Background: Physical restraints, such as bedrails, belts in chairs or beds, and fixed tables, are commonly used for older people in general hospital settings. Reasons given for using physical restraints are to prevent falls and fall-related injuries, to control challenging behavior (such as agitation or wandering), and to ensure the delivery of medical treatments. Clear evidence of their effectiveness is lacking, and potential harms are recognised, including injuries associated with the use of physical restraints and a negative impact on people's well-being. There are widespread recommendations that their use should be reduced or eliminated.

Objectives: To assess the best evidence for the effects and safety of interventions aimed at preventing and reducing the use of physical restraint of older people in general hospital settings. To describe the content, components and processes of these interventions.

Search methods: We searched the Cochrane Dementia and Cognitive Improvement Group's register, MEDLINE (Ovid SP), Embase (Ovid SP), PsycINFO (Ovid SP), CINAHL (EBSCOhost), Web of Science Core Collection (Clarivate), LILACS (BIREME), ClinicalTrials.gov and the World Health Organization's meta-register the International Clinical Trials Registry Portal on 20 April 2022.

Selection criteria: We included randomised controlled trials and controlled clinical trials that investigated the effects of interventions that aimed to prevent or reduce the use of physical restraints in general hospital settings. Eligible settings were acute care and rehabilitation wards. We excluded emergency departments, intensive care and psychiatric units, as well as the use of restrictive measures for penal reasons (e.g. prisoners in general medical wards). We included studies with a mean age of study participants of at least 65 years. Control groups received usual care or active control interventions that were ineligible for inclusion as experimental interventions.

Data collection and analysis: Two review authors independently selected the articles for inclusion, extracted data, and assessed the risk of bias of all included studies. Data were unsuitable for meta-analysis, and we reported results narratively. We used GRADE methods to describe our certainty in the results.

Main results: We included four studies: two randomised controlled trials (one individually-randomised, parallel-group trial and one clustered, stepped-wedge trial) and two controlled clinical trials (both with a clustered design). One study was conducted in general medical wards in Canada and three studies were conducted in rehabilitation hospitals in Hong Kong. A total of 1709 participants were included in three studies; in the fourth study the number of participants was not reported. The mean age ranged from 67 years to 84 years. The duration of follow-up covered the period of patients' hospitalisation in one study (21 days average length of stay) and ranged from 4 to 11 months in the other studies. The definition of physical restraints differed slightly, and one study did not include bedrails. Three studies investigated organisational interventions aimed at implementing a least-restraint policy to reduce physical restraints. The theoretical approach of the interventions and the content of the educational components was comparable across studies. The fourth study investigated the use of pressure sensors for participants with an increased falls risk, which gave an alarm if the participant left the bed or chair. Control groups in all studies received usual care. Three studies were at high risk of selection bias and risk of detection bias was unclear in all studies. Because of very low-certainty evidence, we are uncertain about the effect of organisational interventions aimed at implementing a least-restraint policy on our primary efficacy outcome: the use of physical restraints in general hospital settings. One study found an increase in the number of participants with at least one physical restraint in the intervention and control groups, one study found a small reduction in both groups, and in the third study (the stepped-wedge study), the number of participants with at least one physical restraint decreased in all clusters after implementation of the intervention but no detailed information was reported. For the use of bed or chair pressure sensor alarms for people with an increased fall risk, we found moderate-certainty evidence of little to no effect of the intervention on the number of participants with at least one physical restraint compared with usual care. None of the studies systematically assessed adverse events related to use of physical restraint use, e.g. direct injuries, or reported such events. We are uncertain about the effect of organisational interventions aimed at implementing a least-restraint policy on the number of participants with at least one fall (very low-certainty evidence), and there was no evidence that organisational interventions or the use of bed or chair pressure sensor alarms for people with an increased fall risk reduce the number of falls (low-certainty evidence from one study each). None of the studies reported fall-related injuries. We found low-certainty evidence that organisational interventions may result in little to no difference in functioning (including mobility), and moderate-certainty evidence that the use of bed or chair pressure sensor alarms has little to no effect on mobility. We are uncertain about the effect of organisational interventions on the use of psychotropic medication; one study found no difference in the prescription of psychotropic medication. We are uncertain about the effect of organisational interventions on nurses' attitudes and knowledge about the use of physical restraints (very low-certainty evidence).

Authors' conclusions: We are uncertain whether organisational interventions aimed at implementing a least-restraint policy can reduce physical restraints in general hospital settings. The use of pressure sensor alarms in beds or chairs for people with an increased fall risk has probably little to no effect on the use of physical restraints. Because of the small number of studies and the study limitations, the results should be interpreted with caution. Further research on effective strategies to implement a least-restraint policy and to overcome barriers to physical restraint reduction in general hospital settings is needed.