Management of pulmonary tuberculosis in Tajikistan: which factors determine hospitalization-.

OBJECTIVE: To assess predictors for tuberculosis hospitalization and treatment outcome in Tajikistan. METHODS: Stratified, single stage cluster sample survey of 1495 adult patients with pulmonary TB during 2 calendar years (2005-2006) from the registries of 10 TB centres chosen by simple random sampling. The primary outcome was referral to hospital. Logistic regression was conducted to test associations with the study outcome using linearization and a variance formula. RESULTS: Prevalence of hospitalization for tuberculosis was 58%. The odds of patients with smear-positive tuberculosis being referred were three times those of smear-negative patients [OR 2.99 (95% CI 1.81-4.96)]. Other predictors for hospitalization were the availability of TB hospital beds within the same district [OR 2.15 (95% CI 1.22-3.76)] and male gender [OR 1.46 (95% CI 1.07-2.48)]. The overall treatment success was 80%. CONCLUSIONS: Hospitalization of patients with pulmonary tuberculosis was determined by positive sputum smear, supply of hospital beds, and gender. Reducing hospitalization with support of national guidelines is not expected to have a negative impact on treatment outcome and spread of disease, but could lead to improved efficiency and effectives of health service delivery for pulmonary tuberculosis in Tajikistan.