Comparison of Fallers and Nonfallers at an Inpatient Rehabilitation Facility: A Retrospective Review


  • Frances Kwan PT DPT OCS,

  • Stephanie Kaplan PT DPT ATP,

  • Mary Hudson-McKinney PT MS DPT NCS,

  • Donna Redman-Bentley PT PhD,

  • Emily R. Rosario BS MS PhD


Patients in an inpatient rehabilitation facility (IRF) are at increased fall risk. However, little IRF research has focused on fall risk.


The purpose of this study was to retrospectively examine differences between 35 patients who fell and 35 who did not during their IRF stay .


The following admission data were compared: age, gender, diagnosis, Morse Fall Scale score, and 18 Functional Independence Measure (FIM) scores. Independent t-tests were conducted for age and FIM scores, Mann–Whitney test was conducted for Morse scores, and chi-square tests were conducted for gender and diagnosis to examine differences between fallers and nonfallers.


There were no significant differences between groups for age, gender, diagnosis, or Morse scores. However, there were significant differences for 12 FIM items, and FIM Motor, Cognitive, and Total scores.


The results suggest that the Morse Scale may not be the most appropriate tool for assessing fall risk in an IRF.

Conclusions and clinical relevance

Decisions about fall risk should consider admission FIM scores.


Falls in a hospital setting present a significant risk for injury and potential financial burden. Compared with patients in the acute hospital setting, those admitted to an inpatient rehabilitation facility (IRF) are at equal or greater risk of falling. Studies indicate acute hospital fall rates of 3–6 falls per 1000 patient days (Ang, Mordiffi, Wong, Devi & Evans, 2007; Tutuarima, van der Meulen, de Haan, van Straten & Limburg, 1997), whereas fall rates in IRFs vary from 2.92 to 17.8 falls per 1000 patient days (Gilewski, Roberts, Hirata & Riggs, 2007; Mayo, Korner-Bitensky, Becker & Georges, 1989; Nyberg & Gustafson, 1996). In the IRF setting, 10%–50% of patients will fall at least once during their stay, and the rate of injury from falls vary from 9% to 33% of falls (Aberg, Lundin-Olsson & Rosendahl, 2009).

According to the Centers for Disease Control (CDC), unintentional falls continue to be the leading cause of nonfatal injuries in older adults. Thirty-three percent of adults aged 65 and above fall each year (CDC, 2009a,2009b – Overview). Of those who fall, 20–30% suffer moderate to severe injuries that make it difficult for them to live independently and increases their chances of early death (CDC, 2009a,2009b – Overview). In addition, from 1994 to 2003, death rates related to falls increased significantly for persons aged 65 and above (CDC, 2009a,2009b – Overview).

Falls are responsible for 70% of hospital patient accidents (Sutton, Standen & Wallace, 1994); 30% of these lead to injury (Ash, MacLeod & Clark, 1998). Risk of hip fracture is 11 times higher in the hospital setting compared with the community (Papaioannou et al., 2004).

Falls have financial implications as well. Annual costs related to falls were $19 billion in 2000, and is projected to be greater than $50 billion in 2020 (CDC, 2009a,2009b – Cost of Falls). The average health care cost for one fall for a person aged 72 or above is $19,440 (CDC, 2009a,2009b – Cost of Falls). If a patient falls during a hospital stay, Centers for Medicare and Medicaid Services will no longer pay for injuries related to the fall. Most probably, the hospital will need to absorb the costs. Patients who fall often need longer hospital stays, which increase the expenses (Aberg et al., 2009). In addition, these costs do not include long-term consequences of falls, such as increased dependence on others, lost time from work, and reduced quality of life.

