To design and run an effective pharmacovigilance system, one must understand potential hazards that may cause the system to fail. System-Theoretic Process Analysis (STPA) is a hazard analysis technique that has been successfully applied in domains including aviation, nuclear power, and defence systems; however, it has not yet been applied to a critical pharmacovigilance process. The objective of this project was to assess the value of STPA in pharmacovigilance by using the example of signal management, in order to identify process risks, areas for improvement, and applicable process metrics. Hierarchical control structure (HCS) is the starting point for STPA. The HCS models a process as a set of interacting feedback control systems, which maintain the system in a “safe state” where losses do not occur. “Losses” can be defined as harm to patients, regulatory non-compliance, or other negative consequences. The HCS viewpoint is a considerable departure from the usual consideration of a pharmacovigilance process as a series of sequential steps carried out by responsible individuals. Following the definition of the HCS, a two-step STPA process was carried out, leading to a set of unsafe actions and scenarios by which process failures could lead to loss. Recommendations were then made to prevent those scenarios that were not already covered by existing features of the process. The signal management process was successfully modelled as a feedback control system, which dynamically ensures that an incoming stream of safety data is accurately reflected in the reference safety information for the product. After creating the HCS, we conducted the STPA itself. This yielded 215 scenarios through which the process could end up in an unsafe state where losses could occur. The scenarios included 91 that were covered by existing measures, 28 related to existing or planned metrics, and 25 were scenarios considered implausible, inconsequential, undetectable, or out-of-scope. The remaining 71 scenarios were consolidated into 8 proposed recommendations for enhancements to metrics, 10 proposed recommendations for enhancements to the process itself, and 8 proposed recommendations regarding infrastructure or related processes, including the associated quality management system. STPA considers “soft” factors such as cultural influences, and this was reflected in several recommendations, for instance regarding training. STPA is a labour-intensive and time-consuming process, most appropriate for high-risk, compliance-critical processes. STPA is a powerful technique for identification of risks in pharmacovigilance systems, and can contribute to process reliability with the potential to improve patient safety and the maintenance of regulatory compliance.