Keywords:
side rails;
falls;
nursing homes;
dementia
OBJECTIVES:
Top of page
OBJECTIVES:
DESIGN:
SETTING:
PARTICIPANTS:
MEASUREMENTS:
RESULTS:
CONCLUSION:
METHODS
ANALYSES
RESULTS
DISCUSSION
ACKNOWLEDGMENTS
REFERENCES
To analyze the effect of physical restraint reduction on nighttime side rail use and to examine the relationship between bilateral side rail use and bed-related falls/injuries among nursing home residents.
DESIGN:
Top of page
OBJECTIVES:
DESIGN:
SETTING:
PARTICIPANTS:
MEASUREMENTS:
RESULTS:
CONCLUSION:
METHODS
ANALYSES
RESULTS
DISCUSSION
ACKNOWLEDGMENTS
REFERENCES
Secondary analysis of data collected in a longitudinal, prospective clinical trial designed to reduce restraint use.
SETTING:
Top of page
OBJECTIVES:
DESIGN:
SETTING:
PARTICIPANTS:
MEASUREMENTS:
RESULTS:
CONCLUSION:
METHODS
ANALYSES
RESULTS
DISCUSSION
ACKNOWLEDGMENTS
REFERENCES
Three nonprofit nursing homes.
PARTICIPANTS:
Top of page
OBJECTIVES:
DESIGN:
SETTING:
PARTICIPANTS:
MEASUREMENTS:
RESULTS:
CONCLUSION:
METHODS
ANALYSES
RESULTS
DISCUSSION
ACKNOWLEDGMENTS
REFERENCES
To examine the first question regarding the effect of physical restraint reduction on side rail usage, we included all nursing home residents who survived a 1-year data collection period (n = 463). To answer the second research question concerning the relationship between side rail status and bed-related falls, subjects' side rail status for each of the four data collection periods was compared. The sample for this analysis includes only those with consistent side rail status (n = 319) for the four observations periods: either 0/1 side rail (n = 188) or 2 (bilateral) side rails (n = 131).
MEASUREMENTS:
Top of page
OBJECTIVES:
DESIGN:
SETTING:
PARTICIPANTS:
MEASUREMENTS:
RESULTS:
CONCLUSION:
METHODS
ANALYSES
RESULTS
DISCUSSION
ACKNOWLEDGMENTS
REFERENCES
Side rail and restraint status was directly observed by two research assistants, twice each night shift (10 p.m.–6 a.m.) for three nights at each of four data collection points. Nighttime fall-related outcome data were obtained from a review of nursing home incident reports during the entire 1-year data collection period (T1 through T4). Cognitive status was measured using the Folstein Mini-Mental State Examination. Functional and behavioral status was obtained using subscales of the Psychogeriatric Dependency Rating Scale.
RESULTS:
Top of page
OBJECTIVES:
DESIGN:
SETTING:
PARTICIPANTS:
MEASUREMENTS:
RESULTS:
CONCLUSION:
METHODS
ANALYSES
RESULTS
DISCUSSION
ACKNOWLEDGMENTS
REFERENCES
Over a 1-year period, there was an increase in the proportion of bilateral side rail use for all three nursing homes. Based on the multiple logistic regression analysis, there was no indication of a decreased risk of falls or recurrent falls with bilateral side rail use, controlling for cognition and functional and behavioral status (adjusted odds ratio (AOR) = 1.13, 95% confidence interval (CI) = 0.45,2.03). Similarly, bilateral side rail use did not reduce the risk of recurrent falls, controlling for cognition and functional status (AOR = 1.25, 95% CI = 0.33,4.67).
CONCLUSION:
Top of page
OBJECTIVES:
DESIGN:
SETTING:
PARTICIPANTS:
MEASUREMENTS:
RESULTS:
CONCLUSION:
METHODS
ANALYSES
RESULTS
DISCUSSION
ACKNOWLEDGMENTS
REFERENCES
Despite high usage of bilateral side rails, they do not appear to significantly reduce the likelihood of falls, recurrent falls, or serious injuries. Bed-related falls remain clinically challenging. The data from this study, coupled with increasing reports of side rail–related injuries and deaths, compel us to seek and empirically test alternative interventions to prevent bed-related falls. J Am Geriatr Soc 50:90–96, 2002.
The most frequently used intervention to prevent bed-related falls and injuries among nursing home residents is bilateral, full-length side rails. Although no national figures exist for side rail prevalence, data from an ongoing study of side rail use in three nursing homes suggest that, in 1999, bilateral side rails were used with approximately 40% to 70% of nursing home residents.1 Despite this widespread use, a search of CINAHL (January 1982–May 2001) and Medline (January 1966–May 2001) uncovered fewer than 15 published articles about side rails.
