Table 3.
Odds ratios of TB treatment initiation following negative diagnostic test result: alternative model specification using strict outcome definition and study random effects.
| Adjusted odds ratios for treatment initiation (95% Credible Intervals) |
|||
|---|---|---|---|
| Main analysis (primary outcome*, country random effects) | First alternative specification (alternative outcome**, country random effects) | Second alternative specification (primary outcome, study random effects)*** | |
| Age category | |||
| 18 – 30 years | Ref | Ref | Ref |
| 31 – 40 years | 1.17 (0.92–1.51) | 1.35 (0.90–2.02) | 1.13 (0.88–1.46) |
| 41 years and above | 1.11 (0.87–1.43) | 1.37 (0.94–2.06) | 1.07 (0.84–1.37) |
|
| |||
| Sex | |||
| Female | Ref | Ref | Ref |
| Male | 1.61 (1.31–1.95) | 1.57 (1.17–2.12) | 1.62 (1.34–1.95) |
|
| |||
| History of prior TB | |||
| None | Ref | Ref | Ref |
| Yes | 1.36 (1.06–1.73) | 1.45 (1.03–2.02) | 1.36 (1.08–1.72) |
| Unknown | 0.73 (0.28–1.65) | 0.57 (0.14–1.75) | 0.39 (0.06–1.65) |
|
| |||
| Reported cough | |||
| None | Ref | Ref | Ref |
| Yes | 4.62 (3.42–6.27) | 6.78 (3.78–12.65) | 4.73 (3.50–6.37) |
|
| |||
| Reported night sweats | |||
| None | Ref | Ref | Ref |
| Yes | 1.50 (1.21–1.90) | 1.29 (0.93–1.78) | 1.45 (1.17–1.82) |
|
| |||
| Reported fever | |||
| None | Ref | Ref | Ref |
| Yes | 1.13 (0.91–1.39) | 1.14 (0.84–1.56) | 1.08 (0.87–1.34) |
|
| |||
| HIV | |||
| Negative | Ref | Ref | Ref |
| Positive, not on ART | 1.68 (1.23–2.32) | 1.84 (1.16–2.95) | 1.55 (1.12–2.14) |
| Positive, on ART | 0.90 (0.64–1.30) | 0.92 (0.58–1.48) | 0.86 (0.59–1.24) |
| Unknown | 0.82 (0.55–1.20) | 0.38 (0.20–0.69) | 0.87 (0.59–1.29) |
|
| |||
| Diagnostic test | |||
| Sputum Smear | Ref | Ref | Ref |
| Xpert | 0.77 (0.62–0.96) | 0.65 (0.42–0.98) | 0.79 (0.62–0.99) |
| Xpert Ultra | 0.57 (0.30–1.07) | 0.35 (0.17–0.75) | 0.37 (0.21–0.64) |
|
| |||
| Year | 0.81 (0.74–0.90) | 0.97 (0.85–1.09) | 0.87 (0.74–1.04) |
*
Our primary outcome was the initiation of TB treatment after negative SSM, Xpert, or Xpert Ultra results; using treatment provision on clinical grounds if available, otherwise defining clinical diagnosis as treatment initiation post-negative initial tests prior to culture results.
**
For alternative outcome definition, clinical diagnosis was defined as treatment initiation within 7 days of the initial diagnostic test.
***
Multivariable regression model also included study random effects, coefficients shown in Table S4. Ref = reference category.