Risk management in medicine is often seen as lagging behind other safety-critical domains where there has been considerable research into incident causation models. In this article, incident analysis theory and methodology from fields other than medicine are applied to an incident reporting scheme in an Edinburgh Intensive Care Unit. The incident analysis model used emphasizes the importance of latent organizational factors and complex, multilayered incident causation. It also takes the role of cognitive performance-shaping factors into account. This provides an analytical framework that integrates the identification of distal causal factors and a mechanism for comparing alternative causal hypotheses.