{ "background": "Public health surveillance systems in low-resource settings often face low adoption rates, undermining their effectiveness. In Ethiopia, despite investments in infrastructure, the optimal methodological approach for deploying such systems to maximise uptake by health workers remains unclear.", "purpose and objectives": "This study aimed to determine which of two surveillance system deployment methodologies—centralised cascade training versus decentralised peer-supported implementation—leads to higher rates of adoption within the country's public health infrastructure.", "methodology": "We conducted a parallel, cluster-randomised field trial across 40 rural districts (woredas). Districts were randomly allocated to receive either the centralised training intervention (n=20) or the decentralised peer-support intervention (n=20). Adoption was measured as the proportion of eligible health workers routinely submitting complete surveillance reports. The primary analysis used a generalised linear mixed model: $\\logit(p{ij}) = \\beta0 + \\beta1 T{ij} + ui + e{ij}$, where $p{ij}$ is the probability of adoption for health worker $j$ in district $i$, $T{ij}$ is the treatment indicator, and $u_i$ is a district-level random effect.", "findings": "The decentralised peer-support methodology yielded a significantly higher adoption rate (68%, 95% CI: 64 to 72) compared to the centralised training approach (52%, 95% CI: 48 to 56). The adjusted odds ratio for adoption in the decentralised arm was 2.15 (95% CI: 1.74 to 2.66, p<0.001).", "conclusion": "A decentralised, peer-supported implementation strategy substantially improves the adoption of public health surveillance systems by frontline health workers in a resource-constrained setting.", "recommendations": "National public health programmes should integrate structured peer-support mechanisms into surveillance system roll-outs. Further research should investigate the cost-effectiveness and long-term sustainability of this approach.",