Table 4.
Risk-based screening tools for women ordered by the area under receiver operating characteristic curve (AUC)
| Author | Year | Sample size | Country | Setting (target group) | External validation? | AUC | Sensitivity | Specificity |
|---|---|---|---|---|---|---|---|---|
| Pintye (2017) [46] | 2011-14 | 1304 | Kenya |
Antenatal clinics (pregnant women) |
Yes |
0.84 (95%CI 0.72-0.95) 0.76 (0.67-0.85) – simplified score |
Not reported | Not reported |
| Wand (2018) [47] | 2002-12 | 8982 | South Africa |
Part of trial (sexually active 16+) |
Yes |
0.75 - development 0.71 - validation |
83% (development) 80% (validation) |
33% (development) 32% (validation) |
| Wand (2012) [14] | 2003-06 | 1485 | South Africa |
Unclear (women 18-49) |
Yes |
0.73 (0.66-0.79) - development 0.79 (0.70-0.81) - validation |
88% (development) 90% (validation) |
32% (development) 36% (validation) |
| Balkus (2016) [49] | 2009-11 | 5029 | South Africa, Uganda, Zimbabwe |
Part of trial (women 18-40) |
Yes |
0.67 (0.64-0.70) – development 0.7 (0.65-0.75) – validation with HPTN035 0.58 (0.51-0.65) – validation with FEM-PrEP |
91% (development) 84% (HPTN035) 83% (FEM-PrEP) |
38% (development) 46% (HPTN035) 31% (FEM-PrEP) |
| Balkus (2016) [50] | 2016 | 1269 | Malawi, South Africa, Uganda, Zimbabwe |
Part of trial (women 18-40) to externally validate the tool. |
Yes | 0.66 (0.6-0.73) | 90% | 35% |
| Burgess (2018) [52] | 2018 | 444 | South Africa |
Unclear (women 18-40) |
Unclear |
0.66 (0.54-0.74) – overall 0.69 (0.6-0.78) - age <25 0.49 (0.3-0.63) – age >25 |
64% (overall) 78% (age <25) 58% (age >25) |
57% (overall) 49% (age <25) 38% (age >25) |
| Peebles (2018) [51] | 2015-18 | 5573 | South Africa |
A diverse range of settings across five provinces (women 18-35) |
Yes |
0.64 (0.6-0.67) – age 18-24 0.68 (0.62-0.73) – age 25-35 0.61 (0.58-0.65) – using VOICE score [49] |
48.6% (age 18-24) 78.6% (age 25-35) |
70.8% (age 18-24) 42.7% (age 25-35) |
| Burgess (2017) [48] | 2011-14 | 1115 | South Africa |
9 South African sites (sexually active, 18-30) |
No | 0.56 (0.5-0.62) |
96% (Risk score >3) 84% (Risk score >5) |
84% (Risk score >5) 23% (Risk score >5) |
The risk-based tools are provided for emergency department attendees (Supplementary Tables 4 and 5), primary care attendees (Supplementary Tables 6 and 7), hospital inpatients (Supplementary Tables 8 and 9), adults in the community (Supplementary Tables 10 and 11), STI clinic attendees (Supplementary Tables 12 and 13), incarcerated persons (Supplementary Tables 14 and 15), serodiscordant couples (Supplementary Tables 16 and 17), and people who inject drugs (Supplementary Tables 18 and 19). The risk of bias assessments are provided in Supplementary Tables 20-24.