I N T EG R A T I V E L I T E R A T U R E R E V IE W S A N D M E T A - A N A L Y S E S
Patient injury and physical restraint devices: a systematic review David Evans
PhD RN
Lecturer, Department of Clinical Nursing, University of Adelaide, Adelaide, Australia
Jacquelin Wood
BN MN RN
Nursing Director, Royal Adelaide Hospital, Adelaide, Australia
and Leonnie Lambert
BN RN
Nursing Director, Royal Adelaide Hospital, Adelaide, Australia
Submitted for publication 23 October 2001 Accepted for publication 25 September 2002
Correspondence: David Evans, Department of Clinical Nursing, University of Adelaide, Adelaide, South Australia, Australia 5005. E-mail:
[email protected]
E V A N S D ., W O O D J . & L A M B E R T L . ( 2 0 0 3 ) Journal of Advanced Nursing 41(3), 274–282 Patient injury and physical restraint devices: a systematic review Objective. To investigate physical restraint-related injuries. Areas of interest were the prevalence of injury, types of injuries, risk of sustaining an injury and specific restraint devices associated with injury. Definitions. Injury in the context of this review was considered to be either direct injury, such as lacerations and strangulation, or indirect injury considered to be an adverse outcome such as increased mortality rates or duration of hospitalization. Method. A comprehensive search was undertaken that involved all major databases and the reference list of all relevant papers. To be included in the review studies had to involve people in acute or residential care settings and report data related to injury caused by restraint devices. A number of different types of research designs were included in the review. The findings of studies were pooled using odds ratio and narrative discussion. Results. The search identified 11 papers reporting the findings of 12 observational studies. These studies were supplemented with the findings of a number of other types of studies that reported restraint-related data. The review highlights the potential danger of using physical restraint in acute and residential health care facilities. Observational studies suggest that physical restraint may increase the risk of death, falls, serious injury and increased duration of hospitalization. However, there is little information to enable the magnitude of the problem to be determined. Discussion. Many of the findings highlight the urgent need for further investigation into the use of physical restraint in health care facilities. Further research should investigate the magnitude of the problem and specific restraint devices associated with injury. However, given the limited nature of the evidence, this association should be investigated further using rigorous research methods.
Keywords: physical restraint, injury, adverse outcomes, aged care, acute care
Background The physical restraint of people during admission to acute and residential care facilities has been common practice for 274
many years. Physical restraint devices have been used as a means of preventing confused people from wandering or tampering with medical devices and to minimize the risk of falls and injury. However, while restraint may reduce some 2003 Blackwell Publishing Ltd
Integrative literature reviews and meta-analyses
What is already known about this topic • There is increasing concern about the use of physical restraint in health care. • A growing number of papers have reported a range of different injuries as a result of the use of restraint. • However, there has been no attempt to summarize this literature.
What this study adds • This paper provides a rigorous summary of research investigating injury secondary to physical restraint. • Findings suggest that physical restraint may increase the risk of death, falls and serious injury. • There is little information to enable the magnitude of the problem to be determined. • The review also highlights the urgent need for further investigation into the use of physical restraint in health care facilities.
hazards, being restrained brings other risks. An initial search of the literature found that restraint-related injury and death have been reported in a number of papers (DiMaio et al. 1986, Miles & Irvine 1992, Parker & Miles 1997). However, despite this growing number of reports, there has been no rigorous and systematic evaluation of available evidence. To address this, a systematic review was conducted to investigate the link between physical restraint and adverse events.
