Predicting the use of restraint on adult inpatient psychiatric wards: a four-year incident report analysis

Background: The use of restraint on psychiatric inpatient wards varies widely and is poorly understood, but patient characteristics, situational factors and organisational factors have been implicated. Objectives: To understand whether use of restraint differed systematically between wards and sites and then define the relative contribution of patient, situational and organisational factors. Design: Restraint incidents were extracted from incident reporting over a fouryear period. Incidents were drawn from five acute wards and two psychiatric intensive care units across three hospital sites within an NHS Trust in the UK. Two research samples were created. An overall incident sample (N = 1297) was used to assess the degree of variability between wards and sites. Binary logistic regression in a first use of restraint sample (N = 435) then assessed the contribution made by significant patient, situational and organisational predictors. Results: No significant differences between the wards or sites were found for restraint rates and the levels of restraint used. Staff response and time of day were significant predictors of level of restraint. The regression model explained 16.9% of variance in level of restraint. Conclusion: Future research needs to better measure and then model the dynamic relationships between patient, organisational and situational factors that predict restraint. Intervention programmes that target restraint reduction need to be better evaluated.