Restraint use in inpatient rehabilitation: Incidence, predictors, and implications

427

Restraint Use in Inpatient Rehabilitation: Incidence, Predictors, and Implications Rana’al E. Schleenbaker, MD, Susan M. McDowell, MD, Robert W. Moore, PhD, Julia F. Costich, PhD, Gloria Prater, RN ABSTRACT. Schleenbaker RR, McDowell SM, Moore RW, Costich JF, Prater G. Restraint use in patient rehabilitation: incidence, predictors, and implications. Arch Phys Med Rehabil 1994;75:427-30. a The use of mechanical restraints in rehabilitation facilities focuses attention on the conflict between patient safety and independent physical function. To evaluate restraint use, we reviewed records of 323 inpatient rehabilitation admissions. Restraint orders were written for 78.3% of admissions, but used in only 32.2% of cases (mean duration of use was 16 days). Posey vests were most commonly used (78.2%). Reasons for restraint were previous fall (26.8%), impulsivity (23.7%), and inappropriate self-transfers (19.6%). Male sex, decreased mental status, low admission functional independence measure (FIM) score, stroke, or traumatic brain injury were closely associated with restraint use. Falls occurred in 25 % of restrained and 10.1% of unrestrained patients. Conclusions are as follows: (1) although physician orders are required to apply restraints, nursing staff initiate, monitor, and discontinue restraint use independently; (2) traumatic brain injury or stroke, decreased admission mental status, lower FIM scores, and male sex are indicators of restraint use; (3) age is not associated with restraint use; and (4) restraints may not prevent falls. 0 1994 by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation

Restraint use in various patient care settings has generated significant controversy over the last 10 years.le5 Concerns for self-determination and right to privacy have prompted legal limitations and guidelines that have addressed restraint use in psychiatric and skilled nursing facilities.3s6 Federal regulations require documentation of the medical symptoms that necessitate use of restraints.’ The efficacy of restraints, however, is in question.” Justifications for the use of restraints include prevention of falls,’ maintenance of body alignment,’ and protection of other patients and staff.* Conversely, some observers suggest that restraint use increases injury from falls,‘“~‘lpromotes functional decline,‘* and contributes to skin breakdown,6 as well as to deleterious psychological4 and physiological’3 changes. Furthermore, the use of restraints may be affected by the fear of litigation6 arising from either failure to ensure patient safety or injury attributed to restraint use. A rehabilitation hospital is a unique health care environment where the value conflicts between patient safety and autonomy are evident. The programmed activities that promote and develop patient independence increase the potenFrom the Cardinal Hill RehabiIitation Hospital (Dr. Schleenbaker, Ms. Prater), Department of Physical Medicine and Rehabilitation (Drs. Schleenbaker, McDowell, Costich), Stroke Program, Center on Aging (Dr. Moore), College of Medicine University of Kentucky, Lexington, KY. Submitted for publication May 12, 1993. Accepted in revised form September 28, 1993. No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit upon the authors or upon any organization with which the authors are associated. An abstract summarizing portions of this work appeared in the October 1992 Archives of Physical Medicine & Rehabilitation. It was the subject of an oral presentation at the annual meeting of the American Academy of Physical Medicine & Rehabilitation in San Francisco. Reprint requests to Randal E. Schleenbaker, MD, Department of Rehabilitation Medicine, Kentucky Clinic, University of Kentucky Medical Center, Lexington, KY 40536.0284. 0 1994 by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation 0003-9993/94/7504-0118$3.00/O