The review Method Objectives The objective of this review was to investigate restraintrelated injuries. Specific areas of interest were the prevalence of restraint-related injuries, type of injuries that have been reported, risk of sustaining injury and injuries caused by specific restraint devices. For the purposes of the review, physical restraint was defined as ‘any device, material or equipment attached to or near a person’s body and which cannot be controlled or easily removed by the person and which deliberately prevents or is deliberately intended to prevent a person’s free body movement to a position of choice and/or a person’s normal access to their body’ (Retsas 1998, p. 186). Chemical restraints are also used to control behaviour; however, because of the
Injury and physical restraint devices
different modes of action of physical and chemical restraint, these were excluded from the review. A number of different types of injuries were identified in the literature. To allow this information to be grouped into a logical summary, two broad categories were developed: 1 Direct injury: physical injury caused as a direct result of external pressure from the restraining device, including lacerations, bruising or strangulation. 2 Indirect injury: adverse outcomes related to a number of different events, including increased mortality rate, development of pressure sores, falls, falls-related injury or failure to be discharged home. Inclusion criteria Studies were included in the systematic review if they met the following inclusion criteria: the study involved people in an acute care hospital or residential care facility, investigated the use of physical restraint, and reported outcomes related to direct or indirect injury. To compare outcomes between restrained and unrestrained people all randomized controlled trials (RCT) and observational studies that met the inclusion criteria were considered. To determine the prevalence of restraint-related injury, descriptive and observational studies were considered. To investigate the specific nature of injuries, descriptive studies and case reports were considered. Studies were not formally critically appraised prior to inclusion in the review and this issue is addressed under the limitations of the review. Search strategy A three-step search strategy was used. An initial limited search of MEDLINE and CINAHL was undertaken to identify keywords, followed by a second search using all relevant search terms. The databases searched were CINAHL, MEDLINE, Current Contents, Cochrane Library, PsycInfo, Embase, HealthSTAR, Expanded Academic Index and Dissertation Abstracts International. The search terms used were restrain* in ti, bedrail*, siderail* and cotside*. During the search, the * character truncates the search term and ‘ti’ limits the search to the title. The third stage entailed inspection of the reference lists of all relevant papers to identify additional studies. Data synthesis Odds ratios (OR) and 95% confidence intervals (95% CI) were used to investigate differences between groups. Heterogeneity between combined studies was tested using standard chi-square test. When statistical pooling was not appropriate or possible, the findings were summarized by narrative.
2003 Blackwell Publishing Ltd, Journal of Advanced Nursing, 41(3), 274–282
275
D. Evans et al. Table 1 Details of observational studies reporting restraint-associated injury Author
Setting
Population
Findings
Burton et al. (1992)
Eight residential care facilities
Residents – Restrained, n ¼ 209 – Unrestrained, n ¼ 228
Residents admitted with moderate to no cognitive impairment experienced greater cognitive decline if restrained No difference for severely cognitively impaired residents
Capezuti et al. (1996)
Residential care
Residents – Retrained, n ¼ 119 – Unrestrained, n ¼ 203
Restraint was not associated with a significantly lower risk of falls or injuries in confused ambulatory residents
Folmar and Wilson (1989)
Residential care
Residents – Never restrained, n ¼ 81 – Always restrained, n ¼ 21
More nonsocial behaviour observed in restrained residents Highest level of social behaviour observed in unrestrained residents
Frengley and Mion (1986)
General medical wards of acute care hospital
Patients – Unrestrained, n ¼ 1197 – Restrained, n ¼ 95
Increased length of hospitalization of restrained compared with unrestrained patients Restrained had higher mortality rate than unrestrained patients
Lofgren et al. (1989)
General medical wards of acute care hospital
Patients – Restrained £4 days, n ¼ 67 – Restrained >4 days, n ¼ 35
More new pressures sores and nosocomial infections in patients restrained for more than 4 days
Mion et al. (1989) Study I
Acute medical ward
Patients – Restrained, n ¼ 35 – Unrestrained, n ¼ 243
Higher mortality in restrained patients. Increased length of stay in restrained patients More falls, nosocomial infections and immobility-related complications in restrained patients
Mion et al. (1989) Study II
Acute care medical rehabilitation ward
Patients – Restrained, n ¼ 49 – Unrestrained, n ¼ 143
No difference in length of hospitalization between groups More falls, nosocomial infections and procedure related complications in retrained group
Mosley (1997)
Residential care
Residents – Restrained, n ¼ 740 – Unrestrained, n ¼ 4333
Restraint was associated with disorientation, and greater dependency for both activities of daily living and walking
Robbins et al. (1987)
Acute care hospital
Medical and surgical patients – Restrained, n ¼ 37 – Unrestrained, n ¼ 185
Higher mortality rate for restrained patients. Increased length of stay for restrained patients
Stiebeling et al. (1990)
Residential care
Residents – Restrained, n ¼ not stated – Unrestrained, n ¼ not stated
Restrained residents had more bladder and bowel incontinence, pressure ulcers and reduced mobility (abstract so minimal data provided)
Tinetti et al. (1992)
Twelve skilled nursing facilities
Residents – Restrained, n ¼ 122 – Unrestrained, n ¼ 275
Higher fall-related injury rate for restrained residents
Werner et al. (1989)
Residential care
Residents with high level of agitation and cognitive impairment – Restrained, n ¼ 22 – Never restrained, n ¼ 2
Significantly more strange noises and movement when residents restrained More agitation when resident was restrained Higher rates of agitation immediately following application of restraint
276
2003 Blackwell Publishing Ltd, Journal of Advanced Nursing, 41(3), 274–282
Integrative literature reviews and meta-analyses
Results The search identified 11 papers reporting the findings of 12 observational studies (see Table 1). These findings have been supplemented by those of before and after studies reporting injury data following a reduction in restraint use, descriptive studies reporting on injuries and deaths associated with the use of physical restraint, and case studies reporting specific restraint-related injuries.
