Objective: Crisis plans are a type of psychiatric advance statement (PAS) that describe how to recognize early signs of a psychiatric crisis and how to handle crisis situations. Although PASs may help reduce the occurrence of crises, their implementation in clinical practice is problematic. This study examined patient and clinician characteristics associated with successful completion of a crisis plan and determined how often a plan was consulted in a crisis situation. Methods: Participants were 139 crisis-prone outpatients and their clinicians recruited in the context of a randomized controlled trial in the Netherlands studying the effects of crisis plans. Multivariate logistic regression analyses examined associations between patient and clinician characteristics and completion of a crisis plan. Descriptive and univariate statistics were used to determine subsequent use of crisis plans. Results: Of the 139 patients, 64% completed a crisis plan. Higher completion rates were associated with a better clinician-rated working alliance, a lower education level of the patient, and fewer years of professional experience of the clinician. Of the 89 patients who completed a crisis plan, 38 later experienced a crisis, and plans were consulted for a third of these patients (34%). For those whose crisis involved involuntary admission, a smaller proportion had their plans consulted, compared with those whose crisis involved an outpatient emergency visit or voluntary admission. Conclusions: PAS completion was associated not only with patient and clinician characteristics but also with a positive working alliance. PAS use did not occur in all cases, especially not when patients were involuntarily admitted.