Objectives: Compared with reverse transcription polymerase chain reaction (RT-PCR), rapid antigen detection tests (RADTs) for SARS-CoV-2 diagnostics are faster, less expensive, but also less accurate. Performance of RADTs is context-specific and depends on disease prevalence. In this real-world study, we assessed impact of RADTs in an inpatient setting through the entire COVID-19 emergency phase.
Methods: In this matched retrospective study, data on RT-PCR and RADT laboratory diagnoses of SARS-CoV-2 made between February 2020 and May 2023 in a large hospital were analyzed. To be included in the study, samples used for both RT-PCR and RADT had to be collected on the same day.
Results: Of 278,867 RT-PCR tests available, 13,321 same-day RADTs could be matched to RT-PCR. Over the entire period, RADT sensitivity and specificity were 76.4% and 99.4%, respectively. With the observed positivity rate of 9.4%, positive and negative predictive values were 92.7% and 97.6%, respectively. Compared with the periods dominated by the Alpha and Delta variants of concern, RADT accuracy was slightly lower during the Omicron phase.
Conclusion: This real-world experience demonstrates that despite suboptimal sensitivity and some variation by predominant variants of concern and positivity prevalence, the use of RADTs is useful in hospital settings. Public health implications were discussed.
Keywords: COVID-19; Diagnostic accuracy; Rapid antigen detection test; Real-world performance; SARS-CoV-2.
Copyright © 2023 The Author(s). Published by Elsevier Ltd.. All rights reserved.