Prevalence of injury Few studies have addressed the prevalence of physical restraint-related injury. One cohort study, involving a total of 222 acute care patients of whom 37 were restrained, found no serious restraint-related injuries and reported only one case of minor skin abrasion (Robbins et al. 1987). A prospective study of physical restraint involving 12 skilled nursing facilities and 1756 residents over a one-year period also reported no serious injury resulting from their use (Tinetti et al. 1991). However, this study did not attempt to investigate minor restraint-related injuries. Despite these findings, investigations of death certificates and records clearly demonstrate that deaths have occurred as a consequence of using physical restraints (Joint Commission for the Accreditation of Health Care Organizations 1998, Miles & Irvine 1992, Parker & Miles 1997). Currently there is little information on the prevalence of restraint-related injury. While there is a suggestion that serious injuries are infrequent, this appears to be contradicted by investigations of death records. However, investigations of death records provide little information on prevalence, which is not known. This is an area in need of further investigation to determine the magnitude of the problem.
Type of injuries A large number of studies and reports have addressed physical restraint and direct or indirect injury. Direct injury Direct injury in this review refers to actual physical damage as a result of external pressure from restraint devices. This information was reported in case reports and descriptive studies: A. Nerve injury. One paper reported two cases of nerve injury as a result of the combination of vest and wrist restraint in patients nursed with the head of the bed elevated (Scott & Gross 1989). This report suggested that the
Injury and physical restraint devices
combination of wrist restraint and bed position caused the person to slide down the bed, resulting in the vest restraint moving up into the axillae and thereby putting pressure on the distal brachial plexus. B. Sudden death. Two case reports described sudden death associated with the use of restraint devices. In one case an older man admitted to hospital with atrioventricular conduction abnormalities had limb and vest restraints applied and after a period of prolonged agitation and struggling suffered sustained ventricular tachycardia and died (Robinson et al. 1993). A similar case of an very agitated older women fighting against physical restraint for several hours followed by sudden death has also been reported in a ‘Letter to the Editor’ (Miles 1993). C. Asphyxiation. A small number of papers were identified that linked use of restraint devices to asphyxiation. One reported two deaths attributed to mechanical asphyxiation by a vest restraint (Dube & Mitchell 1986). Another described the death of three nursing home residents found hanging by a vest restraint over the bedrails, and a fourth found in chair with the vest caught against their neck (DiMaio et al. 1986). One report described the death of a 33-year-old man in an acute care hospital found hanging over the bedrails suspended by the vest restraint device (Langslow 1999). Another describing the strangulation of six people by restraint devices was published in 1981, highlighting the fact that restraintrelated deaths are not a new occurrence (Katz et al. 1981). D. Death. Retrospective investigations of death certificates and records have identified a large number of restraint-related deaths. One review of the records of 19 restraint-related deaths found that these have involved people who were in beds and chairs (Miles & Irvine 1992). Another review of death certificates and reports involving bedrails identified 74 deaths, with the most common cause being people getting trapped between the mattress and bedrail (Parker & Miles 1997). An investigation of 20 restraint-related deaths found that most occurred in psychiatric hospitals (12 of 20), although some occurred in acute hospitals (6 of 20) and long-term care facilities (2 of 20) (Joint Commission for the Accreditation of Health Care Organizations 1998). A survey of coroners identified 63 restraint-related asphyxial deaths, of which 61% occurred in nursing homes and 24% in hospitals (Ruben et al. 1993). In summary, these reports demonstrate that physical restraint causes severe injury and death. However, the retrospective nature of studies, and the fact that death records were the most common source of data, means that
