Fig. 1.

Pictogram representing outcomes of screening 100,000 people in (upper panel) higher prevalence setting (UK Office for National Statistics estimate of 1.33% test positivity in healthcare workers during period 27th April–10th May 2020 [30]) and (lower panel) lower prevalence setting (UK Office for National Statistics estimate of 0.052% test positivity in community during period 14th–27th June 2020 [31]), with sensitivity 70%, specificity 99.95%. To the left of each panel is an expanded view showing all true positive, false positive, and false negative results. Small rectangles represent 100 people, and medium rectangles represent 5000 people. In the top panel there are 1830 people with disease, of whom 1281 test positive (true positive) and 549 test negative (false negative), and 98,170 people without disease, of whom 49 test positive (false positive) and 98,121 test negative (true negative). The number needed to isolate to remove one infectious individual is 1.04. In the bottom panel there are 3 people with disease, of whom 2 test positive (true positive) and 1 tests negative (false negative), and 99,997 people without disease, of whom 50 test positive (false positive) and 99,947 test negative (true negative). The number needed to isolate to remove one infectious individual is 25.98. Note that using these data to generate full prevalence estimates with confidence intervals would require knowing the variability in sensitivity, which is difficult to determine.