Physical restraint deaths in a 13-year national cohort of nursing home residents

Abstract

Objective

this paper aims to investigate the nature and extent of physical restraint deaths reported to Coroners in Australia over a 13-year period.

Methods

the study comprised a retrospective cohort study of residents dwelling in accredited nursing homes in Australia whose deaths were reported to the Coroners between 1 July 2000 and 30 June 2013 and was attributed to physical restraint.

Results

five deaths in nursing home residents due to physical restraint were reported in Australia over a 13-year period. The median age of residents was 83 years; all residents had impaired mobility and had restraints applied for falls prevention. Neck compression and entrapment by the restraints was the mechanism of harm in all cases, resulting in restraint asphyxia and mechanical asphyxia, respectively.

Conclusions

this national study confirms that the use of physical restraint does cause fatalities, although rare. Further research is still needed to identify which alternatives strategies to restraint are most effective, and to examine the reporting system for physical restraint-related deaths.

Physical restraint, Nursing homes, Death, Older people

Introduction

Use of physical restraints on nursing home residents is contentious. Historically, physical restraints were applied to improve safety by reducing falls and preventing wandering or aggressive behaviour [1]. Contemporary evidence suggests this reasoning is fallacious [2, 3], with little difference in falls rates between restrained and non-restrained populations [4].

Physical restraint use is concerning due to the potential for physical harm, including death [5]; psychological harm [6]; and infringement of human rights [7, 8]. Over the last decade, a ‘restraint free’ model of care in nursing homes worldwide has been promoted [9–11], however, physical restraint remains overused. The prevalence of physical restraint use varies greatly across countries, from 12% to 49% in Australia [12], over 31% in Canada [13] and up to 37% in the US [14]. Unfortunately, there is no literature reporting the prevalence of restraint use in care homes in the UK [15]. The frequency of deaths from restraints appears to be rarely collated, and if so, it is not often publically available, making it almost impossible for any genuine international comparisons.

In Australia, ‘nursing homes’ provide permanent and respite residential aged care to people who can no longer be supported living in the community [16]. They are similar to ‘care homes with nursing’ in the UK. Residents are predominantly female, aged over 80 years, and 60% have a diagnosis of dementia [17]. Australian nursing homes are regulated by the Australian Government's Aged Care Quality Agency [18]. There are four Accreditation Standards which nursing homes must comply with to qualify for government funding. Although there is growing support to reduce physical restraint use [11], no standards govern the use of restraint in nursing homes. However, the use of physical restraint has been identified by the National Aged Care Quality Indicator Program [19] as an intervention that if reduced will contribute to better quality of care and an improved quality of life for consumers.

This paper aims to investigate the nature and extent of physical restraint deaths of nursing home residents reported to Australian Coroners.

Method

The study comprised a retrospective cohort study of residents dwelling in accredited nursing homes in Australia, whose deaths were reported to Coroners between 1 July 2000 and 30 June 2013 and attributed to physical restraint.

Data from the National Coronial Information System [20] was examined and where a death met the selection criteria (Box 1), the following information was extracted: socio-demographic characteristics; duration of residence; medical history; incident findings; mechanism of death; formal reports attached and Coroners’ recommendations. A descriptive statistical analysis was conducted on the included cases.

Box 1.

Population and nursing home statistics, case identification and inclusion criteria

Australian population characteristics

The population of Australia in 2000 was 19.2 million, with 2.4 million persons over the age of 65 years. In 2013 these figures rose to 23.1 million and 3.3 million, respectively.

Nursing homes statistics

There were 3005 nursing homes in Australia in the year 2000, comprising 135,991 residents. In 2013, 2718 nursing homes housed 173,094 residents.

The NCIS contained 22,204 deaths of persons residing in nursing homes, occurring between 1 July 2000 and 30 June 2013.

Cases

The National Coronial Information System (NCIS) is an internet-based data storage and retrieval system of deaths reported to Australian Coroners since 1 July 2000 (1 January 2001 for Queensland) [20]. A word search of external cause deaths (n = 3,289); deaths from asphyxia and aspiration (n = 23) and deaths from complications of clinical care (n = 39) within the NCIS identified 58 cases where the word ‘restraint’ was included in the cause of death or text-based documents contained within the NCIS (Coroner's findings, police summary of circumstances, autopsy report and toxicology report).

A manual review of the text-based documents was then undertaken for the 58 cases of interest to confirm the contribution of physical restraint to the death. Five cases were identified that fulfilled the inclusion criteria.

Inclusion and exclusion criteria

Cases were included if: death occurred between 1 July 2000 and 30 June 2013; the Coroner's investigation was completed by 31 December 2014; the deceased was a resident of an ACSAA accredited nursing home; the death occurred within the nursing home setting; the death was directly caused by the physical restraint.

