| Parameter | Base case value | Range used in sensitivity analysis | Source |
| Starting age years | 36 | 18–60 | Estimated based on cost surveys |
| TB-related mortality rates % | |||
| Drug-susceptible or mono/polyresistant TB | |||
| Initial Std/EMB initial | 4 | 2–8 | WHO [27] |
| Retreat Std/Str retreat/MDR Std | 7 | 4–11 | WHO [27] |
| RMP-resistant/MDR-TB¶ | |||
| Non-MDR-regimen | 33 | 30–40 | Assumed same as untreated TB [28] |
| MDR Std | 11 | 9–13 | Johnston et al. [29] |
| Other mortality rates | |||
| Untreated TB % | 33 | 30–40 | Rieder [28] |
| Non-TB-related mortality % | Age-specific | Age-specific | Used WHO life table for Ecuador [30] |
| Mortality rate if blind | 3.5× age-specific mortality rate | 3.5× age-specific mortality rate | Remme [31] |
| Probability of EMB-induced blindness % | |||
| Initial Std | 0.0 | 0.0 | EMB in first 2 months assumed no blindness |
| EMB initial/retreat Std/Str retreat (EMB for 6–8 months) | 0.23 | 0–0.61 | Ezer et al. [14] |
| MDR Std (EMB for 24 months) | 0.46 | 0–0.99 | Assumed double the risk from retreatment due to longer treatment duration |
| Treatment default rate+ % | |||
| Drug-susceptible or mono/polyresistant TB | |||
| All initial regimens (6 months) | 8 | 4–15 | WHO [27] |
| All 8-month retreatment regimens | 15 | 4–30 | WHO [27] |
| MDR treatment | 14.7 | 11.2–19.0 | Assumed same as MDR-TB [29] |
| RMP-resistant/MDR-TB¶ | |||
| Non-MDR-regimen | 0.0 | 0.0 | Assumed that patients would only die, fail or be spontaneously cured |
| MDR Std | 14.7 | 11.2–19.0 | Johnston et al. [29] |
| Treatment failure rate % | |||
| Drug-susceptible TB | |||
| Any regimen | 0.3 | 0.1–0.4 | Menzies et al. [32] |
| Mono/polyresistant TB | |||
| Initial Std/EMB initial | 2.8 | 0.7–14.7 | Menzies et al. [32] |
| Retreat Std | 7.4 | 0–15.1 | Menzies et al. [3] |
| Str retreat | 3.7 | 0–15.1 | Assumed half the rate of first-line retreatment [3] |
| MDR Std | 0.4 | 0–1.5 | Estimated from Menzies et al. [3] |
| RMP-resistant/MDR-TB¶ | |||
| Non-MDR-regimen | 42 | 35–45 | Assumed that patients could only die or be spontaneously cured, and the remainder failed [28, 33] |
| MDR Std | 8 | 5–11 | Johnston et al. [29] |
| Probability of relapse % | |||
| Drug-susceptible TB | |||
| Any regimen | 3.7 | 2.8–4.7 | Menzies et al. [32] |
| Mono/polyresistant TB | |||
| Initial Std/EMB initial | 11.4 | 6.5–16.2 | Menzies et al. [32] |
| Retreat Std | 4.6 | 4.0–14.5 | Menzies et al. [3] |
| Str retreat | 2.3 | 2.0–14.5 | Assumed half the rate of first-line retreatment [3] |
| MDR Std | 2.6 | 0–7.9 | Estimated from Menzies et al. [3] |
| RMP-resistant/MDR-TB¶ | |||
| Non-MDR-regimen | 2.5 | 1.3–2.5 | Reported relapse rate among untreated TB cases who had spontaneous cures [33] |
| MDR Std | 2.0 | 4.0–14.5 | Johnston et al. [29] |
| Probability of acquired INH monoresistant TB among treatment failures % | |||
| Drug-susceptible TB | |||
| Initial Std | 44.2 | Calculated using data from Menzies et al. [32] | |
| EMB initial/retreat Std | 22.1 | Assumed half the rate of the initial standardised regimen [32] | |
| Str retreat/MDR Std | 0.0 | 0.0 | No INH in regimen |
| Probability of acquired INH monoresistant TB among those who default/relapse % | |||
| Drug-susceptible TB | |||
| Initial Std | 3.0 | Calculated using data from Menzies et al. [32] | |
| EMB initial/retreat Std | 1.5 | Assumed half the rate of initial standardised regimen [32] | |
| Str retreat/MDR Std | 0.0 | 0.0 | No INH in regimen |
| Probability of acquired RMP-resistant/MDR-TB among treatment failures % | |||
| Drug-susceptible TB | |||
| Initial Std | 17.3 | Calculated using data from Menzies et al. [32] | |
| EMB initial/retreat Std | 8.7 | Assumed half the rate of initial standardised regimen [32] | |
| Str retreat/MDR Std | 0.0 | 0.0 | No INH in regimen |
| Mono/polyresistant TB | |||
| Initial Std | 85.7 | Calculated from data in Menzies et al. [3] | |
| EMB initial/retreat Std | 42.9 | Assumed half the rate of initial standardised regimen [3] | |
| Str retreat | 21.5 | Assumed half the rate of first-line retreatment [3] | |
| MDR Std | 0.0 | 0.0 | No INH or RMP in regimen |
| Probability of acquired RMP-resistant/MDR-TB among those who default/relapse % | |||
| Drug-susceptible TB | |||
| Initial Std | 2.3 | Calculated from data in Menzies et al. [32] | |
| EMB initial/retreat Std | 1.15 | Assumed half the rate of initial regimen [32] | |
| Str retreat/MDR Std | 0.0 | 0.0 | No INH or RMP in regimen |
| Mono/polyresistant TB | |||
| Initial Std | 8.8 | Calculated from data in Menzies et al. [3] | |
| EMB initial/retreat Std | 4.4 | Assumed half the rate of initial regimen [3] | |
| Str retreat | 2.2 | Assumed half the rate of first-line retreatment [3] | |
| MDR Std | 0.0 | 0.0 | No INH or RMP in regimen |
| Cure rate after treatment default % | |||
| Drug-susceptible or mono/polyresistant TB | 62 | 25–70 | Chee et al. [34] |
TB: tuberculosis; Std: standardised; EMB: ethambutol; Str retreat: strengthened retreatment; MDR: multidrug-resistant; RMP: rifampicin; WHO: World Health Organization; INH: isoniazid. #: all probabilities are based on end-of-treatment outcomes unless otherwise noted; ¶: we assumed that patients with MDR-TB not treated with the standardised MDR regimen experience only three treatment outcomes, fail, die or spontaneously cured; 0% of MDR-TB patients default; +: treatment default rates include patients who default treatment and those who transfer out. We assumed that treatment outcomes of transfers-out were equivalent to those of default [35].