2003 Blackwell Publishing Ltd, Journal of Advanced Nursing, 41(3), 274–282
277
D. Evans et al.
Study
Restrained
Unrestrained
Odds ratio (95% CI)
Failure to be discharged home Mion et al. (1989) (medical) Robbins et al. (1987) Pooled results
20/37 24/35
22/185 46/428
8Æ72 (3Æ98–19Æ11) 18Æ12 (8Æ34–39Æ39) 12Æ42 (7Æ16–21Æ52)
Death during hospitalization Frengley et al. (1986) Mion et al. (1989) (medical) Robbins et al. (1987) Pooled results
11/95 5/35 9/37
13/1197 3/243 6/185
11Æ93 13Æ33 9Æ59 11Æ24
Nosocomial infection Mion et al. (1989) (rehabilitation) Mion et al. (1989) (medical) Pooled results
29/49 8/35
31/94 13/243
2Æ95 (1Æ44–6Æ02) 5Æ24 (1Æ99–13Æ78) 3Æ46 (1Æ93–6Æ22)
Fall during hospitalization Mion et al. (1989) (rehabilitation) Mion et al. (1989) (medical) Pooled results
28/49 6/35
18/94 3/243
5Æ63 (2Æ62–12Æ09) 16Æ55 (3Æ93–69Æ75) 6Æ79 (3Æ44–13Æ39)
the magnitude of the problem cannot be determined from these reports. Additionally, it is also possible that many other deaths and severe injuries have occurred, but as a result of issues such as under-reporting of adverse events, have escaped detection by coroners’ officers and therefore researchers. Some of the adverse events identified were described in single case study reports. However, the search had a time frame of approximately 40 years. Therefore, the small number of reports identified during the search is surprising. The importance of one publication reporting the occurrence of an adverse event is difficult to determine. Events such as sudden death or nerve injury may be rare and so unlikely to occur a second time. However, the small number of reports identified may also reflect under-reporting of adverse events.
Table 2 Indirect injury and use of restraint in acute care settings
(5Æ19–27Æ43) (3Æ03–58Æ62) (3Æ17–29Æ01) (6Æ07–20Æ83)
indirect injuries are summarized by clinical settings, acute or residential care:
Indirect injury A number of indirect injuries have been linked to the use of physical restraint devices. Indirect injury in the context of this review refers to adverse outcomes such as increased mortality rate or development of pressure sores. These
A. Acute care setting. Based on observational studies (see Table 2), findings suggest that restrained patients are less likely to survive to discharge (OR ¼ 12Æ42, 95% CI 7Æ16–21Æ52), and more likely to die (OR ¼ 11Æ24, 95% CI 6Æ07–20Æ83), develop a nosocomial infection (OR ¼ 3Æ46, 95% CI 1Æ93–6Æ22) and fall during hospitalization (OR ¼ 6Æ79, 95% CI 3Æ44–13Æ39). Additionally, four observational studies suggest that restrained patients also have an increased length of hospitalization (Frengley & Mion 1986, Robbins et al. 1987, Lofgren et al. 1989, Mion et al. 1989). However, these findings could not be pooled in a metaanalysis due to incomplete reporting. Restrained patients were also found to be more likely to develop bladder and bowel incontinence (Lofgren et al. 1989) and have mobility-related problems (Mion et al. 1989). A single observational study suggests that patients restrained for periods of more than 4 days have a higher incidence of
Study
Restrained
Unrestrained
Odds ratio (95% CI)
Falls Capezuti et al. (1996) Tinetti et al. (1992) Pooled results
49/119 92/122
100/203 116/275
0Æ72 (0Æ46–1Æ14) 4Æ20 (2Æ61–6Æ77) 1Æ72 (1Æ26–2Æ35)
Serious falls-related injury Tinetti et al. (1992)
21/122
15/275
3Æ60 (1Æ79–7Æ27)
Falls-related fractures Tinetti et al. (1992)
12/122
6/275
4Æ89 (1Æ79–13Æ36)
278
Table 3 Indirect injury and use of restraint in residential care settings
2003 Blackwell Publishing Ltd, Journal of Advanced Nursing, 41(3), 274–282
Integrative literature reviews and meta-analyses
nosocomial infections and new pressure ulcers than those restrained for four days or less (Lofgren et al. 1989). B. Residential care setting. Based on the findings of observational studies (see Table 3) restrained residents are more likely to fall (OR ¼ 1Æ72, 95% CI 1Æ26–2Æ35), suffer serious fall-related injury (OR ¼ 3Æ60, 95% CI 1Æ79–7Æ27) or fallrelated fracture (OR ¼ 4Æ89, 95% CI 1Æ79–13Æ36) than unrestrained residents. A single observational study found those restrained on admission had an increased incidence of bladder and bowel incontinence, pressure ulcers and immobility (Stiebeling et al. 1990). One study investigated the impact of being restrained intermittently or continuously on fall-related injury (Tinetti et al. 1992). Findings suggest that residents who were restrained continuously were more likely to suffer a serious fall-related