There has been growing concern about the levels of violence on psychiatric wards in recent years, which has led to increased emphasis on the safe management of such behavior. Physical intervention (PI) is a form of “hands on restraint” to manage the aggressive patient. PI was originally introduced as control and restraint in high security special hospitals in the mid-1980s to enable nursing staff to manage aggressive behavior in a systematic manner. The use of PI is highly emotive and controversial, yet staff are trained in the techniques across all mental health services in the U.K. (Wright et al., Reference Wright, Sayer, Par, Gray, Southern and Gournay2005). To be legally and ethically justifiable, the form of PI employed must be an appropriate, proportionate and reasonable response to the risk posed and be terminated at the earliest possible opportunity (National Institute of Clinical Excellence, 2005).

There has been a rapid growth in the training and use of PI in all sectors of mental health; in a recent national survey, Wright et al. (Reference Wright, Sayer, Par, Gray, Southern and Gournay2005) reported that 77% of staff were trained in PI. Despite this rapid growth there remains a paucity of literature detailing the safety and effectiveness of PI. In a Cochrane Systematic Review, Sailas and Fenton (Reference Sailas and Fenton2000) stated there were no trials of sufficient quality with regard to effectiveness to make strong recommendations about the procedure. In a national survey of PI training it was revealed that only 23% of participants were taught how to deal with “physical handicap” (Lee et al., Reference Lee, Wright, Sayer, Parr, Gray and Gournay2001). This is concerning as it is common for older adults to have restricted movement and arthritic joints or other physical complaints.

Few studies have published data on the frequency of PI use and much of the literature focuses on its use on adults with a mental illness and/or a learning disability. The survey by Wright et al. (Reference Wright, Sayer, Par, Gray, Southern and Gournay2005) reported that PI was used at a mean frequency of 3.13 times per month (range 0.1–100). However, no studies have yet been published on the incidence of PI in the older adult inpatient setting.

During their training, up to 29% of staff sustain an injury; this is alarming as the process is meant to be passive (Lee et al., Reference Lee, Wright, Sayer, Parr, Gray and Gournay2001). Questions must therefore be raised about the suitability of the techniques for use with aggressive older adults who are likely to have a range of physical health problems. No data are yet available on injuries sustained within any patient group from the application of PI. Currently, the only single piece of advice given to professionals working with aggressive older adults is that given by the National Institute for Mental Health in England (2004) which recommends that “extra care” is taken with those in physical decline. This is open to interpretation and has the potential to increase the risk of PI techniques being abused.

St Andrews Healthcare is a leading independent charitable provider of specialist inpatient mental health services in the U.K. An audit of PI incidents was completed in the Townsend Division, a 102-bed unit for older adults with complex mental health needs. Between January 2004 and December 2006 there were 292 incidents of PI involving 44 patients (male = 32; female = 12): 61 incidents of PI were reported in 2004 (20.8%), 116 in 2005 (39.7%) and 115 in 2006 (39.3%). During the audit period, the mean incidence of PI was 5.02 per month (range 1–22). The mean incidence among the 44 patients was 4.1 (range 1–29). The large majority of PI incidents lasted for less than 10 minutes (217 incidents = 74.3%).

During the 292 incidents of PI, 33 injuries were sustained (i.e. 11.3% of PI incidents resulted in injury). Of the 33 injuries, six were to patients (18.2%) and 27 were to staff (81.8%); thus, of all of PI incidents, 2% resulted in injury to patients and 9.2% in injury to staff. The arms were the most common site of injury in patients (3/6 = 50%), while on two occasions the site of injury was not recorded (2/6 = 33.3%) and one injury was sustained to a patient's face (1/6 = 16.6%). Healthcare assistants were most likely to be injured (16/27 = 59.2%) as they tend to be involved in administering PI. Ten qualified nurses received an injury (37.0%) and one member of the conflict management team who is responsible for training staff in PI was injured (3.7%). Injuries to staff were most common to the arms (10/27 = 37.0%); four staff sustained injuries to the face (4/27 = 14.8%) and it is reasonable to assume this was caused by patient aggression and not from applying the techniques. There were three cases of multiple injuries (3/27 = 11.1%), two to the legs (2/27 = 7.4%), one to the back (1/27 = 3.7%); in seven incidents the site of injury was not recorded (7/27 = 25.9%).

This audit provides the first data on the incidence of PI among older adults and of staff injuries acquired while implementing it. It also provides the first data on patient injuries sustained from implementing PI in any mental health population. We think that the low injury rate to patients and staff, an incidence of 2% and 9.2% respectively, is attributable to the detailed individual care plans adopted for the older adult patients. A blanket approach when employing PI is avoided; rather, the PI techniques employed are adapted to meet the individual needs of each patient. As recommended by the National Institute of Clinical Excellence (2005), PI is always the last resort. Future research is needed to establish the best techniques suitable for PI in older adults.