ORs, relative ranks and absolute proportion difference from fixed-effects and random-effects network meta-analyses across all isoniazid resistance profiles
| OR (95% CrI) | Rank (95% CrI) | Proportion difference (95% CrI) | ||||
|---|---|---|---|---|---|---|
| Treatment | Fixed effects | Random effects | Fixed effects | Random effects | Fixed effects | Random effects |
| ED<3 D<6m | 3.47 (1.08–11.55) | 4.38 (0.50–52.41) | 11 (8–11) | 11 (4–11) | 0.26 (0.01–0.52) | 0.29 (−0.09–0.74) |
| ED<3 D≥6m | 2.54 (1.19–5.59) | 2.50 (0.62–10.36) | 10 (8–11) | 10 (5–11) | 0.18 (0.03–0.34) | 0.16 (−0.07–0.43) |
| ED≥3 D≥6m | 0.91 (0.35–2.30) | 0.95 (0.15–6.21) | 6 (3–9) | 6 (2–10) | −0.01 (−0.14–0.14) | −0.01 (−0.20–0.31) |
| RIF ED<3 D<6m | 1.00 (0.52–1.95) | 1.32 (0.44–5.52) | 6 (4–9) | 7 (4–10) | 0.00 (−0.11–0.10) | 0.04 (−0.13–0.25) |
| RIF ED<3 D=6m | 0.48 (0.20–1.14) | 0.53 (0.13–2.20) | 3 (2–6) | 3 (1–7) | −0.09 (−0.20–0.02) | −0.07 (−0.23–0.09) |
| RIF ED<3 D>6m | 0.17 (0.02–1.07) | 0.19 (0.01–3.36) | 1 (1–6) | 1 (1–9) | −0.15 (−0.26–0.01) | −0.12 (−0.28–0.19) |
| RIF ED<3 Pr6 D=6m | Baseline | Baseline | 6 (3–9) | 6 (2–9) | Baseline | Baseline |
| RIF ED<3 Pr6 D>6m | 1.01 (0.39–2.62) | 0.85 (0.12–5.25) | 6 (3–10) | 5 (1–11) | 0.00 (−0.13–0.17) | −0.02 (−0.20–0.30) |
| RIF ED≥3 D<6m | 1.60 (0.59–4.35) | 1.89 (0.40–10.27) | 9 (4–10) | 9 (3–11) | 0.08 (−0.08–0.28) | 0.10 (−0.13–0.43) |
| RIF ED≥3 Pr6 D=6m | 0.98 (0.24–4.57) | 1.04 (0.13–9.25) | 6 (2–11) | 6 (1–11) | 0.00 (−–0.16–0.31) | 0.01 (−0.19–0.44) |
| RIF ED≥3 Pr6 D>6m | 0.31 (0.12–0.81) | 0.45 (0.07–4.32) | 2 (1–4) | 3 (1–9) | −0.12 (−0.23–0.02) | −0.08 (−0.24–0.23) |
Main results from the network depicted in figure 3B. Baseline proportion with a negative outcome 0.19 (95% CrI 0.12–0.29) in the fixed-effects model and 0.16 (95% CrI 0.07–0.31) in the random-effects model.
CrI, credible interval; ED, effective drugs; Pr6, protected at 6 months; RIF, rifampicin.