Predictors of unsuccessful tuberculosis treatment outcomes in children from a prospective cohort study in Pakistan

BACKGROUND: Every year, about 239 000 children die from tuberculosis (TB), despite availability of highly effective regimens. Few studies have evaluated predictors for poor treatment outcomes in children treated for TB. METHODS: We assessed predictors of unsuccessful TB treatment outcomes in a prospective cohort of children diagnosed by an intensified TB patient-finding intervention at four facilities in Pakistan between 2014 and 2016. A case of TB disease was determined through either bacteriologic confirmation of disease or a clinical diagnosis. To estimate characteristics predictive of experiencing an unsuccessful treatment outcome, we used a multi-level model with a modified Poisson approach, accounting for clustering at the facility level. We report estimated relative risks (RR) and 95% confidence intervals (CI). RESULTS: = 0.03). CONCLUSIONS: More than 94% of children experienced successful treatment outcomes. We identified individual-, facility-, and clinical-factors predictive of experiencing unsuccessful treatment outcomes. Children with fevers and abdominal findings suggestive of TB disease should be tested for TB and followed closely throughout treatment to ensure necessary support for successful completion of treatment.