Accuracy of Rapid Antigen Screening Tests for SARS-CoV-2 Infection at Correctional Facilities in Korea: March - May 2022.

Background: The number of confirmed cases of individuals with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection increased rapidly due to the Omicron variant. Correctional facilities are vulnerable to infectious diseases, and they introduced rapid antigen tests (RATs) to allow for early detection and rapid response. We aimed to evaluate the diagnostic performance and usefulness of SARS-CoV-2 RATs in newly incarcerated people.

Materials and methods: We conducted a cross-sectional study at correctional facilities in Korea from 9 March to 22 May 2022. The study population was newly incarcerated people who were divided into two groups. In one group, 799 paired SARS-CoV-2 RATs and real-time polymerase chain reaction (RT-PCR) were conducted simultaneously in 522 individuals in March 2022. In the other group, 4,034 paired RATs and RT-PCR consecutively in 4,034 participants; only individuals with negative RATs results underwent RT-PCR from April to May 2022. We analyzed data using descriptive statistics and a logistic regression model.

Results: Among the 799 specimens in March, RT-PCR was positive in 72 (9.0%), and among the 4,034 specimens in April - May 2022, RT-PCR was positive in 40 (1.0%). Overall, the RATs had a sensitivity of 58.3%, specificity of 100.0%, positive predictive value (PPV) of 100.0%, and negative predictive value (NPV) of 96.0%. Asymptomatic individuals constituted 98.2% of the study group, and symptomatic individuals 1.8%. In asymptomatic cases, the sensitivity of RATs was 52.5%, specificity was 100.0%, PPV was 100.0%, and NPV was 96.3%. In symptomatic cases, the sensitivity of RATs was 84.6%, specificity was 100.0%, PPV was 100.0%, and NPV was 33.3%. Sensitivity (P = 0.034) and NPV (P = 0.004) differed significantly according to the presence and absence of symptoms, and the F1 score was the highest at 0.9 in symptomatic individuals in March. There was a positive linear trend in the proportion of false-negative RATs in newly incarcerated people following the weekly incidence of SARS-CoV-2 (P = 0.033). The best-associated predictors of RATs for SARS-CoV-2 infection involved symptoms, timing of sample collection, and repeat testing.

Conclusion: Sensitivity and NPV significantly depend on whether symptoms are present, and the percentage of false negatives is correlated with the incidence. Thus, using RATs should be adjusted according to the presence or absence of symptoms and the incidence of SARS-CoV-2 infection in the community. RATs could be a useful screening tool as an effective first-line countermeasure because they can rapidly identify infectious patients and minimize SARS-CoV-2 transmission in correctional facilities.