tial for injury, particularly from falls. Rehabilitation professionals face the dilemma of promoting independence while ensuring patient safety. The use of mechanical restraints in the rehabilitation setting brings into focus the problems of patient safety, independent function, and the psychological effects of being tied down. A review of the literature on mechanical restraints found no studies of restraint use in rehabilitation hospitals. Mion5 studied restraint use in a geriatric population from general medical and rehabilitation wards in an acute care hospital. He found that 34% of rehabilitation patients were restrained as opposed to 13% of the general medicine patients; he suggested that the reason more rehabilitation ward patients were restrained was because of the greater number of patients with compromised balance and cognition resulting from stroke. Because there has been no research on restraints in a rehabilitation hospital setting, we did a retrospective chart review to: describe the prevalence and incidence of restraint use among patients in a rehabilitation hospital; describe the patterns of restraint use, including type and duration of use once initiated; identify the reasons for using restraints; and to identify patient characteristics associated with restraint use. METHODS The continuous sample consisted of patients admitted to a 90-bed rehabilitation hospital from September through December, 1991. Of the 329 patients admitted, 6 (1.8%) were excluded from this analysis because of missing data, leaving a sample size of 323. Data were abstracted from the admission history and physical, initial nursing assessment, physician orders, team notes, incident reports, and the bedside nursing chart. Data were obtained on patient characteristics (age, gender, diagnosis), admission status (functional independence measure [FIM], Arch Phys Med Rehabil Vol75, April 1994

RESTRAINT USE IN INPATIENT REHABILITATION, Schleenbaker

428

Table 2: Admission FIM Scores

Table 1: Characteristics of the Patients N

Percent

179 144

55.4 44.6

95 110 118

29.4 34.1 36.5

131 61 31 16 19 65

40.6 18.9 9.6 5.0 5.9 20.1

109 19 31 10 10 8

33.7 5.9 9.6 3.1 3.1 2.5

Sex Male Female Age 45 or younger 46 to 70 71 or younger Diagnosis Stroke Closed head injury Spinal cord injury Hip fracture Amputation Other* Mental status deficits Not oriented X 3 Unable to follow simple commands Confused Agitated Uncooperative Restless * Parkinson’s disease, Guillain-Barr6 matoid arthritis, multiple fractures.

Syndrome,

multiple

sclerosis,

rheu-

mental status), restraint orders, occurrence and type of restraint use, justification for initiating restraint use, and the occurrence of falls. The overall FIM14 score was obtained along with the six subscores in the areas of self-care, bowel and bladder control, transfer skills, locomotion, communication, and social interaction and cognitive skills. In the absence of consistent Mini-Mental Status scores in patient records, we used six items (alert and oriented times three, able to follow simple commands, agitation, restlessness, confusion, uncooperative) to indicated mental status on admission. This mental status assessment score was derived by abstracting from the medical record these six observations that are routinely documented by physicians and nurses as part of our admission process. From the admission physical, data was abstracted about the patient’s orientation (person, place, and time); the patients’ ability to follow simple commands was also abstracted. From the initial nursing evaluation, data was abstracted regarding the patients’ agitation, restlessness, confusion, and uncooperativeness. Patients were given a score of one for each indicator if they were not oriented times three, could not follow simple commands, or were agitated, restless, confused, or uncooperative. The indicators were summed with a possible score range from 0 (no mental deficits) to 6 (a deficit on each indicator). The resulting scale from 0 to 6 was calculated for each patient to give some indication of the mental status of the patient. The six admission diagnoses-stroke, closed head injury, spinal cord injury (SCI), hip fracture, and other (Parkinson disease, Guillain-Barr6 Syndrome, multiple sclerosis, rheumatoid arthritis, and multiple fracture)-were coded as dummy variables for multiple reyession analysis. Descriptive statistics, t test, x , and multiple regression were used to analyze the data. FINDINGS Table 1 shows the characteristics of the patient population studied. Men accounted for 55% of the subjects. The average Arch Phys Med Rehabil Vol75,

April 1994

Self-care Bowel & bladder Transfer Locomotion Communication Social/cognitive Overall

control

Mean

SD

Minimum

Maximum

24.8 6.8 10.2 4.7 11.7 15.8 74.0

9.7 3.7 4.7 2.4 3.7 5.1 23.3

6 2 3 2 2 3 18

42 14 21 14 14 21 124

N = 323.

