Probabilities (per person per year) and relative risks used in the decision model
| Base value | Range for one-way sensitivity analysis | Distribution for probabilistic sensitivity analysis | Source | |
| Risk of developing active TB | ||||
| Baseline risk | 0.001 | 0.00065–0.005 | Individual risk estimates calculated for every case | Comstock et al. [16] 0.00065: TST converters, after 2 years [17] |
| Risk for TB contacts during the first 2 years | 0.05 for 2 years | 0.0325–0.25 for 2 years | Individual risk estimates calculated for every case | 0.05: risk for TB contacts with TST ≥10 mm during the first 2 years [18, 19] |
| TB case fatality rate (primary cause and contributory) | ||||
| 0–19 years 20–49 years 50–64 years 65–74 years ≥75 years | 0.01 (p1) 0.036 (p1×3.6) 0.048 (p1×4.8) 0.112 (p1×11.2) 0.303 (p1×30.3) | Sensitivity analysis range for p1=0.005–0.02 | Distribution of p1: β 0.01±0.001 Low (2.5th percentile): 0.008124227 High (97.5th percentile): 0.012056 | Walpola et al. [20] Range for sensitivity analysis: 50–200% of base estimate |
| Isoniazid-induced hepatitis requiring hospital admission | ||||
| ≤35 years 36–50 years 51–65 years >65 years | 0.000994915 (p2) 0.001776549 (p2×1.7856289) 0.006493506 (p2×6.526793) 0.023920653 (p2×24.04327) | Sensitivity analysis range for p2=0.0004974575–0.0019898300 | Distribution of p2: β 0.0009949155±0.00035 Low (2.5th percentile): 0.000430165 High (97.5th percentile): 0.001791393 | Smith et al. [21] Range for sensitivity analysis: 50–200% of base estimate |
| Hepatitis case fatality rate | ||||
| All ages | 0.0085 | 0.00425–0.017 | β-distribution 0.0085±0.003 Low (2.5th percentile): 0.003670959 High (97.5th percentile): 0.015292 | Smith et al. [21] for estimate of hepatitis cases Millard et al. [22] for estimate of hepatitis deaths in isoniazid recipients Range for sensitivity analysis: 50–200% of base estimate |
| Isoniazid-associated adverse events other than hepatitis requiring hospital admission | ||||
| ≤35 years 36–50 years 51–65 years >65 years | 0.001768738 (p3) 0.005527043 (p3×3.124851954) 0.010552 (p3×5.965808) 0.027421237 (p3×15.5032818) | Sensitivity analysis range for p3=0.000884369–0.003537475 | Distribution of p3: β 0.001768738±0.00055 Low (2.5th percentile): 0.000858887 High (97.5th percentile): 0.002994176 | Smith et al. [21] Range for sensitivity analysis: 50–200% of base estimate |
| Relative risk of TB with a 6-month course of LTBI treatment | ||||
| All ages | Relative risk reduction of 56%=0.44 (relative risk) | 0.27–0.73 | β-distribution 0.44±0.125 Low (2.5th percentile): 0.21394 High (97.5th percentile): 0.68223 | Smieja et al. [23] Range for sensitivity analysis: 95% confidence intervals in the original dataset |
| Relative risk of TB with LTBI treatment prematurely ceased due to hepatitis | ||||
| All ages | 0.84 (relative risk reduction of 16%) | 0.42–1.68 | β-distribution with 0.84±0.18 Low (2.5th percentile): 0.41628 High (97.5th percentile): 0.99906 | Rose et al. [6] Range for sensitivity analysis: 50–200% of base estimate |
Data are presented as mean±sd unless otherwise stated. TB: tuberculosis; LTBI: latent tuberculosis infection; TST: tuberculin skin test.