Test performance of lateral flow rapid antigen tests for COVID-19 in Welsh adult care home staff using routine surveillance data

Background Lateral flow tests (LFTs) have been used to screen for SARS-CoV2 in Wales since January 2021. Between May and August 2021, adult care home staff policy was for biweekly Innova LFT and weekly Polymerase Chain Reaction (PCR) testing while asymptomatic. We estimated test performance of LFTs conducted in adult care home staff using PCR tests as a reference standard. Methods Test results from surveillance data were matched by individual where both LFT and PCR were taken on the same day. We calculated sensitivity, specificity, positive and negative predictive values, and agreement using Matthew's correlation coefficient. Ct values of positive PCR results were compared by matched LFT result. Analysis was conducted using R v4.1.3. Results We analysed 115,593 test pairs, 499 (0.43%) of which were PCR positive. Median age was 48 (IQR: 22) and 85.00% of the study population were female. Test result agreement was 99.59% (95%CI 99.55-99.63; MCC: 0.38, p< 0.001). Sensitivity and specificity were 25.65% (95%CI 22.02-29.67) and 99.91% (95%CI 99.89-99.93), respectively. PPV was 55.90% (95%CI 49.42-62.17) and NPV was 99.68% (95%CI 99.64-99.71). Crude Ct values were significantly lower in positive PCR tests matched to a positive LFT compared to a negative LFT. Conclusions Specificity and negative predictive value were high in an asymptomatic population of care home staff indicating this test is an effective tool for identifying cases of SARS-CoV-2 infection during periods of high prevalence where transmission is likely, due to the presence of high viral loads. Positive predictive value results are lower than existing literature yet should be considered in light of the asymptomatic study population and low prevalence (under 1%) at the time most of these tests were conducted. Performance improved at times of higher prevalence during the study. These results suggest that whilst lateral flow tests are effective for identifying SARS-COV-2 infections with high viral loads, they are not effective at identifying cases with a low viral load. When an LFT provides a negative result, false negatives should be considered and additional diagnostic tests performed.