Table 4 Comparison of multivariate logistic regression models using complete case and multiple imputation analysis
Risk factors | Successn (%) | Non-success n (%) | Complete case analysisa | Multiple imputation analysis | ||
|---|---|---|---|---|---|---|
aOR | 95% CI | aOR | 95% CI | |||
Gender: male | 1231 (75.0%) | 471 (78.8%) | 1.55 | 0.99–2.42 | 1.10 | 0.84–1.44 |
Age ≤ 60 years |  |  | 0.98 | 0.90–1.07 | 1.01 | 0.96–1.06 |
Age > 60 years |  |  | 1.79 | 1.26–2.54 | 1.45 | 1.14–1.79 |
|  | For every increase of 5 years of age | |||||
Diabetes | 285 (34.1%) | 71 (21.3%) | 0.84 | 0.55–1.31 | 0.81 | 0.61–1.08 |
HIV positive | 102 (6.5%) | 99 (17.9%) | 3.12 | 1.66–5.84 | 2.94 | 2.07–4.16 |
History of previous MDR-TB treatment | 18 (1.2%) | 38 (7.0%) | 20.37 | 5.52–75.17 | 5.53 | 2.85–10.72 |
Low smear gradeb | 788 (48.9%) | 259 (44.1%) | 1.72 | 1.00–2.95 | 1.47 | 1.08–2.00 |
High smear gradec | 416 (25.8%) | 209 (35.6%) | 2.25 | 1.28–3.93 | 2.06 | 1.49–2.87 |
AFB positive unknown | 40 (2.5%) | 22 (3.7%) | 2.70 | 1.08–6.78 | 2.80 | 1.47–5.36 |
BMI < 21 |  |  | 0.82 | 0.76–0.89 | 0.83 | 0.79–0.87 |
BMI ≥ 21 |  |  | 0.96 | 0.81–1.15 | 1.06 | 0.93–1.2 |
| Â | For every 1 increase of BMI | |||||