P1-261 Prediction of dropout tuberculosis treatment on prioritary cities to control in EspÍrito Santo State, Brazil

• Published in: Journal of epidemiology and community health (1979) Vol. 65; no. Suppl 1; p. A138
• Main Authors: do Prado, T N, Dietze, R, Netto, A R, Zandonade, E, Maciel, E L N
• Format: Journal Article
• Language: English
• Published: London BMJ Publishing Group Ltd 01.08.2011; BMJ Publishing Group LTD
• Subjects: Mycobacterium; Regression analysis
• ISSN: 0143-005X; 1470-2738
• DOI: 10.1136/jech.2011.142976e.53

ObjectivesTo identify the epidemiological factors related to the abandonment of TB treatment in the priority municipalities for TB control and to establish a score for use in TB control programs to identify patients most likely to abandon treatment.Methodscase-control study matched by sex and notification place to compare patients who dropped out of the treatment (cases) with those who were cured (control) in eight priority municipalities for TB control in the Espirito Santo state, from January 2006 to July 2008. Patients were interviewed directly by one of the researchers, at the clinic or at home. To data analysis we performed a bivariate analysis and the significant results obtained from these analysis were for the logistic regression analysis and network neural artificial (NNA). The questionnaire score was created and validated.ResultsThe study involved 21 cases and 41 controls. In the bivariate analysis, the epidemiological factors involved in the TB treatment dropout were identified as follows: average income, smoking, age, occupation, religion, drugs, previous treatment for TB and willingness to abandon. The logistic regression analysis and neural network revealed that the use of illicit drugs and patient without religion was strongly related to the abandonment. It was found that the neural classification was not more effective than logistic regression for the score marks preparation.ConclusionsThe score created was able to estimate the treatment dropout cases identified in the study and can be used in programs of TB control to identify patients most likely to abandon TB treatment.