OBJECTIVE: To estimate the Tuberculosis treatment drop out prevalence and the variables associated in the patients registered in the Tuberculosis Control Program in Sapucaia do Sul (Brazil), between 2000 and 2008. METHOD: A cross-sectional study was conducted, which was based on the notified data in Information System for Disease Surveillance of the City Health Secretariat. RESULTS: From the 632 cases included in the study, 65 (10.3%; CI95%=7.9-12.7) were classified as treatment abandonment. Between 2000 and 2004, the prevalence of noncompliance was 12.7% (95%CI=9.1-16.2), and in the period 2005 to 2008 decreased to 7.0% (95%CI=4.0-9.9). In the crude analysis, we find association with sex, age and AIDS presence. The adjusted analysis with the Poisson regression didn't show significant differences between the independent variables. CONCLUSION: The analysis showed reduction in the prevalence of noncompliance with the creation of Tuberculosis Control Program, from 2005, although the confidence intervals are shown superimposed. Still, the prevalence of noncompliance was high and stove above the 5% target, agreed between levels of government.
tuberculosis; cross-sectional studies; medication adherence; treatment refusal; patients dropouts; therapeutics