Cost-Effectiveness Analysis of Water Treatment Systems in Ethiopia: A Multilevel Regression Modelling Case Study

Evaluating the cost-effectiveness of water treatment infrastructure in low-resource settings is critical for sustainable development. Existing analyses often fail to account for the hierarchical nature of cost data, which is clustered by geographical region and facility type, leading to potentially biased estimates. This case study aims to demonstrate a robust methodological framework for assessing the cost-effectiveness of diverse water treatment systems. Its objective is to quantify the influence of facility-level and regional-level factors on unit treatment costs. A multilevel regression model was applied to operational cost data from a sample of treatment facilities. The core statistical model is specified as $\text{Cost}_{ij} = \beta_{0} + \beta_{1}X_{ij} + u_{j} + e_{ij}$, where $i$ denotes facilities and $j$ denotes regions, with $u_{j}$ as the region-level random effect. Inference was based on robust standard errors. The analysis revealed that membrane-based systems were, on average, 34% more cost-effective per cubic metre treated than conventional coagulation plants, after controlling for capacity and energy source. The regional random effects were significant, indicating that unobserved geographical factors account for a substantial portion of cost variation. Multilevel modelling provides a superior analytical approach for infrastructure cost-effectiveness studies where data has a nested structure, yielding more reliable and actionable insights for engineers and policymakers. Future project appraisals should adopt hierarchical modelling techniques. Investment should be prioritised towards membrane technologies in settings with reliable power, while also funding region-specific cost-driver analyses. cost-effectiveness analysis, multilevel modelling, water treatment infrastructure, random effects, sustainable engineering This study provides a novel application of multilevel regression to civil engineering economic analysis, demonstrating that accounting for data hierarchy significantly alters cost-effectiveness rankings for water treatment technologies.