Little research was found on fall risk and fall assessment tools in the general IRF setting. Instead, current research has focused on acute care, skilled nursing facilities, stroke-specific rehabilitation settings, or with community-dwelling older adults. Scott, Votova, Scanlan and Close (2007) reviewed the literature for studies between 1980 and 2004, where the primary or secondary purpose was to test the predictive value of one or more fall assessment tools. Thirty-eight tools were evaluated in settings such as community dwelling, acute, and long-term care (nursing homes). In this article, the “acute” setting included acute care, emergency departments, or geriatric inpatient or outpatient rehabilitation. The authors found 12 studies and 8 tools utilized in the acute setting, one of which was the Morse Fall Scale. The Downton index was assessed for the rehabilitation setting, but only for patients diagnosed with a stroke . No assessment tools were used to predict falls in a general IRF, which admitted more than one diagnosis. Scott et al. (2007) also reported that none of the tools demonstrated consistently strong predictive values across two or more settings, and only two studies, Functional Reach and STRATIFY, showed good predictive validity in repeated studies in one setting. Perell et al. (2001) systematically reviewed the literature from 1984 to 2000, reviewing 20 fall assessment tools for content and validity. Tools were categorized into acute, outpatient, or extended care with no specific differentiation for inpatient rehabilitation. These authors stated that most patients in the extended care setting, including nursing home and rehabilitation units, were at high fall risk. Therefore, they believed it to be more useful to institute a fall prevention program for all patients, but commented that further research was needed to determine the effectiveness. However, having universal fall precautions for all patients may not be realistic or effective in reducing the number of falls.

At the facility of interest, nursing staff utilized the Morse Fall Scale to assess fall risk for each patient within 24 hours of admission. The Morse Fall Scale was developed in 1989 using data from 100 patients who fell and 100 randomly selected patients who did not fall at a primarily acute care facility that also contained a 50-bed long-term geriatric center and a 140-bed Veteran's Home (Morse, 1996). In light of the number and frequency of patient falls, the facility wanted to critically re-assess the effectiveness of the fall prevention program and identification of patients at high fall risk.

The purpose of this study was to retrospectively evaluate and discriminate between patients who did and did not fall during their stay in a 68-bed IRF, examining demographic variables, Functional Independence Measure (FIM) scores, and Morse scores.



This study included 35 patients (18 males, 17 females) who had fallen and 35 patients (17 males, 18 females) who had not fallen during their 2007 stay at a 68-bed IRF located in Southern California. Patients ranged in age from 13 to 93 years (mean = 65.0, SD = 19.7) and presented with diagnoses that included stroke, traumatic brain injury, spinal cord injury, or orthopedic surgeries. (See Table 2 for demographic data of fallers and non-fallers.) In this study, a fall was defined as unintentionally coming to rest on the ground, floor, or other lower level (Gilewski et al., 2007). Falls that were attended (in the presence of another person) and unattended (patient was found on the ground or fall reported by the patient) were included. Of the 108 patients who fell during 2007, 35 patients were randomly selected to serve as the fallers group, and of the 1194 patients who experienced no falls, 35 patients were randomly selected to serve as the nonfallers group. Both groups were chosen using a random numbers table.


The following data were collected through retrospective chart review for each patient: age at admission, gender, diagnosis, Morse Fall Scale score, and 18 admission FIM scores. FIM Motor (sum of motor items), Cognitive (sum of cognitive items), and Total (sum of Motor and Cognitive) scores were also calculated. Staff, including nursing and therapists, gathered the original data between 24 and 72 hours of admission.

The Morse Fall Scale identified six variables of interest (Figure 1). Each variable is assigned a numerical score, the sum of which is the total score ranging from 0 to 125. Patients who scored 25 or higher on the Morse were categorized as high fall risk. Interrater reliability was reported as = .96 (Morse, 1996). To validate the scale, it was prospectively tested using combined data from an acute care hospital, a long-term care (a psychogeriatric unit and a nursing home), and a rehabilitation hospital (Morse, 1996).

Figure 1.

Morse Fall Scale. Adapted from Preventing Patient Falls, by Morse (1996). Copyright 1996 by Sage Publications

The FIM is the most widely used measure of disability in rehabilitation hospitals (Table 1) and has demonstrated high reliability. Median interrater reliability was .95 for total FIM and ranged from .78 for Social Cognition to .95 for Self-Care for the six FIM subscales (Ottenbacher, Hsu, Granger & Fiedler, 1996). Each of the 18 items is assigned a FIM level based on the amount of assistance required to perform basic activities (0–7). Total FIM score is the sum of Motor and Cognitive FIM scores (0–126).

