Health care is delivered through an assemblage of systems and minisystems. If a devicerelated minisystem of an assemblage fails to deliver its clinical benefit or delivers a physical insult to a patient, an incident investigation will begin. Information will be gathered, devices will be tested, and personnel will be interviewed. Some conclusions will be reached, and corrective actions will be taken. However, when the investigative process is discussed with in-house investigators, the process is frequently found to be incomplete, flawed, or biased. These deficiencies appear to be related to the infrequency with which most in-hospital investigators perform investigations, a lack of education and training in investigative techniques, and a lack of understanding of analytical methods for determining the cause(s) of an event. A generic system’s risk model has been developed for analyzing the performance of a minisystem. It provides the investigator with a mental model of the interacting components of the mini-system and provides a logical pathway toward the root causes of an adverse event. In addition, at the time of a minisystem failure, the model can be used as a checklist to assure that information about each component has been gathered and analyzed for its contribution to the event. Although errors still will be made in causal determinations, they will not be due to a forgotten or neglected component or condition.