Evaluating real-world data in COVID-19 antigen and PCR testing

COVID-19 testing is an important pillar in fighting the SARS-CoV-2 pandemic. Even though billions of tests have been conducted, little is known on performance of testing sites. This is a retrospective observational study with real-world data from a pharmacist-led COVID-19 testing center in Germany during the Omicron subvariant BA.4 and BA.5 wave in June 2022. The rapid antigen tests employed were purchased from Anbio Biotechnology (Xiamen, China). The RT-PCR was performed on Egens YS-qPCR-1 real-time system with Kewei multiple real-time PCR kits for detection of SARS-CoV-2 (Beijing Kewei Clinical Diagnostic Reagent Inc., Beijing, China). The study followed the STARD 2015 guideline. In this retrospective cohort study, the performance of testing sites was compared. During the study period, 7,112 patients were tested by rapid antigen tests and 1,025 RT-PCR tests conducted. Included were 233 patients who were referred by other testing sites for confirmation of positive results. A positive predictive value of 99.6% was calculated for the antigen tests in the pharmacist-led testing center. Referred positive patients from non-medical sites were antigen and RT-PCR negative in 16 cases, which led to a positive predictive value of 88.8%. Difference between site performance was statistically significant (p < 0.05). Results indicate that nucleic acid amplification confirmation is crucial in the context of the current testing strategy in Germany. Higher standards in antigen-testing, however, can make nucleic acid amplification in active COVID-19 infections unnecessary and testing cost efficient. This study provides the first data in the world on COVID-19 testing performance, and how it can be optimized.