Valid knowledge about risk stratification is essential for the development of preventive approaches to reduce restraints. This retrospective, monocentric study examined all 248 restraining measures used on 79 patients of a university psychiatric department in 2015. Eight institutional and seven patient-related factors were statistically calculated using logistic regression analysis and matched pairs. A statistically significant increased probability of fixation was found in connection with a lower number of nurses per patient, a higher number of patients admitted by law, lower professional experience of nurses and in the nursing late shift as well as when admission occurred via emergency medical service. Based on the results, possible preventive organizational measures are elaborated and preconditions for a prospective evaluation of the complex factors influencing coercive measures are discussed.