An early descriptive study evaluating the effectiveness of side rails in fall prevention found that among 16 hospitalized adults who fell from their beds, side rails were raised in 14 (88%) of the cases.2 In numerous studies of falls and injuries, researchers have noted a significant incidence of falls and injuries with side rail use,3–11 and several editorials and reviews of the falls literature discourage their use.12–17 It is well recognized by physicians and nurses working in British geriatric facilities that side rails lack any known benefit in fall prevention,18 and their prevalence is modest, only 8.4% (56/668) in one study.19 A study conducted in a New Zealand hospital found a decease in falls and injuries after a policy to reduce side rail use.20 Thus, use of side rails has no empirical basis and may contribute to negative outcomes in frail older adults.
Side rails have been correlated with adverse events including agitation, urinary and fecal incontinence, and injuries and death from entrapment.19,21–25 In 1995, the Food and Drug Administration (FDA) issued a Safety Alert concerning hazards associated with side rail use based on 102 reports of head and body entrapment resulting in 68 deaths and 22 injuries.23 A review by Parker et al.24 of 74 deaths caused by side rails illustrates the various positions leading to death from side rail entrapment. The risk of injury and death related to side rails has also been the subject of recent investigative journalism.26–32 With increasing reports of deaths and other serious side rail–related injuries, the FDA in April 1999 convened the Hospital Bed Safety Workgroup, which includes representatives from the medical equipment industry, academia, and consumer and professional organizations to address these problems.33
In October of 1990, the Nursing Home Reform Act34 was implemented, mandating that nursing homes reduce physical restraints. Although interpretive guidelines issued by the Health Care Financing Administration (HCFA) in April 199235 identified side rails as restraints in certain situations, nursing homes in only a few states changed their practices.36–38 In 1997 and 1999, HCFA distributed Side Rails Guidelines to nursing homes. These defined side rails as a restraint if the side rail is used to prevent a resident's desired movement out of bed.39–41 Nursing home staff is required to conduct an individualized assessment and document a plan of care providing a rationale for side rail use.42
None of the more recent studies describing the relationship between falls and physical restraints has included side rails in their definition of restraint.43–49 Thus, this study was designed to analyze the effect of physical restraint reduction on nighttime side rail use and examine the relationship between bilateral side rail use and bed-related falls and injuries among nursing home residents.
METHODS
Top of page
OBJECTIVES:
DESIGN:
SETTING:
PARTICIPANTS:
MEASUREMENTS:
RESULTS:
CONCLUSION:
METHODS
ANALYSES
RESULTS
DISCUSSION
ACKNOWLEDGMENTS
REFERENCES
Design
The present study is a secondary analysis of data collected in a longitudinal prospective clinical trial designed to test the effects of three interventions to reduce physical restraint use in nursing homes: restraint education, restraint education with consultation from an advanced practice nurse, or no intervention.47 The clinical trial was conducted in three large nonprofit nursing homes between 1990 and 1992.
Subjects and Setting
To examine the first question regarding the effect of physical restraint reduction on side rail usage, we included all nursing home residents who survived a 1-year data collection period (n = 463). Data were collected four times: baseline (T1, immediately after the October 1990 HCFA mandate to reduce physical restraint use and immediately preceding the primary study's intervention), immediately after the intervention (T2), and 3 (T3) and 6 months (T4) postintervention.
To answer the second research question concerning the relationship between side rail status and bed-related falls, subject's side rail status for each of the four data collection periods was compared. The sample for this analysis includes only those with consistent nighttime side rail status (n = 319) for the four observation periods: either 0/1 side rails (n = 188) or 2 (bilateral) side rails (n = 131). Thus, 144 subjects of the 463 were excluded due to variation in side rail status (n = 80) or any use of nighttime physical restraint (n = 64) during the 1-year data collection period. The latter group was excluded because nighttime physical restraints were used to prevent bed-related falls, although such restraint use does not decrease the risk of bed-related falls.4
Measures
Nighttime Side Rail/Restraint Use
Nighttime bilateral side rails are two raised full-length (versus half or partial) side rails used between 10 p.m. and 6 a.m. The three nursing homes in this sample used full-length side rails only. Nighttime physical restraints used in bed included vest and wrist/mitt restraints. Nighttime physical restraint was defined as restraint use between 10 p.m. and 6 a.m. while lying in bed. Although some residents went to bed as early as 8 p.m. and some did not rise until 8 a.m., we observed that almost all residents were in bed between 10 p.m. and 6 a.m. Each resident's side rail and restraint status were directly observed and confirmed by two research assistants, twice per night shift (10 p.m.–6 a.m.) for three consecutive nights at each of four data collection points.40
Bed-Related Fall Outcomes
Nighttime fall–related outcome data were obtained from a review of nursing home incident reports during the entire 1-year data collection period (T1 through T4). All bed-related falls occurring between 10 p.m. and 6 a.m. were included. Time of day was recorded for each fall, along with location, to verify that the fall occurred from bed. Three bed-related fall outcomes were used: any fall, fall resulting in serious injury (all fractures, dislocated joint, subdural hematoma, or laceration requiring sutures) and recurrent falls, defined as two or more falls in the 1-year data collection period.
