Background
Despite robust Tuberculosis (TB) program with effective chemotherapy and high coverage, treatment interruption remains a serious problem. Interrupting TB treatment means that patients remain infectious for longer time and are at risk of developing drug resistance and death. This study was conducted to identify and describe predictors of TB treatment interruption.
Methods
A cohort of 291 notified TB patients from 20 selected health facilities in Vihiga County were enrolled in to the study and followed up until the end of treatment. Patient characteristics that potentially predict treatment interruption were recorded during treatment initiation using structured questionnaires. Patients who interrupted treatment were traced and reasons for stoppage of treatment recorded. Kaplan Meier method was used to estimate probabilities of treatment interruption by patient characteristics and determine time intervals. The Log rank test for the equality of survival distributions analyzed significance of survival differences among categorical variables. For multivariable analysis, Cox proportional hazard model, was fitted to identify predictors of TB treatment interruption through calculation of hazard ratios with 95% Confidence Intervals (CIs). For variable analysis, statistical significance was set at P ≤ 0.05. Reasons for treatment interruption were categorized according to most recurrent behavioral or experiential characteristics.
Results
Participants’ median age was 40 years (IQR = 32–53) and 72% were male. Of the 291 patients, 11% (n = 32) interrupted treatment. Incidences of treatment interruption significantly occurred during intensive phase of treatment. Independent predictors of treatment interruption included alcohol consumption (HR = 9.2, 95% CI; 2.6–32.5, p < 0.001), being female (HR = 5.01, 95% CI; 1.68–15.0, p = 0.004), having primary or lower education level (HR = 3.09, 95% CI; 1.13–8.49, p < 0.029) and having a treatment supporter (HR = 0.33, 95% CI; 0.14–0.76, p = 0.009). Reasons for interrupting treatment were categorized as: alcoholism, feeling better after treatment initiation, associated TB stigma, long distance to health facility, lack of food, perception of not having TB and pill burden.
Conclusion
TB treatment interruption was high and largely associated with patients’ socio-demographic and behavioral characteristics. These multidimensional factors suggest the need for interventions that not only target individual patients but also environment in which they live and receive healthcare services.