Fig 1.
Each point shows paired values from the same donor, and the contours are of probability density. The red points, above the solid orange diagonal line, have fold changes greater than the threshold value in Figure A in S1 File, i.e. are inferred seroconversions. Those below the orange diagonal line have a bimodal distribution. The vertical and horizonal dashed lines show S/C = 0.49. So, for example, those to the left of the vertical dashed line are defined as initially negative. Dark blue points, with first S/C ≥0.49, are defined as positive on the first donation, with any increase to the second donation being less than the threshold for fold change. Light blue points are negative for both timepoints, except for a small proportion in the right triangle below the orange line but in the upper left quadrant: These are above the fixed positivity threshold but with a low fold change. The contours show that the blue points are segregated into a bimodal distribution, shown as light and dark blue.
Fig 2.
Force of infection over calendar time.
The right axis, and the red line (red), show the daily force of infection estimated from within-person changes in Abbott S/C values. Dashed lines are 95% confidence limits. The blue line is the rolling daily average, over one week, of suspected COVID-19 hospitalizations in the SIVEP-Gripe database, lagged by one week [24] to reflect the incubation period, with a smoothed line in the same color. The ratio of the peak of the smooth line in the second wave to the peak in the first wave is 4.1.
Table 1.
Participants’ characteristics.
Table 2.
Force of infection by age and sex.
The event is inferred seroconversion based either on fold change, or change from below to above the absolute threshold. The person-years includes time between donations, and also from an inferred negative value to the first donation. In order to estimate the force of infection, periods after the first positive result for each person are not included, on the basis of not being at risk. People with missing age or sex are not included. Note that the overall rate cannot be obtained by simple division of the events and person-years between donations, because the exact times of the events were not observed.