age was 58 years, (SD = 21.6) with a range of from 1 to 89 years. Stroke was the primary diagnosis for 40.6% of the cases, and closed head injury accounted for 18.9%. Together, these two types of acute brain injury made up more than half (59.4%) of the study subjects. More than one third (38.4%) of the patients had one or more deficits in mental status. The average of the mental status score was 5.42 (SD = 0.95). Table 2 shows the means, standard deviations, and ranges of the FIM subscores. The overall initial FIM score averaged 74.0 (SD = 23.3) with a range from 18 to 124. Table 3 shows means and standard deviations of admission FIM scores by diagnosis. Subjects with traumatic brain injury (TBI) (62.7) and stroke (68.7) had the lowest overall FIM scores. Combined, these patients with damage to the brain had average scores of 67.0 on the admission FIM, compared with an average of 84.3 for subjects with other diagnoses (t = 7.05, df = 321 p < .OOOl). Restraints were ordered for 78.3% of the patients admitted. Restraints were used on 32.2% (n = 104) of patients at some time during their rehabilitation hospitalization (40.7% of those for whom restraints were ordered). All restraint orders were written “restraint as needed.” Of the restraints used, 78.2% were Posey or waist restraints, 18.3% were Posey restraints with pelvic strap and wrist restraint or mitts, and 3.5% were Posey restraints with pelvic strap. Patients with stroke (94.7%), TBI (90.2%), and hip fracture (75.0%) were most likely to have restraint orders written. Patients with TBI (63.9%) and stroke (45.8%) were the most likely to have restraints actually used. Table 3: Restraint Use by Diagnosis

All Traumatic brain injury Stroke Spinal cord injury Hip fracture Amputation Other7 F Probability RZ

Admission FIM Mean (SD)

Ordered (%)*

74.0 (23.3)

78.3

32.2

5.4 (9.8)

14.9

62.3 (27.8) 69.1 (17.7)

90.2 94.7

63.9 45.8

11.6 (14.3) 7.3 (10.2)

24.6 16.0

0.0

0.0 (0.0) 3.4 (7.9) 0.0 (0.0) 0.4 (2.9) 13.9 .0001 ,180

81.5 (24.6) 71.1 (24.2) 94.2 (19.7) 85.8 (19.2) 13.16 .OOOl ,172

Restraints

51.6 75.0 68.4 50.8 17.20 .OOOl ,213

* Coded as 0 = no, 1 = yes. t Parkinson’s disease, Guillain-Barr6 matoid arthritis, multiple fractures.

Mean (SD) Days of use (%I

used (BY

25.0 0.0 1.5 25.01 .OOOl ,283

Syndrome,

multiple

sclerosis,

Who Fell (so)*

12.9 0.0 5.3 10.8 1.99 NS .030 rheu-

RESTRAINT

USE IN INPATIENT

Table 4: Multiple Regression Analysis of Restraint and Days of Restraint Use Restrain Use (0 = no, 1 = yes) Independent Variables Intercept Sex (0 = female. 1 = male) Age Mental status Admission FIM TBI* CVA’ SW HIP” AMP”

Estimate

F

1 134 0.127 0.000 -0.127 -0.005 0.310 0.319 -0.078 0.170 -0.027

Days of Restraint p

Estimate

F

Use p

6.96

.OOOl

23.16

6.61

.OOOl

3.07 0.41 4.71 4.7 1 4.22 5.47 0.98 1.66 0.29

.0023 .6840 .OOOl .OOOl .OOOl .OOOl .3296 .0971 .7730

2.49 -0.007 -1.992 -0.147 4.636 3.810 - 1.752 1.083 -0.011

2.57 0.28 3.16 6.17 2.70 2.79 0.94 0.45 0.01

.0106 .7821 .0017 .OOOl .0073 .OO55 .3500 .6503 .9958

R’ = ,436 * Traumatic

Use

R2 = ,350

brain injury

’ Cerebral vascular accident. * Spinal cord injury. I’ Hip fractures. p Amputation, coded as 0 = not present,

1 = present.

