OBJECTIVE - To identify factors that predict noncompliance with tuberculosis treatment in the province of Ica, Peru.
METHODS - Between 1998 and 2000 a case-control study (1:1 ratio) was conducted in the province of Ica, with 55 cases (persons who dropped out of treatment) being identified. The factors evaluated were chosen from Lalonde's model of the field of health. The respective odds ratios were calculated by means of univariate analysis and multivariate analysis.
RESULTS - The following factors were identified as being predictive of noncompliance with tuberculosis treatment: thinking that the information that health workers provided on treatment was insufficient (odds ratio (OR) = 4.20; 95% confidence interval (95% CI), 1.77-10.02), thinking that the office hours for receiving treatment were inadequate (OR = 9.95; 95% CI, 1.97-50.21), and consuming illegal drugs (OR = 7.15; 95% CI, 1.69-30.23).
CONCLUSIONS - To improve compliance with tuberculosis treatment, it is necessary to provide patients with personalized information on the disease and its treatment, and to offer them flexible, appropriate times to receive treatment. Drug consumption is the most serious risk factor for noncompliance, and identifying and following up with drug users is crucial.