Turning points, reproduction number, and impact of climatological events for multi-wave dengue outbreaks

To study climatological and public health events which might have affected the 2007 two-wave dengue outbreak in Taiwan, an island with both tropical and subtropical regions, where the 2007 dengue incidence exceeded the combined total of the previous four years. A multi-phase Richards model was fitted to weekly cumulative dengue data to pinpoint the turning points of the outbreak. We obtained the 'initial' reproduction numbers for the two waves of the outbreak. By means of correlation analysis we explored the possible impact of climatological events on the occurrence of turning points. Three turning points occurred around early August, late August/early September, and late October/early November. The 'initial' reproduction number for the first wave was Ri = 4.67 (95% CI: 0*-10.92), where 0* = max{0, lower bound}, and Ri = 3.93 (95% CI: 1.74-6.13) for the second wave. The highest correlation was between dengue incidence and two climatological variables: maximum temperature at a lag of 5 weeks (r = 0.66 and 0.71) and total precipitation at a lag of seven weeks (r = 0.53). The first two turning points were partially attributable to two typhoons around early to mid-August that brought a sharp drop in temperature and substantial rainfall. The drop in temperature first drove the dengue incidence down, then the rainfall drove it up at the beginning of fall. In recent years, Taiwan has witnessed increasingly frequent large summer dengue outbreaks that persisted into early winter, perhaps due to warmer autumns. This highlights the possible impact of global warming on the spread of infectious diseases.