Public health surveillance systems are crucial for monitoring disease outbreaks in Kenya. However, their reliability and effectiveness vary widely. A randomized field trial was conducted in two districts across Kenya. System reliability was evaluated using a logistic regression model with robust standard errors. The system demonstrated an accuracy rate of 85% in detecting outbreaks, indicating moderate effectiveness overall. The findings suggest that improving training and resource allocation could enhance the reliability of public health surveillance systems in Kenya. Public health officials should prioritise investing in staff training and enhancing infrastructure to improve system performance. public health surveillance, Kenya, randomized field trial, reliability Treatment effect was estimated with $\text{logit}(p_i)=\beta_0+\beta^\top X_i$, and uncertainty reported using confidence-interval based inference.