injury (OR ¼ 3Æ65, 95% CI 1Æ39–9Æ64) or a falls-related fracture (OR ¼ 3Æ41, 95% CI 1Æ01–11Æ55). One observational study found that residents exhibited the same amount, or more, agitated behaviours when they were restrained than when they were not restrained (Werner et al. 1989). Another found that restrained residents experienced greater cognitive decline than residents who were not restrained (Burton et al. 1992). One study addressing social behaviour and restraint found striking differences between residents who were restrained and unrestrained (Folmar & Wilson 1989), some engaging in significant social activity when not restrained and ceasing to socialize completely when they were put in restraints. Physical restraint has also been associated with an increase in disorientation, and dependence during physical activities of daily living and walking (Moseley 1997). A before and after study investigated the impact of discontinuing restraints on the likelihood of residents sustaining injury and found a reduced risk of sustaining serious injuries (Neufeld et al. 1999). However, the number of hip fractures remained constant and there was an increase in minor injuries. From a different perspective, a number of studies have evaluated restraint minimization programmes in residential care settings and reported data on injury before and after restraint reduction. However, the findings of the impact of restraint minimization on injury are contradictory. Some studies report no change in the incidence of falls (Werner et al. 1994, Levine et al. 1995, Si et al. 1999), while another reported an increase in nonserious falls (Ejaz et al. 1994). Two studies suggest that the incidence of serious injury does not change following a reduction in restraint use (Levine et al. 1995, Si et al. 1999). A single study evaluating reduction in use of bedrails reported a reduced incidence of serious injuries and no change in number of falls (Hanger et al. 1999).
Injury and physical restraint devices
In summary, use of restraint in acute care settings may increase the risk of adverse outcomes, including death, infection and falls. Restrained patients are also less likely to survive and be discharged home and more likely to have an increased duration of hospitalization. Additionally, prolonged use of physical restraints may further increase these risks. Similarly, in residential settings the use of physical restraint has been associated with a number of adverse outcomes such as falls and falls-related injury. The findings also suggest that restraints may increase agitation and cognitive decline, and reduce social interaction.
Injuries and specific restraint devices Vest restraints A number of reports have implicated vest restraints in deaths (Katz et al. 1981, DiMaio et al. 1986, Dube & Mitchell 1986, Berrol 1988, Miles & Irvine 1992, Langslow 1999). These deaths have involved people in beds and chairs, and a small number were related to the incorrectly fitted of the vest (Ruben et al. 1993). Reports commonly describe the person being partially able to climb out of the bed or chair, resulting in the vest being caught around their neck (DiMaio et al. 1986, Langslow 1999). Alternatively, the person partially left the bed and was left suspended with the vest around their chest, impairing breathing (Miles & Irvine 1992). However, it should be noted that while a large number of deaths and injuries have been linked to vest restraints, these are one of the more commonly used devices. Bedrails Bedrails have also been linked to a number of deaths (Ruben et al. 1993, Parker & Miles 1997, Miles & Parker 1998). A descriptive study of 63 reports of restraint-related deaths found that 21% involved bedrails (Ruben et al. 1993). These commonly involved compression of the chest or abdomen by the bedrail. One report described a person slipping between two half-length bedrails, while in another the person became trapped between the bedrail and mattress (Miles & Parker 1998). It was noted that the raised head of the bed assisted slipping down between the two half-length bedrails. An investigation of death certificates involving bedrails identified 74 cases (Parker & Miles 1997), described as three distinct types: • being trapped between the mattress and bedrail with the face buried in the mattress, resulting in asphyxiation; • being trapped by the bedrail while still in the bed; and • slipping from bed and being trapped by the bedrail, for example by the head or hips (Parker & Miles 1997).