Cases were excluded if: physical restraint was used in the residents care but was not involved in the death; there was insufficient information to determine the mechanism and cause of death; the incident did not occur within the nursing home.

In the 53 excluded cases, physical restraints were utilised as an intervention in the residents’ care; however, no adverse outcomes or contribution to the residents’ deaths were reported. They types of physical restraints used in these cases were primarily lap belts and bed rails.

Results

Of the 58 cases identified for detailed review, five were deaths of nursing home residents due to physical restraint (Table 1).

Table 1.

Case findings

Case	1	2	3	4	5
Gender	F	M	F	M	M
Age (years)	86	56	86	83	79
Organic disorders	D	AD	D	AD; PD	Nil
Medical history	NIDDM; stroke; HTN; osteoporosis; OA; depression; GORD	Down Syndrome; Epilepsy; CVA; feeding tube	Anxiety	Coronary artery atherosclerosis	Stroke; HTN; renal impairment
Duration of residence (months)	8	7	NR	24	2 days
Resident status	P	P	P	P	R
Falls risk	H; HoF	H; HoF	H; HoF	H; HoF	H
Mobility	A	A	A	IM	W
Year of incident	2012	2000	2000	2000	2009
NH location	Metro NSW	Metro VIC	Metro VIC	Metro VIC	Metro QLD
Incident day of week	W	W	Th	Th	M
Incident time	20:00–22:00	05:00–06:20	16:00	07:00–07:45	01:30–02:00
Type of restraint	B	C	B, TC	C	BR
Restraint application	C	C	I	C	C
Reason for restraint	F	F	F, W	F	F
Restraint authorisation	FA	FA, MP	NR	FA	NR
Previously restrained	Y	Y	Y	Y	Y
Mechanism of harm	Slipped down in chair with restraint tightly around neck.	Fell out of bed and became caught on bedside netting with neck wedged between bed rail and steal bedhead.	Slipped down in chair under the table with lap belt tightly around throat.	Fell out of bed and became trapped in mesh/webbing of cotside face down with head protruding through the mesh and neck caught on the end of the bed railing.	Fell out of bed with body over the side of the bed rail and head and shoulders face down on the floor.
Time last seen to discovery (h)	2	4	0.25	1.5	0.5
Time of death	22:20	08:00	16:15	07:45	02:00
Cause of death (autopsy)	NC	MA	NC	MA	MA
Reports attached	A, T, P	A, T, P, C	A, T, P, C	A, P, C	A, P, C
Inquest held	N	N	Y	Y	N
Recommendations	N	N	N	N	N

NR, not recorded.

Gender: F, female; M, male.

Organic disorders: D, dementia; AD, Alzheimer's disease; PD, Parkinson's disease.

Resident status: P, permanent; R, respite.

Falls risk: H, high; HoF, history of falls.

Mobility: IM, immobile; W, wheelchair dependent; A, walks with assistance (person or aid).

Incident day of week: M, Monday; W, Wednesday; Th, Thursday.

Type of restraint: B, seat belt/lap belt; C, cotsides with webbing; BR, bed rails; TC, table over chair.

Restraint application: C, correct; I, improvised.

Reason for restraint: F, falls; W, wandering.

Restraint authorisation: FA, family; MP, medical practitioner; NR, not recorded.

Previously restrained: Y, yes.

Cause of death by autopsy: NC, neck compression; MA, mechanical asphyxia.

Reports attached: A, autopsy findings; T, toxicology; P, police report; C, Coroner's findings.

Inquest held: Y, yes; N, no.

Recommendations made: N, no.

Demographics

Age and gender

The age of residents ranged from 56 to 86 years; the median age was 83 years (IQR = 18.5). Three subjects were male.

Nursing home location and residential status

All five deaths occurred in metropolitan regions. Four occurred amongst individuals residing as permanent residents for seven months or longer. One resident was a temporary resident admitted two days prior to the incident for respite care.

Medical history

All five residents had multiple co-morbidities (Table 1); and four diagnosed with dementia.

Falls risk and mobility

All residents had impaired mobility. The resident who was immobile had not walked for 18 months prior to the incident. Residents were considered high falls risks; four had a documented history of repeated falls.

Incident and death details

Timing of incident

All incidents occurred on weekdays (Table 1). The time the resident was last seen alive was documented and ranged from 15 min to 4 h.

Type and application of physical restraint

Types of restraints included seat belts, bed rails and cotsides.¹ In four cases the Coroner deemed the use of the restraint appropriate. In Case-3 the Coroner concluded the restraint was inappropriate because staff had improvised a belt from a lifting machine. This belt was too short and was fastened across the resident's legs instead of behind the chair.