The average number of days of restraint use for all patients was 5.4 days (SD = 9.8), with a range from 0 to 56 days. For the 104 patients actually restrained the average number of days of use was 16.8 days and 29 (30.5%) were restrained at the time of discharge. The reason for restraint use was documented in the initial nursing assessment for 97 (93.4%) of the cases in which restraints were used. The reasons were inappropriate self-transfer (19.6%), impulsive behavior (23.7%), agitation (8.2%), previous fall (26.8%), decreased cognition (11.3%), confusion (4.1%), and safety risk (6.2%). We noted that 48 patients (14.9%) suffered falls during their hospitalization. Falls occurred in 25% of patients who used restraints compared with only 10.1% of those who did not use restraints (x2 = 12.464, df = 1, p < .OOl). The average duration of restraint use for patients who fell was 11.7 days, compared with 4.2 days for those who did not fall (t test = 3.02, p = <.Ol). Table 4 shows the results of multiple regression of restraint use (coded as 0 = no, 1 = yes) and number of days of restraint use on sex, age, mental status, admission FIM score, and five dummy variables (0 = not present, 1 = present) for diagnoses of TBI, stroke, XI, hip fracture, and amputation. For both dependent variables the patient’s sex, mental status score, admission FIM score, presence of TBI, and presence of stroke are statistically significant independent predictors. The model accounts for 43.6% of the variation in restraint use and 35% of the variation in days of restraint use. Men were restrained more than women, as were patients with impaired mental status, lower admission FIM scores, and diagnoses of TBI or stroke. DISCUSSION Thirty-three percent of the patients in this study were restrained at some point during their hospitalization. This rate is comparable to the 34% incidence reported by Mion et al.’ However, care must be taken in making comparisons because there are significant age differences between our

REHABILITATION,

Schleenbaker

429

patient population and that studied by Mion. In Mion’s population the mean age was 70 years (SD = 11.5), whereas the mean age in our population was 58 years (SD = 21.6). Mion’s explanation of the higher rate in restraint use in rehabilitation patients compared with that found in the acute hospital setting is the cognitive and balance problems found in the elderly stroke patients. Our data suggest that the strongest predictors of restraint use are sex, mental status, and functional status, which are associated with closed head injury in the young and stroke in the elderly. The average duration of restraint use in our sample was 16.8 days. This finding refutes the notion of “once restrained, always restrained.” Although our study was not designed to evaluate the temporal order of applying and removing restraints, chart review indicated that the removal of restraints occurred in 69.5% of cases during the hospitalization. Seventy-eight percent of the restraints used were waist restraints, suggesting that the level of restraint required was minimal. Restraints were used on only 40.7% of the patients for whom they were ordered. All patients on the stroke, closed head injury, and general rehabilitation units received restraint as needed orders on admission, written by the physician. Restraints were applied only after the nursing staff completed the safety portion of their assessment and documented their finding that the patient was unsafe. Our review of the charts indicated that the decision to apply and remove restraints was a nursing decision 95.8% of the time. In only one case was the use of restraints identified as a team decision, whereas in three cases the decision to restrain a patient was specifically the physician’s. The patients most likely to be restrained were those in the diagnostic categories of TBI or stroke with poor mental assessment scores and low functional independence ratings. In our study there was no intrinsic age factor associated with restraint use. This finding contradicts the strong association found between age and restraint use in the studies done in skilled nursing facilities. One explanation is that both stroke and TBI are associated with low FIM scores and poor mental assessment scores, and these factors correlate significantly with restraint use. The bimodal age distribution of these two diagnoses dilutes the effect of age on restraint use. Conversely, the association of age with restraint use in skilled nursing facilities may be caused by as yet unidentified covariables. The independent association of male gender with restraint use, raises questions of nursing perceptions in regard to restraint and gender. It may be that large patients are restrained more than small patients, and that men, being larger, will be restrained more. It is also plausible that the gender differences found in restraint use are because of sex role stereotypes that lead to different staff expectations from male and female patients. The association of restraint use with falls is most easily explained by assuming that patients were restrained as a result of their falls or that the patients most likely to fall are those most likely to be restrained. The former explanation seems unlikely because 36% of those who used restraints and fell were using their restraints at the time of the fall. If the latter explanation is true, the efficacy of restraints for Arch Phys Med Rehabil Vol75,

