Predicting mechanical restraint of psychiatric inpatients by applying machine learning on electronic health data.

These findings open for the development of an early warning system that may guide interventions to reduce the use of MR.