Fig. 2. Efficacy of Ag-RDT screening interventions for reducing nosocomial SARS-CoV-2 incidence.

Points represent mean efficacy (across n = 10,000 simulations) for each of 26 screening interventions, arranged by timing of the screening intervention (days since initial outbreak detection, x-axis) and coloured by screening implementation (either as 1-round screening with no other testing, orange; as 1-round screening in combination with routine RT-PCR testing, purple; or as 2-round screening with routine RT-PCR testing, black). For 2-round screening, the first round was conducted on day 1, with points arranged according to the date of the second round (days 2–9). The solid horizontal line corresponds to mean efficacy of routine RT-PCR testing in absence of screening, which is conducted continuously over time and does not correspond to a specific date. Relative reductions in incidence were similar across LTCFs, but there was significant variation in the number of infections averted (Supplementary Fig. S8). Error bars (dashed lines for routine testing) correspond to 95% confidence intervals estimated by bootstrap resampling (n = 10,000). Baseline assumptions underlying simulations include: “low” community SARS-CoV-2 incidence; time-varying Ag-RDT sensitivity relative to RT-PCR (Ag-RDT A); and screening interventions that target all patients and staff in the LTCF. RT-PCR = reverse transcriptase polymerase chain reaction; Ag-RDT = antigen rapid diagnostic testing; LTCF = long-term care facility.