Demographic/Clinical Characteristics
Review of healthcare records provided demographic information. Clinical characteristics (functional, cognitive, and behavioral status) were collected three times during the 1-year data collection period: T1, T2, and T4. Trained research assistants directly interviewed resident-subjects to determine cognitive status using the Folstein Mini-Mental State Examination (MMSE).50 The MMSE has been used extensively in research as a screening instrument to determine cognitive impairment in older people. Validity measures of sensitivity ranging from .82 to .87 have been reported and specificity ranges from .80 to .82.51,52 Nursing home staff was interviewed for functional and behavioral status using subscales of the Psychogeriatric Dependency Rating Scale, which have reported interrater reliability of 0.94.53
As has been identified in other studies, including other analyses of this data set, clinical characteristics affect both fall risk and use of restraints and, therefore, may potentially confound the relationship between side rail status and bed-related fall outcomes.4,45,49 Side rails are most likely used with cognitively and physically impaired older persons.16,17,19,24,54,55 Voluntary reports to the FDA suggest that advanced age and cognitive impairment are risk factors for injuries from side rail entrapment.25 Similarly, falls by nursing home residents are strongly correlated with physical and cognitive impairment, but the relationship is not linear. Those with severe dementia56,57 and those who are immobile58 are less likely to fall. Because physical and cognitive impairment was high in this sample, we included behavioral status as a covariate. We decided a priori on a profile of three clinical characteristics (functional, cognitive, and behavioral status) that are closely associated with both side rail use and falls.
We then compared the three clinical characteristics (functional, cognitive, and behavioral status) of those with consistent bilateral side rail use to those with consistent use of 0/1 side rails. Specifically, we compared the mean score of these characteristics computed from three available data points (T1, T2, and T4) to produce a summary measure reflecting clinical characteristics over the 1-year data-collection period. This approach was validated by observing large intertime correlations for all measures, even while noting significant average declines over time. Thus, use of the mean score resulted in a reliable measure of average resident-specific characteristics during the same 1-year data collection period in which we observed residents for bed-related falls.
ANALYSES
Top of page
OBJECTIVES:
DESIGN:
SETTING:
PARTICIPANTS:
MEASUREMENTS:
RESULTS:
CONCLUSION:
METHODS
ANALYSES
RESULTS
DISCUSSION
ACKNOWLEDGMENTS
REFERENCES
Side Rail Usage
To answer the first question, repeated measures Grizzle-Starmer-Koch analyses for categorical data were used to assess the statistical significance of changes over time in the proportions of residents using bilateral side rails. This permitted estimation of linear trends and a test for differences in trends among the sites (i.e., site-by-time interactions59,60). The repeated measures analyses were necessarily confined to residents with data at T1, T2, T3, and T4 (n = 463).
Side Rail Status and Bed-Related Fall Outcomes
Tests for the statistical significance of the bivariate associations between side rail status, clinical characteristics, and bed-related falls were constructed using chi-square t statistics. Mantel-Haenszel stratified analyses were used to examine the association between bilateral side rail use and bed-related falls adjusting for one clinical characteristic (functional, cognitive, and behavioral status) at a time.61 Multiple logistic regression was used to assess the effect of bilateral side rail use on bed-related falls, controlling simultaneously for the influences of multiple clinical characteristics. A Hosmer-Lemeshow goodness-of-fit test62 was performed to test the null hypothesis that the predicted outcomes of the model adequately fit the data. This null hypothesis was not rejected (χ2 = 8.1, df = 8, P = .4254).
RESULTS
Top of page
OBJECTIVES:
DESIGN:
SETTING:
PARTICIPANTS:
MEASUREMENTS:
RESULTS:
CONCLUSION:
METHODS
ANALYSES
RESULTS
DISCUSSION
ACKNOWLEDGMENTS
REFERENCES
Side Rail Usage
T1 immediately followed the October 1990 mandate under the Omnibus Budget Reconciliation Act (OBRA 1987) to reduce physical restraints. As previously reported,4 at baseline (T1), 166 of 463 residents in three nursing homes were restrained at any time of the day. Of these restrained residents, 38.5% (64/166) were physically restrained (use of a vest or wrist restraint) in bed at night. Nighttime vest/wrist restraint use was reduced to 8.8% (11/125) of restrained residents 6 months after the intervention (T4) and 1 year after implementation of OBRA–87. Neither OBRA–87 nor the intervention specifically addressed bilateral side rail use. As demonstrated in Table 1, each of the three nursing homes increased bilateral side rail usage between T1 and 1 year later, at T4. For the total sample of 463 residents, bilateral side rail use increased from 58.7% to 64.1%.