A Bayesian Hierarchical Modelling Approach to Evaluating Urban Primary Care Networks: Clinical Outcomes in the Ethiopian Health System

"background": "Urban primary care networks are a critical component of health system strengthening in sub-Saharan Africa, yet robust methodologies for evaluating their impact on clinical outcomes are underdeveloped. Existing evaluations often rely on aggregated data, failing to account for heterogeneity across facilities and patient populations. ", "purpose and objectives": "This study aimed to develop and apply a Bayesian hierarchical modelling framework to quantify the effect of urban primary care network integration on key clinical outcomes within a national health system, using antenatal care completion as a primary indicator. ", "methodology": "We utilised a longitudinal, facility-level dataset from urban health centres. The core model was specified as y{it \ (, nit), () = \ + \ Xit + ui + vt +, where ui and vt are structured random effects for facility and time, and X₈ₓ denotes network integration status. Posterior distributions were estimated using Hamiltonian Monte Carlo. ", "findings": "Network integration was associated with a 14. 2 percentage point increase (95% credible interval: 8. 7 to 19. 1) in the probability of complete antenatal care attendance. The facility-level random effects showed substantial variation, indicating that contextual factors modify the network's effectiveness. ", "conclusion": "The Bayesian hierarchical model provides a robust, nuanced tool for health systems research, revealing that while primary care network integration significantly improves clinical service completion, the magnitude of benefit is context-dependent. ", "recommendations": "Health policymakers should prioritise network integration in urban health strategies but couple it with targeted support for lower-performing facilities. Future evaluations should adopt similar multi-level frameworks to guide resource allocation. ", "key words": "Health systems research, primary health care, Bayesian statistics, hierarchical model, clinical outcomes, sub-Saharan Africa", "contribution statement": "This paper provides a novel methodological framework for health