2003 Blackwell Publishing Ltd, Journal of Advanced Nursing, 41(3), 274–282
279
D. Evans et al.
From a different perspective, one study evaluated the impact of limiting the use of bedrails in a rehabilitation unit and reported a reduction in serious injuries and increase in minor injuries (Hanger et al. 1999). Waist restraints There is little information specifically addressing waist restraints. A descriptive study of 63 reports of deaths related to restraint devices reported that 18% involved these (Ruben et al. 1993). Wrist restraints There is little information specifically addressing wrist restraints. A descriptive study of 63 reports of restraintrelated deaths, found two deaths involving these (Ruben et al. 1993). Other restraint devices A number of other restraint devices are used in acute and residential care settings, including ‘gerichairs’, ‘geritables’ and even bed sheets. However, the search failed to identify any reports linking these to injury. In summary, most device-specific reports on injuries concern vest restraints and bedrails. While this may mean that these pose a greater risk, it may also reflect their common use. The significance of these reports is difficult to determine in the absence of any systematic evaluation of specific restraint devices.
these retrospective studies. Secondly, investigations of coroners’ death reports would not identify serious nonfatal injuries. Consequently, the only serious nonfatal injuries identified during this review were those published in a small number of case reports. A number of indirect injuries were also identified. For hospital patients, use of restraints was linked to prolonged hospitalization, increased mortality, increased nosocomial infection, and being less likely to survive and be discharged home. Restrained patients were also more likely to fall during hospitalization. Prolonged periods of restraint may also increase the risk of adverse outcomes. For nursing home residents, the use of restraints was linked to increased agitation, cognitive decline, reduced social behaviour and reduced mobility. Continuously restrained residents appeared to be at greater risk than those restrained only intermittently. Finally, discontinuation of restraints may reduce some of these risks.
Under-reporting
While this review of restraint-related injuries provides some information on the occurrence and nature of these injuries, many issues have yet to be adequately addressed.
It has been suggested that there are ethical, humanitarian and financial imperatives to finding out what is going wrong, collate and analyse the information and devise strategies to prevent the problems (Runciman 1999). However, it appears that there has been only minimal sharing of information addressing restraint-related injuries in the literature. While investigations of records, such as those kept by coroners, have identified deaths linked to restraint devices, these have not been matched by primary reports in the health care literature. There is also scant information about nonfatal restraintrelated injuries. This suggests that restraint-related injury is under-reported. If this problem is to be adequately explored and strategies developed to minimize the risks, determining its magnitude is vital.
How often do injuries occur?
Validity of findings
While there are a number of papers addressing different aspects of restraint-related injury, there is currently little information on its prevalence. As a consequence, it is impossible to determine the magnitude of the problem.
Some areas of restraint-related injuries have yet to be subject to systematic evaluation and so the validity of findings reported here must be interpreted with caution. However, given the nature of the issue, evaluation of injuries by RCTs may not be ethically appropriate. Despite this, further investigation is urgently needed to confirm or refute the findings of existing studies. This should focus on concerns raised by this review that restraint use may increase the risk of injury. Studies have reported descriptive investigations of death records (Miles & Irvine 1992, Parker & Miles 1997), but this provides little information on prevalence. Case reports have demonstrated that adverse events such as nerve injury (Scott & Gross 1989), sudden death (Robinson et al.
Discussion
What injuries? Death is the most commonly reported adverse event directly related to the use of physical restraint. However, it should be noted that because of the seriousness of this event, it is far more likely to be identified during retrospective record reviews than less serious injuries. It is possible that some serious but nonfatal injuries may not have been detected by 280
2003 Blackwell Publishing Ltd, Journal of Advanced Nursing, 41(3), 274–282
Integrative literature reviews and meta-analyses
1993) and asphyxiation (Dube & Mitchell 1986) have occurred, but this type of report provides little information on relative risk.
Demonstrating causality Caution is needed when interpreting findings generated by observational studies because assigning causality is difficult when the intervention is closely linked to the outcomes. Determining cause and effect relationships is best achieved using randomized controlled trials, as this ensures that study groups are equal for all variables other than the intervention. For observational studies, determining the population of study groups is achieved through processes other than chance. This means that unknown variables may have an influence on findings. In the context of injury and physical restraint, it may be that other factors are responsible for the adverse outcomes. For example, deteriorating physiological status may not only result in increased risk of adverse outcomes, but may also increase the risk of being subject to physical restraint. In this example, restraint would be more a ‘marker’ of impending physiological decline rather than its cause. This limitation is acknowledged in presenting the present findings. However, evaluation of this relationship using the randomized controlled trial design may be difficult for ethical reasons.
