Abstract Aggression is a common problem that confronts nurses within inpatient psychiatric units, yet there is limited evidence to definitively guide identification of consistent precursors to aggression and interventions to prevent aggression. This dissertation included two foci: an integrative literature review on the precursors to aggression, and a qualitative study examining the experiences of expert psychiatric nurses preventing acutely ill patients from transitioning to aggression. The integrative literature review of 17 studies identified numerous precursors to aggression. These were organized into three categories adapted from the socio ecological model: individual, interpersonal, and organizational (Dahlberg & Krug, 2002). The review of available studies reflected the complexity of aggression research in the patient population; as well as the delicate nature of studying interactions of patients and nurses in a locked psychiatric unit. Precursors identified were not outlined or linked to specific phases of aggression. The qualitative study using phenomenological analysis explored the experiences of 10 inpatient psychiatric nurses in preventing patient aggression during a situation of escalating behavior. Experienced psychiatric nurses described interventions used to prevent patient aggression. Three themes emerged: 1) connected knowing, piecing together the puzzle of the patient’s situation; 2) moral commitment and action; and 3) managing uncertainty. Within the three themes, subthemes also emerged. Connected knowing included subthemes: 1) knowing from past experience with the specific patient; 2) knowing from past experiences with other patients with similar diagnoses and behaviors; and 3) knowing through presence with patients. Subthemes embedded within moral commitment and action were: 1) empathic understanding; and 2) advocacy. Managing uncertainty included: 1) priority of creating and maintaining a safe environment; and 2) using time in clinical reasoning. Applying interventions in an uncertain situation requires patience, advocacy, and empathic understanding. The results suggest that nurses need to be allowed time to work with their patients to allow interventions to unfold and continue the connection with the patient. Future research is needed to test whether educational interventions that teach how to establish and maintain connections with patients during behavioral escalation and managing the uncertainty of the situation are effective in supporting nurses and other health care staff in intervening during the “middle phase” of aggression. Additional opportunities for research include the explicitly defining the how expert psychiatric nurses teach “middle phase” aggression prevention interventions to novice nurses.