{ "background": "The expansion of urban populations in sub-Saharan Africa has strained primary healthcare delivery. Primary care networks (PCNs) are promoted as a model to integrate services and improve efficiency, but robust longitudinal evidence on their implementation and impact is scarce.", "purpose and objectives": "This study aimed to evaluate the comparative effectiveness of two distinct PCN implementation methodologies—centralised coordination versus facilitated peer collaboration—on clinical outcomes for chronic disease management in an urban African setting.", "methodology": "A longitudinal, cluster-randomised field trial was conducted across 24 public primary health facilities. Facilities were randomised to either the centralised coordination or facilitated peer collaboration arm. Clinical outcomes for hypertensive patients were tracked. The primary outcome was the proportion of patients with controlled blood pressure (<140/90 mmHg) at 18 months. Analysis used a generalised linear mixed model: $\\text{logit}(P(Y{ij}=1)) = \\beta0 + \\beta1 \\text{Arm}j + \\gamma X{ij} + uj$, where $u_j \\sim N(0, \\sigma^2)$, with robust standard errors clustered at facility level.", "findings": "The facilitated peer collaboration arm demonstrated a statistically significant improvement in hypertension control compared to centralised coordination (adjusted odds ratio 1.82, 95% CI 1.24 to 2.68). Specifically, the proportion of patients achieving controlled blood pressure was 58.3% in the peer collaboration arm versus 42.1% in the centralised coordination arm at final follow-up.", "conclusion": "The methodology for implementing primary care networks significantly influences their clinical effectiveness. A facilitated peer collaboration approach yielded superior chronic disease management outcomes compared to a top-down, centrally coordinated model in this urban context.", "recommendations": "Policymakers should prioritise implementation strategies that foster direct peer learning and collaboration among primary care facilities when scaling network models. Future programme design should incorporate embedded, pragmatic trials to evaluate implementation processes.",