Restraint use and authorisation

Physical restraints were applied predominantly for falls prevention. The residents’ family had agreed with use of restraint in three cases. Two residents had documented evidence of restraint-free care trials. On review, restraint-free care was deemed high risk in both residents and restraints were reintroduced.

Mechanism of harm and cause of death

The mechanism of harm and cause of death were ascertained by a forensic pathologist following autopsy and formulated as ‘neck compression and entrapment by the restraints’ in all cases. Neck compression from slipping down in their chair resulting in the lap restraint applying direct pressure on the neck occurred in two residents. Entrapment occurred in three residents where the residents fell out of bed becoming entrapped in the cotside webbing or between the bed rails, resulting in mechanical asphyxia.

Forensic medical and scientific reports and Coroners’ recommendations

Toxicological investigation occurred in three of the deaths revealing prescription medications, predominantly antipsychotics, used at appropriate therapeutic levels.

The Coroners’ did not make recommendations in any of the five cases.

Discussion

Five nursing home residents died from physical restraints over a 13-year period in Australia. This confirms that the use of physical restraint does cause fatalities, though this is rare.

Interpretation

Risk of falling was the most common indication for physical restraint use in these cases. Nursing staff have heightened concern about harm from falls, as one in two residents will experience falls in a 12-month period [21]. The perception that restraining residents prevents falls is misplaced as despite the use of restraint, residents still fall [22] from chairs or beds resulting in death (Table 1).

Four of the five residents who died from physical restraint had diagnosed dementia. This is consistent with the observation that residents with dementia are three times more likely to be physically restrained [23, 24]. There is an urgent need to address this issue, as the number of people with dementia is estimated at 225 million globally by 2050 [25], over half residing in nursing homes [19]. If no changes occur in practice, the use of physical restraints will increase. Nursing homes must focus on creating a dementia-enabling environment and utilising alternatives to restraint [24]. Behavioural symptoms exhibited by persons with dementia must cue staff to consider unmet needs of the resident [24] and not simply trigger restraint.

Coroners’ did not make recommendations in these five deaths. In contrast, three deaths from physical restraint investigated in Canada [26] lead to recommendations about: employing a policy of least restraint; regular review of restraint policies; exploration of alternatives to restraint and staff education about risks of restraint and viable alternatives.

Strengths and limitations

This national study is the first of its kind and provides population estimates. These are an underestimate of physical restraint-related deaths. Berzlanovich et al.’s study in Germany had a substantially greater number of deaths suggesting significant under-reporting in Australia. Under-reporting may be due to a reflexive emotional response to conceal the event [27] or failure of providers to provide guidance on reporting requirements [28]. In Victoria, Australia over 12 years (2000–12) we estimate approximately 4% of deaths occurring in nursing homes are reported to the Coroner, accounting for 7% (4,027/56,855) of all deaths reported [29].

Study design limitations include the retrospective evaluation of prospectively collected data creates a bias due to missing or misclassification of causes of death [30]. The small number of cases makes findings difficult to generalise.

Implications and conclusion

Although there are few deaths occurring due to physical restraint, the seriousness of the outcome, and questionable efficacy of restraint use for falls prevention, should prompt policy and practice reform towards a restraint free model of care. A potential significant policy reform in Australia is introducing ‘incidence of physical restraint use’ as a voluntary quality indicator for nursing homes. Further research is still needed to identify effective alternatives to restraint and to examine the reporting system around physical restraint-related deaths.

Key points

• Historically, use of physical restraints was advocated on the assumption of improving resident safety.

• There were five deaths of nursing home residents due to physical restraint reported to Coroners in Australia over a 13-year period.

• All deaths were from neck compression and entrapment by the restraints. Median age of residents was 83 years; all residents had impaired mobility and, had restraints applied for falls prevention.

• This number of fatalities is likely to be an underestimate.

• A ‘restraint free’ model of care in nursing homes should be promoted.

Conflict of interest

The authors have no potential financial or personal interests that may constitute a source of bias.

Funding

Funding support for this work was provided by Commonwealth Department of Social Services; Department of Health and Human Services, Ageing and Aged Care Branch, Victoria, Australia and the Department of Forensic Medicine, Monash University. None of the funders influenced the design, methods, subject recruitment, data collections, analysis and preparation of paper.

Ethics approval

The Victorian Institute of Forensic Medicine Research Advisory Committee granted approval for this study on 14 April 2015 (RAC 011/13) and the Department of Justice Human Research Ethics Committee granted ethics approval for access to the NCIS on 12 May 2015 (CF/13/8187).

Footnotes

1

Cotsides were considered different to bed rails as they included a mesh webbing that acted as a net.