Table 6. Performance results of the proposed diagnostic algorithm with different ACR score thresholds, T, where only HIV-uninfected and HIV-unknown subjects get ACR pre-screening and all HIV-infected subjects get Xpert testing.
| ACR threshold T | Selected for Xpert (%) | Sensitivity (%) | Specificity (%) | PPV (%) | NPV (%) | CSS ($) | CNTBC ($) | Throughput per day |
|---|---|---|---|---|---|---|---|---|
| 0 | 100 | 78.9 | 98.1 | 90.3 | 95.4 | 13.09 | 90.70 | 45 |
| 30 | 90 | 78.9 | 98.4 | 91.8 | 95.4 | 12.72 | 88.15 | 50 |
| 49 | 80 | 78.9 | 98.4 | 91.8 | 95.4 | 11.41 | 79.04 | 56 |
| 70 | 70 | 77.5 | 98.4 | 91.7 | 95.1 | 10.13 | 71.43 | 64 |
| 84 | 60 | 76.1 | 98.7 | 93.1 | 94.8 | 8.81 | 63.30 | 75 |
| 95 | 50 | 70.4* | 98.7 | 92.6 | 93.7 | 7.49 | 58.15 | 90 |
| 99 | 42 | 59.2* | 98.7 | 91.3 | 91.5 | 6.45 | 59.57 | 107 |
| 100 | 33 | 35.2* | 98.7 | 86.2 | 87.2 | 5.27 | 81.75 | 136 |
The first row matches a scenario where all subjects undergo Xpert and no ACR.
ACR, automated chest radiography; CNTBC, cost per notified TB case; CSS, cost per screened subject; NPV, negative predictive value; PPV, positive predictive value; Xpert, Xpert MTB/RIF.
*Sensitivity significantly different from Xpert standalone, T = 0. (p < 0.05 considered significant).