Table 3.
GRADE Evidence Profile: Predictors of in-hospital mortality among TB patients.
| Predictor/Time/N° of patients | N° of studies | Risk of bias | Inconsistancy | Indirectness | Imprecision | Quality | Relative effect (95% CI) |
|---|---|---|---|---|---|---|---|
| AFB smear positive/At baseline/1116 patients | 2 | No serious risk of bias | No serious inconsistancy | No serious indirectness | Serious imprecision | Low | OR 0.99 (0.40–2.48) |
| DM/At baseline/2165 patients | 4 | No serious risk of bias | Serious inconsistancy | No serious indirectness | Serious imprecision | Very low | OR 1.31 (0.38–4.46) |
| Hx of previous TB/At baseline/1675 patients | 2 | No serious risk of bias | Serious inconsistancy | No serious indirectness | Serious imprecision | Very low | OR 2.66 (0.48–14.87) |
| Male sex/At baseline/1880 patients | 7 | No serious risk of bias | No serious inconsistancy | No serious indirectness | Serious imprecision | Low | OR 1.09 (0.84–1.41) |
| Malignancy/At baseline/694 patients | 3 | No serious risk of bias | No serious inconsistancy | No serious indirectness | No serious imprecision | Moderate | OR 1.85 (1.01–3.40) |
*DM = Diabetes Mellitus; Hx = history.