Diagnostic technologies for power-distribution equipment are promoted to enhance grid reliability, but rigorous field evidence on their cost-effectiveness in low-resource settings is scarce. This study aimed to conduct a randomised field trial to evaluate the cost-effectiveness of implementing a diagnostic system for transformers and switchgear within a national utility. A randomised controlled trial was designed, allocating 200 primary substations to either a diagnostic intervention group or a standard practice control group. Cost data and failure events were recorded over an operational period. Cost-effectiveness was evaluated using a Bayesian hierarchical model: $\lambda_{ij} = \exp(\beta_0 + \beta_1 T_i + u_j)$, where $\lambda_{ij}$ is the failure rate for substation $i$ in region $j$, $T_i$ is the treatment indicator, and $u_j$ are regional random effects. Robust standard errors were calculated. The diagnostic system reduced unplanned outages by an estimated 18% (95% credible interval: 12% to 24%). However, the net present cost per major failure averted exceeded the local benchmark for capital projects, indicating poor cost-effectiveness under current tariff structures. While technically effective in improving reliability, the diagnostic intervention was not cost-effective for the utility under prevailing economic conditions. Deployment should be contingent on tariff reforms or targeted at critical network segments. Utilities should prioritise cost-benefit analyses before large-scale adoption of such technologies. power distribution, diagnostic systems, randomised trial, cost-benefit analysis, grid reliability, Bayesian modelling This paper provides the first experimental evidence from a randomised field trial on the economic efficacy of distribution equipment diagnostics in a sub-Saharan African context.