Fig. 5.

Protective effect through time. For each sera-antigen pair analysed in Bedford et al., this shows how the average protection (defined as reduction in nasal shedding) predicted by the HI titre of serum against antigen is affected by the of number of years difference between the date of isolation of the serum and the date of isolation of the antigen (positive numbers mean that the antigen was collected after the serum). Comparisons between sera and antigen that were isolated close together in time are much more common than between those isolated far apart in time and so this represents the median and 50% CI for predicted protection based on a bootstrap of 1000 samples for each measurement.