Table 2 Parameter values used and estimated in the model. Numbers in brackets show the uncertainty ranges used in the simulations (for input parameters) or Bayesian credible intervals (for parameters being estimated)
From: Counting the lives saved by DOTS in India: a model-based approach
Parameter name | Symbol | Value | Note/source |
|---|---|---|---|
Average number of infections per drug-susceptible (DS) TB case per year | β | 10.7 [5.8–13.6] | Estimated |
Average number of infections per MDR-TB case per year | β mdr | 2.00 [1.62–2.62] | Estimated |
Per care seeking attempt, probability of seeking care in the public sector (following RNTCP scale-up) | p max | 0.34 [0.25–0.58] | Estimated to get reported notifications from 1997–2015 |
Proportion of MDR-TB cases whose drug resistance is recognized at the point of TB diagnosis and who start appropriate treatment | q max | 0.07 [0.06–0.09] | Estimated to get reported notifications from 2007–2015 |
Reduction in force of infection owing to previous infection | C | 0.5 | Assumed |
Proportion of infections undergoing ‘rapid’ progression | k | 0.15 | Vynnycky and Fine, 1997 [42] |
Rate of breakdown from remote infection to active disease | h | 0.001 y-1 | Horsburgh et al., 2010 [43] |
Rate corresponding to the delay from the start of symptoms to the initiation of treatment (whether in public or private sector) | r | 3.29 y-1 [0.83–5.70] | Estimated |
Mean duration of first-line treatment | \( {\uptau}_{\mathrm{FL}} \) | 2 y-1 | Corresponding to 6Â months of treatment duration |
Rate of default from non-RNTCP treatment | d RNTCP | 1.06 y-1 | Uplekar et al. 1998 [29] |
Rate of default from RNTCP treatment | d non-RNTCP | 0.049 y-1 | Corresponds to 4.8% default in RNTCP (TB India, 2015 [34]) (averaged over smear-positive, smear-negative and extrapulmonary TB) |
Rate of repeat care seeking after recurrence or failure | w | 4 y-1 | Corresponds to 3Â months of delay period |
Annual recurrence rate | \( \uprho \) | 0.003 y-1 | Corresponds to lifetime recurrence risk of 17% (Sun et al., 2013 [44]) |
Rate of primary MDR acquisition from patient treated under RNTCP | m | 0.02 y-1 | TB India, 2015 [34] |
Mean duration of second-line treatment | \( {\uptau}_{\mathrm{SL}} \) | 0.5 y-1 | Corresponding to 2Â years of treatment duration |
Spontaneous cure rate | \( \upsigma \) | 0.166 y-1 | Corresponds to 50% spontaneous cure in 3Â years alongside with TB mortality (Tiemersma et al., 2011 [26]) |
Proportion cure of drug-susceptible (DS)-TB in RNTCP after first-line treatment | α pub | 0.87 | TB India, 2015 [34] |
Proportion cure of DS-TB in non-RNTCP after first-line treatment | α prv | 0.51 | Uplekar et al., 1998 [29] |
Proportion cure of MDR-TB in RNTCP after first-line treatment (excluding self-cure) | α’ pub | 0.24 | TB India, 2015 [34] |
Proportion cure of MDR-TB in non-RNTCP after first-line treatment | α’ prv | 0 | Assumed |
Proportion cure of MDR-TB with second-line treatment (excluding self-cure) | α’ pub2 | 0.48 | TB India, 2014 [45] |
Per-capita mortality hazard before diagnosis | μ UTB | 0.086 (95% CI 0.075–0.11) y-1 | See Additional file 1 |
Mortality hazard during RNTCP treatment | μ RNTCP | 0.076 (95% CI 0.069–0.095) y-1 | See Additional file 1 |
Mortality hazard during non-RNTCP treatment | μ non-RNTCP | 0.27 (95% CI 0.22–0.33) y-1 | See Additional file 1 |
Mortality hazard following default and treatment failure | μ B | 0.28 (95% CI 0.22–0.36) y-1 | See Additional file 1 |