% threshold criterion.cNo covariate remained in final model.dAnnual number of passengers leaving or entering the province, divided by province-specific population size.