Patient characteristics associated with the use of mechanical restraints

To compare the characteristics of restrained patients with those of unrestrained patients by assessing a number of medical, behavioral, and cognitive variables including a disruptive-behavior inventory.

Case-control study.

A 719-bed university-affiliated teaching hospital.

The 80 cases were patients identified by the nursing staff as having had a restraint applied within the last 24 hours prior to entry in the study. The 80 unrestrained controls were selected from the rooms adjacent to the cases' in order to match for proximity to the nursing station and nurse staffing.

Demographic data, data on diagnoses and treatments, results of the Folstein Mini-Mental State (MMS) test and an eight-item disruptive-behavior inventory, and outcome information were obtained for each patient using a standardized procedure. Three important patient characteristics were significantly associated with restraint use in a multiple logistic regression model: disruptive behaviors, nursing assessment of risk of falling, and cognitive impairment. Cases were older than controls, but age was not an independent characteristic associated with restraint use when controlling for cognitive impairment, risk of falling, and disruptive behaviors.

Restraint use was more likely in patients with disruptive behaviors, at risk of falling, and with cognitive impairment. Attention to these factors and alternative strategies for dealing with them may reduce the use of physical restraints.