RESUMO
SETTING: Tibetan refugees in India, 1994-1996. OBJECTIVE: To determine tuberculosis (TB) incidence, independent risk factors for TB, and predictors of adverse outcomes. DESIGN: Data from a house-to-house census/demographic survey were merged with TB patient data. Separate multivariable models for each birthplace were developed for outcomes of interest. RESULTS: From 1994 to 1996, 47,491 Tibetans were surveyed and 1197 TB cases confirmed (incidence 835/ 100,000). Risk factors for TB in separate multivariable models differed by place of birth. Independent predictors of death for Tibet-born refugees included age >50 years, extra-pulmonary TB, and second-line therapy, while for India-born refugees they included second-line therapy and no improvement at the end of treatment. No significant risk factors for default were identified for Tibet-born refugees, while region of residence and the absence of a BCG scar were independent predictors among those born in India. Predictors of receipt of second-line therapy among Tibet-born refugees included region, years in camps, and prior TB, while among those born in India they were region, age > or =20 years, sputum-positive at diagnosis, and previous TB. CONCLUSIONS: TB incidence in Tibetan refugee settlements exceeds the highest national TB rates, and country of birth determines risk factors. TB control efforts in India should include this population.