5. RDT types and brands verified with microscopy.
| RDT Brand | Study cohorts (n) | Patients (n) | P. falciparum cases (n) | Pooled sensitivity (95% CI) | Pooled specificity (95% CI) | Test1 |
| Type 1 Brands | ||||||
| Paracheck‐Pf | 27 | 22,319 | 6929 | 93.2 (89.7, 95.6) | 95.6 (92.8, 97.3) | P = 0.15 |
| ParaSight‐F | 17 | 12,521 | 3261 | 94.1 (89.9, 96.6) | 94.6 (90.4, 96.8) | |
| ICT Malaria‐Pf | 16 | 2955 | 1200 | 97.6 (95.5, 98.8) | 94.5 (90.5, 96.9) | |
| ParaHIT‐F | 4 | 1119 | 192 | 92.3 (74.9, 98.0) | 98.9 (94.9, 99.8) | |
| Determine Malaria‐Pf | 1 | 526 | 262 | 98.2 (85.4, 99.8) | 86.8 (35.1, 98.8) | |
| PATH | 2 | 378 | 180 | 96.6 (83.8, 99.3) | 93.3 (68.6, 98.9) | |
| Rapid Test Malaria | 1 | 306 | 36 | 97.8 (70.1, 100.0) | 96.1 (65.6, 99.7) | |
| DiaSpot Malaria | 1 | 153 | 63 | 71.8 (23.1, 95.6) | 82.6 (27.3, 98.4) | |
| Hexagon Malaria | 1 | 119 | 32 | 100.0 (.) | 65.7 (13.4, 96.0) | |
| New Pf‐1 mini | 1 | 10 | 6 | 100.0 (0, 100.0) | 100.0 (.) | |
| Combined2 | 65 | 40,062 | 11,966 | 94.8 (93.1, 96.1) | 95.2 (93.2, 96.7) | |
| Type 2 Brands | ||||||
| ICT Malaria Pf/Pv | 6 | 2255 | 600 | 96.0 (93.6, 97.5) | 95.6 (86.1, 98.7) | P = 1.0 |
| Now Malaria ICT | 2 | 1142 | 190 | 96.0 (91.6, 98.1) | 94.1 (66.6, 99.2) | |
| Combined | 8 | 3397 | 790 | 96.0 (94.0, 97.3) | 95.3 (87.3, 98.3) | |
| Type 3 (too few studies to stratify by brand) | ||||||
| Combined | 5 | 958 | 330 | 99.5 (71.0, 100.0) | 90.6 (80.5, 95.7) | |
| Type 4 Brands | ||||||
| OptiMAL | 10 | 3393 | 833 | 90.1 (86.3, 92.9) | 99.3 (98.0, 99.8) | P = 0.009 |
| Carestart Pf/Pan | 2 | 537 | 240 | 97.8 (94.1, 99.2) | 92.2 (72.4, 98.1) | |
| OptiMAL‐IT | 3 | 1356 | 280 | 87.4 (79.9, 92.4) | 97.0 (88.4, 99.3) | |
| Parabank | 2 | 7918 | 2992 | 87.9 (82.0, 92.0) | 98.8 (90.9, 99.9) | |
| Combined3 | 16 | 13,010 | 4274 | 91.5 (84.7, 95.3) | 98.7 (96.9, 99.5) | |
| Type 5 Brands (too few studies to stratify by brand) | ||||||
| Combined4 | 3 | 1777 | 400 | 98.4 (95.1, 99.5) | 97.5 (93.5, 99.1) | |
1 Likelihood ratio test for evidence of a difference between brands.
2 65 study cohorts evaluated 71 different tests. Only one test (selected randomly) from each cohort is included in the combined analysis.
3 16 study cohorts evaluated 17 different tests. Only one test (selected randomly) from each cohort is included in the overall analysis.
4 HSROC model fitted assuming no correlation between sensitivity and specificity.