Table 2.
Diagnostic accuracy variables of three rapid antigen tests, with different sampling methods. Values are percentages (95% confidence interval) unless stated otherwise
| Analysis | Sampling method | No. | Prevalencea (%) | Sensitivity | Specificity | PPV | NPV |
|---|---|---|---|---|---|---|---|
| BD-Veritor System (Beckton Dickinson) | |||||||
| Primary analysis | OP-N | 1441 | 13.0 | 68.6 (61.5 to 75.2) | 99.8 (99.4 to 100.0) | 98.5 (94.6 to 99.8) | 95.5 (94.2 to 96.6) |
| Secondary (stratified) analysis | |||||||
| Viral load above the cut-offc | OP-N | 1441 | 10.1 | 85.6 (78.9 to 90.9) | 99.5 (99.0 to 99.8) | 95.4 (90.3 to 98.3) | 98.4 (97.6 to 99.0) |
| Symptoms present at samplingb | |||||||
| Yes | OP-N | 662 | 18.1 | 75.8 (67.2 to 83.2) | 99.8 (99.0 to 100.0) | 98.9 (94.1 to 100.0) | 94.9 (92.8 to 96.6) |
| No | OP-N | 742 | 8.0 | 55.9 (42.4 to 68.8) | 99.9 (99.2 to 100.0) | 97.1 (84.7 to 99.9) | 96.3 (94.7 to 97.6) |
| SD-Biosensor (Roche Diagnostics) | |||||||
| Primary analysis | NP | 1769 | 12.2 | 74.4 (68.0 to 80.1) | 99.8 (99.4 to 100.0) | 98.2 (94.7 to 99.6) | 96.6 (95.6 to 97.4) |
| OP-N | 1689 | 9.7 | 75.0 (67.7 to 81.4) | 99.8 (99.4 to 100.0) | 97.6 (93.2 to 99.5) | 97.4 (96.5 to 98.1) | |
| Secondary (stratified) analysis | |||||||
| Viral load above the cut-offc | NP | 1769 | 10.3 | 87.9 (82.3 to 92.3) | 99.8 (99.4 to 100.0) | 98.2 (94.7 to 99.6) | 98.6 (97.9 to 99.1) |
| OP-N | 1689 | 8.3 | 83.7 (76.5 to 89.4) | 99.5 (99.0 to 99.8) | 93.7 (87.9 to 97.2) | 98.5 (97.8 to 99.1) | |
| Symptoms present at samplingb | |||||||
| Yes | NP | 1091 | 13.8 | 83.4 (76.5 to 89.0) | 99.8 (99.2 to 100.0) | 98.4 (94.5 to 99.8) | 97.4 (96.2 to 98.3) |
| OP-N | 900 | 12.7 | 78.9 (70.3 to 86.0) | 99.7 (99.1 to 100.0) | 97.8 (92.4 to 99.7) | 97.0 (95.6 to 98.1) | |
| No | NP | 658 | 9.6 | 54.0 (40.9 to 66.6) | 99.8 (99.1 to 100.0) | 97.1 (85.1 to 99.9) | 95.3 (93.4 to 96.9) |
| OP-N | 735 | 6.3 | 63.0 (47.5 to 76.8) | 100.0 (99.5 to 100.0) | 100.0 (88.1 to 100.0) | 97.6 (96.2 to 98.6) | |
| PanBio (Abbott) | |||||||
| Primary analysis | NP | 2056 | 8.4 | 68.8 (61.3 to 75.6) | 99.9 (99.7 to 100.0) | 99.2 (95.4 to 100.0) | 97.2 (96.4 to 97.9) |
| Secondary (stratified) analysis: | |||||||
| Viral load above the cut-offcd | NP | 2039 | 5.9 | 89.3 (82.3 to 94.2) | 99.9 (99.6 to 100.0) | 98.2 (93.6 to 99.8) | 99.3 (98.9 to 99.6) |
| Symptoms present at samplingb | |||||||
| Yes | NP | 1470 | 9.0 | 72.2 (63.7 to 79.6) | 99.9 (99.6 to 100.0) | 99.0 (94.4 to 100.0) | 97.3 (96.3 to 98.1) |
| No | NP | 511 | 6.7 | 55.9 (37.9 to 72.8) | 100.0 (99.2 to 100.0) | 100.0 (82.4 to 100.0) | 97.0 (95.0 to 98.3) |
NC not calculated because all Ag-RDT results were negative, NP deep nasopharyngeal, OP-N combined oropharyngeal and nasal sampling, PPV positive predictive value, NPV negative predictive value
aSARS-CoV-2 infection based on the molecular test result
bSymptoms not available for 37, 20, 53, and 75 participants, including 9, 1, 4, and 7 with a positive molecular test result, in the BD-Veritor group, SD-Biosensor NP group, SD-Biosensor OP-N group, and PanBio NP group, respectively
cThe viral load cut-off was 5.2 log10 SARS-CoV-2 E-gene copies/mL. This was the viral load above which 95% of people with a positive RT-PCR test result had a positive viral culture in a recent study by our group [6]
dViral load unavailable for 17 participants in the PanBio NP group