Objectives: The WHO recommends that children and adolescents living with HIV (CALHIV) complete TB symptom screening at every clinical encounter but evidence supporting this recommendation is limited. We evaluated the performance of the recommended TB symptom screening in six high-burden TB/HIV countries.
Design: Retrospective longitudinal cohort.
Methods: We extracted data from electronic medical records of CALHIV receiving care from clinics in Botswana, Eswatini, Lesotho, Malawi, Tanzania, and Uganda from January 2014 to June 2017. We defined incident TB cases as those prescribed TB treatment within 30 days of TB diagnosis. We analyzed the most recent symptom screen preceding a TB diagnosis. In accordance with WHO guidelines, positive screens were defined as current fever, cough, poor weight gain, or recent TB contact. Odds of TB disease was modeled by screen result and age at which screening was conducted.
Results: Twenty thousand seven hundred and six patients collectively had 316 740 clinic visits, of which 240 161 (75.8%) had documented TB symptom screens. There were 35 701 (14.9%) positive TB symptom screens, and 1212 incident TB diagnoses. Sensitivity and specificity of the TB symptom screen to diagnose TB were 61.2% (95% CI 58.4--64.0) and 88.8% (95% CI 88.7--88.9), respectively. Log odds of documented TB for positive or negative screens was statistically different only for screens conducted at ages 7--17.
Conclusion: Although specificity was high, the sensitivity of the TB symptom screen to detect TB in CALHIV was low. More accurate screening approaches are needed to optimally identify TB disease in CALHIV.