POCT | Sample | Parameters | Assuming imperfect reference standards (95% Crl)* |
Alere Determine | Serum | Sensitivity | 90.04% (80.45, 95.21) |
Specificity | 94.15% (89.26, 97.66) | ||
Whole blood | Sensitivity | 86.32% (77.26, 91.70) | |
Specificity | 95.85% (92.42, 97.74) | ||
SD Syphilis 3.0 | Serum | Sensitivity | 87.06% (75.67, 94.50) |
Specificity | 95.85% (89.89, 99.53) | ||
Whole blood | Sensitivity | 84.50% (78.81, 92.61) | |
Specificity | 97.95% (92.54, 99.33) | ||
Syphicheck-WB | Serum | Sensitivity | 74.48% (56.85, 88.44) |
Specificity | 99.14% (96.37, 100.0) | ||
Whole blood | Sensitivity | 74.47% (63.94, 82.13) | |
Specificity | 99.58% (98.91, 99.96) | ||
Visitect Syphilis | Serum | Sensitivity | 85.13% (72.83, 92.57) |
Specificity | 96.45% (91.92, 99.29) | ||
Whole blood | Sensitivity | 74.26% (53.62, 83.68) | |
Specificity | 99.43%, (98.22, 99.98) |
Adapted from Jafari et al.17
*Adjustments were made for imperfect reference standards using the Bayesian hierarchical summary receiver operating characteristic curve method. The results are point estimates of sensitivity and specificity for each test, using serum and whole blood, around a 95% credible interval (as opposed to a CI).
Crl, credible interval; NA, not available; POCT, point-of-care testing; TP, treponemal.