April 1994

RESTRAINT USE IN INPATIENT REHABILITATION, Schleenbaker

fall reduction in the rehabilitation setting is in question. The retrospective manner in which this data was collected precludes further analysis of this question. However, our findings suggest that it is worthwhile to consider the question of whether restraint use might be reduced to the benefit of patients in the acute rehabilitation hospital if staff members were more selective in the cases and circumstances in which they applied restraints. There are several limitations of this study that should be considered. First, the retrospective design limits the specificity of our data. We are unable to answer questions regarding the temporal course of restraint use, the specific variables conditioning the initiation or discontinuation of restraint use and the specifics of the relationship between restraint use and falls. Second, we did not take into account the effect medication may have on restraint use. The type and dosage of medication used by the patient could have an independent effect on the profile of restraint use, and the possibility must be taken into account when interpreting our results. Third, our mental assessment score was derived using observations that are routinely documented in the medical record by clinical staff as part of the admission process. However, the assessment score is not a standardized instrument, it’s validity and reliability have not been established. CONCLUSIONS The pattern of restraint use in a rehabilitation hospital is different from published accounts of restraint use in acute care hospitals or nursing homes. Sex, mental status, Admission Functional Independence Measure score, and diagnosis of TBI or stroke are independent predictors of restraint use. Restraint application and removal seems to be a nursing

Arch Phys Mod Rehabil Vol75,

April 1994

decision rather than one made by the team or the physician. The relationship between restraint use and falls requires further evaluation. References 1. Rader J, Donius M. Leveling off restraints. Geriatr Nurs 1991;12: 71-3. 2. Masters R, Marks SF. The use of restraints. Rehabil Nurs 1990;15: 22-5. 3. Evans LK, Strumpf NE. Tying down the elderly. J Am Geriatr Sot 1989;37:65-73. 4. Robbins IJ, Boyko E, Lane J, Cooper D, Jahnigen DW. Binding the elderlv: a orosnective studv of the use of mechanical restraints in an acute carehospital. J Am Geriatr Sot 1987; 35:290-6. 5. Mion LC, Frengley JD, Jakovcic CA, Marino JA. A further exploration of the use of physical restraints in hospitalized patients. J Am Geriatr Sot 1989;37:949-56. 6. Johnson SH. The fear of liability and the use of restraints in nursing homes. Law Med Health Care 1990; 18:263-73. 7. Food and Drug Administration. Revocation of exemptions from the 510(k) premarket notification procedures and current good manufacturing practice regulations. Federal Register June 19,1992;57(119):27397401. 8. Cape R. Freedom from restraints [abstract]. Gerontologist 1983;23:217. 9. Strumpf NE, Evans LK. Physical restraint of the hospitalized elderly: perceptions of patients and nurses. Nurs Res 1988;37:132. 10. Jancken JK, Reynolds BA, Swiech K. Patient falls in the acute care setting: identifying risk factors. Nurs Res 1986;35:215-9. 11. Dube AH, Mitchell EK. Accidental strangulation from vest restraints. JAMA 1986;256:2725-6. 12. Warshaw GA, Moore J, Friedman J, et al. Functional disability in the hospitalized elderly. JAMA 1982; 248:847-50. 13. Miller M. Iatrogenic and nursigenic effects of prolonged immobilization of the ill aged. J Am Geriatr Sot 1975;23:360-9. 14. Hamilton BB, Granger CV, Sherwin FS, Zielezny M. A uniform national data system for medical rehabilitation. In: Fuhrer MJ, ed. Rehabilitation outcomes: analysis and measurement Baltimore: Paul H. Brooks, 1983:137-47.