Multilevel Analysis of Challenges in Diagnosis and Treatment of Pediatric Tuberculosis at Facility and Patient Levels in Lusaka Province, Zambia

Abstract

Tuberculosis (TB) remains a major public health challenge globally, with children disproportionately affected due to diagnostic and treatment complexities. In Zambia, pediatric TB continues to be underdiagnosed and undertreated, particularly in high-burden settings such as Lusaka Province. This study investigated the challenges affecting the diagnosis and treatment outcomes of pediatric tuberculosis at both patient and facility levels using a multilevel analytical approach. A cross-sectional quantitative study was conducted using secondary data extracted from tuberculosis registers in 20 health facilities across Lusaka Province. The study included 700 children aged 0–14 years who received TB treatment between January 2023 and December 2024. Descriptive statistics were used to summarize patient and facility characteristics, while inferential analyses, including chi-square tests, logistic regression, and ordinal logistic regression, were performed. Multilevel modeling was explored to account for the hierarchical structure of patients nested within facilities. Overall, 75.0% of children successfully completed treatment, while 15.3% defaulted and 9.7% died. Male children and younger age groups exhibited slightly poorer treatment outcomes. Facility-level clustering was minimal, with an intraclass correlation coefficient of 0.6%, indicating that most variation in treatment outcomes was attributable to patient-level factors rather than differences between facilities. A likelihood ratio test confirmed that single-level ordinal logistic regression was sufficient for analysis, although multilevel modeling was retained for robustness. All model assumptions were satisfied. The findings demonstrate that pediatric TB treatment outcomes in Lusaka Province are primarily driven by patient-level characteristics, including age, sex, and HIV status, while facility-level factors play a supportive but less decisive role. These results underscore the need for patient-centered, multi-level interventions that strengthen caregiver support, improve adherence, and complement existing health system investments. The study contributes methodologically by applying multilevel analysis and provides evidence to inform policy, program design, and future pediatric TB research in Zambia. Keywords: Pediatric Tuberculosis, Treatment Outcomes, Multilevel Analysis, Patient-Level Factors, Facility-Level Factors.