Critical appraisal Studies were not critically appraised before inclusion in the review. This decision was the result of the difficulty in identifying appropriate criteria to appraise the range of different types of research included. For RCTs, a large number of appraisal tools exist that have been subject to varying degrees of evaluation. However, how best to distinguish good quality observational studies from poor is less clear. There is also little information on how best to appraise descriptive studies. Determining their validity is further complicated because some used data sources (such as death records) where accuracy and completeness of information could not be determined. However, in this review these descriptive studies provided the only information concerning restraint-related deaths. As a result of these factors studies were not critically appraised before inclusion and this limitation is acknowledged.
Injury and physical restraint devices
as increased mortality. While the findings of this review are based on a range of different types of research, it is clear that physical restraint is not a benign intervention. Based on current evidence, a number of recommendations can reasonably be made. Additionally, a number of published recommendations were identified during the review that addressed restraint-related injury and these have been used to guide the following recommendations (Joint Commission for the Accreditation of Health Care Organizations 1998, Katz et al. 1981, Berrol 1988, Scott & Gross 1989, Parker & Miles 1997): 1 As there are risks associated with the use of restraint: • physical restraint should only be used as a last resort; • physical restraint should only be used when potential benefits outweigh potential harm. 2 As prolonged use of restraint may increase risk: • the need for continued physical restraint should be assessed regularly. 3 As multiple restraint devices may increase risk: • only the minimal level of physical restraint should be used to ensure a person’s safety. 4 As some injuries are related to attempts to exit the device: • the physically restrained person should be continuously observed and isolation should be avoided. In addition, it seems prudent of offer two further general recommendations: • If physical restraint devices are used, then manufacturers’ recommendations must be followed. • Health care workers should be educated about the dangers of physical restraint, correct application of restraining devices and safe management of the restrained person.
Implications for research Many issues arise from this review which highlight the urgent need for further research into the use of physical restraint in health care facilities. The magnitude of the problem is not known and warrants further investigation. There are some suggestions in the literature that minimizing use of physical restraint reduces risk of injury. The issues associated with short versus long and intermittent versus continuous restraint are in need of further investigation. Finally, restraint minimization and investigation into methods by which this can be safely achieved should be considered a research priority.
Implications for practice
Conclusions
The most important implication for practice is that use of physical restraint brings a risk of adverse outcomes. These may be direct injury, such as strangulation, or indirect, such
The findings of this review highlight the potential danger of using physical restraint in acute and residential health care facilities. The major adverse event reported in the literature is
2003 Blackwell Publishing Ltd, Journal of Advanced Nursing, 41(3), 274–282
281
D. Evans et al.
death. Observational studies suggest that restraints may also increase the risk of death or prolonged hospitalization. Studies suggest that prolonged use of restraint may increase these risks, and that continuous restraint may be pose greater a risk than intermittent restraint. However, there is little information to enable the magnitude of the problem to be determined. Based on the findings of this review, recommendations proposed are that restraint should only be used as a last resort, minimal level of restraint be used, for the minimal duration and that the restrained person should be closely monitored. In conclusion, the review highlights many areas for further investigation. Most important is the need to explore methods by which use of physical restraint can be minimized because physical restraint devices may be doing more harm than good.
References Berrol S. (1988) Risks of restraints in head injury. Archives of Physical Medicine and Rehabilitation 69, 537–538. Burton L.C., German P.S., Rovner B.W. & Brant. L.J. (1992) Physical restraint use and cognitive decline among nursing home residents. Journal of the American Geriatrics Society 40, 811–816. Capezuti E., Evans L., Strumpf N. & Maislin G. (1996) Physical restraint use and falls in nursing home residents. Journal of the American Geriatrics Society 44, 627–633. DiMaio V.J., Dana S.E. & Bux R.C. (1986) Deaths caused by restraint vests. Journal of the American Medical Association 255, 905. Dube A. & Mitchell E. (1986) Accidental strangulation from vest restraints. Journal of the American Medical Association 256, 2725–2726. Ejaz F.K., Folmar S.J., Kaufmann M., Rose M.S. & Goldman B. (1994) Restraint reduction: can it be achieved? Gerontologist 34, 694–699. Folmar S. & Wilson H. (1989) Social behaviour and physical restraints. Gerontologist 29, 650–653. Frengley J.D. & Mion L.C. (1986) Incidence of physical restraints on acute general medical wards. Journal of the American Geriatric Society 34, 565–568. Hanger H.C., Ball M.C. & Wood L.A. (1999) An analysis of falls in the hospital: can we do without bedrails. Journal of the American Geriatrics Society 47, 529–531. Joint Commission for the Accreditation of Health Care Organisations (1998) Preventing restraint deaths. Sentinel Event Alert, JCAHO, Washington. Katz L., Weber F. & Dodge P. (1981) Patient restraint and safety vests. Dimensions of Health Service 58, 10–11. Langslow A. (1999) Safety and physical restraint. Australian Nurses Journal 7, 34–35. Levine J.M., Marchello V. & Totolos E. (1995) Progress toward a restraint-free environment in a large academic nursing facility. Journal of the American Geriatrics Society 43, 914–918.
