Volume 8, Number 11—November 2002
Research
Tuberculosis-Related Deaths within a Well-Functioning DOTS Control Program
Table 2
Population attributable-risk percent and hazard ratios for death among smear-positive tuberculosis (TB) patients, Orizaba, Veracruz, 1995–2000a
| Variables | Population attributable risk (%) | Adjusted hazard ratio | 95% CI | p valueb |
|---|---|---|---|---|
| Death due to TBc | ||||
| Treatment default | 28.7 | 8.9 | 3.3 to 24.4 | <0.0001 |
| Resistance to isoniazid and rifampin | 25.9 | 5.7 | 2.0 to 16.3 | <0.001 |
| Clustered | 18.8 | 4.1 | 1.6 to 10.0 | 0.002 |
| Weight loss (>15%) | — | 3.9 | 1.5 to 10.9 | 0.007 |
| Formal education <6 yr | — | 1.8 | 0.6 to 5.2 | 0.3 |
| Death due to other causes | ||||
| HIV/AIDS | 11.1 | 33.1 | 11.4 to 95.4 | <0.0001 |
| Hepatic cirrhosis | 6.6 | 5.7 | 1.6 to 19.7 | 0.006 |
| Weight loss (>15% ) | — | 3.3 | 1.6 to 6.7 | 0.001 |
| Age (yrs) | — | 1.02 | 0.99 to 1.04 | 0.07 |
aCI, confidence interval; —, not applicable.
bCox proportional hazards model.
cControlling for death before or after treatment completion, default or failure.
