Table 4.
Subgroup analyses of reported time delay for TB diagnosis
| Subgroup | # of studies | Median reduction (95% CI) | I2 | p-value |
|---|---|---|---|---|
| Year | ||||
| Pre-2015 | 8 | 1.21 (− 0.67 to 3.10) | 98.99% | 0.20 |
| Post-2015 | 4 | 3.15 (− 2.72 to 9.01) | 98.56% | 0.29 |
| Empiric treatment rate | ||||
| High ≥ 50% | 5 | 1.58 (0.55–2.61) | 89.18% | 0.003 |
| Low < 50% | 7 | 1.85 (− 1.91 to 5.6) | 99.31% | 0.34 |
| HIV prevalence | ||||
| High ≥ 50% | 5 | 1.12 (− 1.19 to 3.43) | 96.31% | 0.34 |
| Low < 50% | 7 | 2.31 (− 1.09 to 5.71) | 99.62% | 0.18 |
| Healthcare level | ||||
| Primary | 4 | 0.83 (− 1.83 to 3.49) | 96.25% | 0.54 |
| Tertiary | 7 | 2.54 (− 0.84 to 5.92) | 98.87% | 0.14 |
| Study design | ||||
| RCT | 3 | 4.56 (− 2.72 to 11.43) | 99.19% | 0.19 |
| Observational | 6 | 1.17 (− 1.18 to 4.1) | 97.85% | 0.44 |
| Overall | 12 | 1.78 (− 0.27 to 3.85) | 99.25% | 0.089 |
Pre-2015 refer to studies with data from before 2015 when Xpert capacity was limited. For empiric treatment rate and HIV prevalence, 50% or greater was considered to be high
RCT Randomized Controlled Trial