282
Lofgren R.P., MacPherson D.S., Granieri R., Myllenbeck S. & Sprafka J.M. (1989) Mechanical restraints on the medical wards: are protective devices safe? American Journal of Public Health 79, 735–738. Miles S.H. (1993) Restraint and sudden death. Journal of the American Geriatric Society 41, 1013. Miles S.H. & Irvine P. (1992) Deaths caused by physical restraints. Gerontologist 32, 762–766. Miles S. & Parker K. (1998) Pictures of fatal bedrail entrapment. American Family Physician 58, 1755, 1759–1760. Mion L., Frengley D., Jakovcic C.A. & Marino J.A. (1989) A further exploration of the use of physical restraints in hospitalised patients. Journal of the American Geriatrics Society 37, 949–956. Moseley C.B. (1997) The impact of restraints on nursing home resident outcomes. American Journal of Medical Quality 12, 94–102. Neufeld R.R., Libow L.S., Foley W.J., Dunbar J.M., Cohen C. & Breuer B. (1999) Restraint reduction reduces serious injuries among nursing home residents. Journal of the American Geriatrics Society 47, 1202–1207. Parker K. & Miles S.H. (1997) Deaths caused by bedrails. Journal of the American Geriatric Society 45, 797–802. Retsas A.P. (1998) Survey findings describing the use of physical restraints in nursing homes in Victoria, Australia. International Journal of Nursing Studies 35, 184–191. Robbins L.J., Boyko E., Lane J., Cooper D. & Jahnigen D.W. (1987) Binding the elderly: a prospective study of the use of mechanical restraints in an acute care hospital. Journal of the American Geriatric Society 35, 290–296. Robinson B.E., Sucholeiki R. & Schocken D.D. (1993) Sudden death and resisted mechanical restraint: a case report. Journal of the American Geriatric Society 41, 424–425. Ruben B.S., Dube A.H. & Mitchell A.K. (1993) Asphyxial death due to physical restraint: a case series. Archives of Family Medicine 2, 405–408. Runciman W.B. (1999) incidents and accidents in health care: its time (Editorial). Journal of Quality in Clinical Practice 19, 1–2. Scott T.F. & Gross. J.A. (1989) Bracheal plexus injury due to vest restraints. New England Journal of Medicine 320, 598. Si M., Neufeld R.R. & Dunbar J. (1999) Removal of bedrails on a shortterm nursing home rehabilitation unit. Gerontologist 39, 611–614. Stiebeling M., Schor J., Morris J. & Lipsitz L. (1990) Morbidity of physical restraints among institutionalized elderly. Journal of the American Geriatric Society 38, 45A. Tinetti M.E., Liu W. & Ginter S.F. (1992) Mechanical restraint use and fall-related injuries among residents of skilled nursing facilities. Annals of Internal Medicine 116, 369–374. Tinetti M.E., Liu W.L., Marottoli R.A. & Ginter S.F. (1991) Mechanical restraint use among residents of skilled nursing facilities: prevalence patterns and predictors. Journal of the American Medical Association 265, 468–471. Werner P., Cohen-Mansfield J., Braun J. & Marx M.S. (1989) Physical restraints and agitation in nursing home residents. Journal of the American Geriatrics Society 37, 1122–1126. Werner P., Cohen-Mansfield J., Koroknay V. & Braun J. (1994) The impact of a restraint-reduction program on nursing home residents. Geriatric Nursing: American Journal of Care for the Aging 15, 142–146.
2003 Blackwell Publishing Ltd, Journal of Advanced Nursing, 41(3), 274–282
