Evaluation of risk factors and development of a predictive clinical score for TB contact tracing in Angola

Abstract

Introduction

Contact tracing for tuberculosis (TB) is an essential component of public health efforts to detect secondary cases and interrupt transmission chains. This study aimed to assess the risk factors associated with TB diagnosis and to develop a predictive clinical score for TB diagnosis among contacts of people with TB attending the outpatient clinic at Hospital Nossa Senhora da Paz (HNSP) in Cubal, Angola.

Methods

Prospective observational study including TB contacts of all ages. Data were analyzed using descriptive and inferential statistics, including multivariable logistic regression with multiple imputation for missing data. The performance of the proposed clinical score was assessed through ROC analysis, calibration, bootstrap internal validation, and comparison with the WHO symptom-based screening approach.

Results

Two-hundred and eighty-five TB contacts were included. Fifty (17.5%) contacts were diagnosed with TB, 40 (80%) of whom were children ≤15 years old. Multivariate analysis identified cough, fever, undernutrition, and peripheral lymphadenopathy as independent risk factors for TB diagnosis. These variables were incorporated into a predictive clinical score ranging from 0 to 17 points. Using a cut-off of ≥6 points, the score showed a sensitivity of 94% and a specificity of 75%, with an area under the ROC curve of 0.89 (95% CI: 0.85–0.94).

Conclusions

These findings highlight the high burden of TB among close contacts and the challenges in properly assessing them. An easy-to-use clinical score may support frontline health workers in improving TB contact tracing in settings with high burden of disease and limited resources, such as Angola.