Table 2. Meta-regression: Association between pooled prevalence of new MDR TB cases and study characteristics.
| CHARACTERISTICS | ¶ Univariate coefficient 95% CI: () | P value | Adjusted coefficient | P value |
|---|---|---|---|---|
| Year of study | -0.0018 (-.0037 .00015) | 0.070 | ||
| Study design | ||||
| Cross sectional | Ref | - | ||
| Prospective | -0.0056 (-0.026 0.016) | 0.598 | - | |
| Retrospective | 0.016 (-0.0149 .0482) | 0.296 | ||
| Region, | - | |||
| Central Africa | Ref | - | ||
| East Africa | -0.0023 (-0.026–0.022) | 0.850 | - | |
| Southern Africa | 0.016 (-0.012–0.043) | 0.259 | - | |
| West Africa | .0019786 (-0.025–0.029) | 0.882 | - | |
| Population | Collinearity | - | - | |
| Age, | Collinearity | - | - | |
| Method of screening | - | |||
| BACTEC 460 | Ref | - | ||
| BACTEC 960 | .00233(-0.020–0.024) | 0.836 | - | |
| LJ | .00906 (-0.050 0.068) | 0.762 | - | |
| HIV prevalence in study | -0.0001 (-0.0006 .0005) | 0.696 | - |
¶The regression coefficients are the estimated increase in the MDR TB prevalence per unit increase in the study characteristics