Public health surveillance systems are critical for disease control, yet their methodological reliability in low-resource settings is often inadequately quantified. In Ghana, despite long-standing system implementation, a rigorous, longitudinal evaluation of surveillance data quality and consistency is lacking. This study aimed to methodologically evaluate the reliability of the national public health surveillance system and to identify the key facility- and district-level determinants of reporting consistency over an extended period. We conducted a longitudinal analysis of surveillance performance metrics from district health records. System reliability was operationalised as the consistency of key indicator reporting. A three-level hierarchical linear model was fitted to assess determinants of reliability: $Y_{ijk} = \beta_0 + \beta_1X_{1ijk} + u_{jk} + v_k + \epsilon_{ijk}$, where $u_{jk}$ and $v_k$ are random intercepts for facility $j$ in district $k$. Inference was based on robust standard errors. District-level logistical capacity was the strongest predictor of reliability (β = 0.42, 95% CI: 0.31, 0.53). Facilities with dedicated surveillance officers showed a 24% higher mean reliability score compared to those without. Significant variability was attributed to the district level, indicating systemic influences. The reliability of surveillance data is predominantly influenced by higher-level systemic and resource factors, rather than facility-specific characteristics alone. Investment should prioritise strengthening district-level logistical and human resource capacity. A revised supervisory framework with targeted support for low-performing districts is required to homogenise system performance. health surveillance, system reliability, multilevel modelling, health systems research, data quality This study provides a novel methodological framework for longitudinally quantifying surveillance reliability and is the first to apply multilevel regression to isolate district- and facility-level determinants in this context.