Table 4.
Performance of the COR signature as a triage test in HIV-infected and uninfected persons based on results from our diagnostic cohort.
| HIV status | Sensitivity | Specificity | Number needed to screen∗ | False positives | False negatives |
|---|---|---|---|---|---|
| WHO TTP for Triage test, optimal characteristics | >95% | >80% | <55 | <11 | 1 |
| ACS COR signature HIV-uninfected | 100% | 80% | 50 | 10 | 0 |
| ACS COR signature HIV-infected | 65% | 80% | 77 | 16 | 1 |
| Performance of current screening tools in HIV-infected persons | |||||
| Symptom screening only∗∗ | 79% | 50% | 64 | 32 | 1 |
| Symptom screening and chest radiography∗∗ | 91% | 39% | 55 | 33 | 1 |
TTP-Target product profile (World Health Organization [WHO], 2014). ∗Number needed to screen to detect one case of TB. For these calculations a TB prevalence of 2% was used. False positives and negatives were calculated per number screened to detect one case of TB. ∗∗Estimates for symptom screening and chest radiography are based on results from Getahun et al. (2011).