Table 1. Functional Independence Measure (FIM)


  1. FIM consisting of motor and cognitive domains. Each item is scored a FIM level from 0 to 7. Adapted from “The reliability of the functional independence measure: A quantitative review” by Ottenbacher et al., 1996. Copyright (1996) by Elsevier.

Motor domain
A. Eating
B. Grooming
C. Bathing
D. Dressing – upper body
E. Dressing – lower body
F. Toileting
Sphincter control
G. Bladder management
H. Bowel management
I. Bed, chair, wheelchair
J. Toilet
K. Tub, shower
L. Walk/wheelchair
M. Stairs
Cognitive domain
N. Comprehension
O. Expression
Social cognition
P. Social interaction
Q. Problem solving
R. Memory
FIM Levels of scoring
Independence (no helper)
7 Complete independence (timely, safely)
6 Modified independence (device)
Modified dependence (helper)
5 Supervision (subject=100%)
4 Minimal assistance (subject=75% or more)
3 Moderate assistance (subject=50% or more)
Complete dependence (helper)
2 Maximal assistance (subject=25% or more)
1 Total assistance (subject <25%)
0 Activity does not occur; Use this code only at admission

Statistical analyses were chosen in consultation with a statistician and calculated using SPSS version 17.0 (Armonk, NY). To examine differences in age and FIM scores between fallers and nonfallers, independent t-tests were calculated. Non-parametric analyses were used to assess for differences in the Morse Fall Scale scores (Mann–Whitney), gender, diagnosis, and number of participants rated as high fall risk (chi-square tests). When applicable, data were presented as mean with SD or as the number of patients and percent per group.


As stated above, this study included 35 patients who had fallen and 35 patients who had not fallen during their 2007 IRF stay. There were 41 fall events with four patients experiencing two falls each and one patient experiencing three falls. Nine patients (26%) from the fallers group reported negative consequences (pain, laceration, skin break, or tear). Falls were categorized as unattended, attended/assisted by staff, attended by nonstaff, and reported by patient/family. The most common type of fall was unattended with 61% of fall events (Figure 2).

Figure 2.

Type of Fall

Demographic variables, the Morse Fall Scale, and FIM scores were evaluated to assess differences between fallers and nonfallers. Mean (SD) ages for fallers and nonfallers were 63.3 (20.4) and 66.7 (19.1) years, respectively, with no significant difference between groups (Table 2). There were no significant gender differences between groups (Table 2). The groups also differed in the number of orthopedic diagnoses, but the difference was not statistically significant (Figure 3). Mann–Whitney tests found no significant differences between groups for each of the Morse Fall Scale items and the Morse Fall Scale total score (Table 3). The Morse rated 86% of fallers and 91% of nonfallers as high fall risk, based on the cut-off score of 25. Significant differences were found between fallers and nonfallers when comparing individual FIM scores including eating, grooming, bathing, dressing-upper body, toileting, transfers to the bed, chair or wheelchair, toileting transfers, stairs, comprehension, expression, social cognition, problem solving, and combined scores for motor, cognitive, and cumulative total, all p < .05 (Table 4).

Table 2. Mean (SD) or Number (%) of Demographic Characteristics for Fallers and Nonfallers
CharacteristicFallersNonfallersp value
Mean age (years)63.3 (20.4)66.7 (19.1).48
Number of females17 (48.6)18 (51.4).81
Morse high fall risk30 (85.7)32 (91.4).45
Stroke8 (22.9)6 (17.1).34
Brain injury3 (8.6)1 (2.9).34
Spinal cord injury6 (17.1)3 (8.6).34
Orthopedic6 (17.1)13 (37.1).34
Amputee0 (0)2 (5.7).34
Neurologic3 (8.6)1 (2.9).34
Pulmonary2 (5.7)2 (5.7).34
Other7 (20.0)7 (20.0).34
Figure 3.

Differentiation by diagnosis. BI, brain injury; SCI, spinal cord injury; Ortho, orthopedic diagnosis; Neuro, neurologic disorders not including Stroke, BI, or SCI

Table 3. Mean Rank for Morse scores for Fallers and Nonfallers
VariableFallersNonfallersp value
Morse 1: History of falling34.037.0.45
Morse 2: Secondary diagnosis36.534.5.62
Morse 3: Ambulatory aid34.536.5.65
Morse 4: IV/Heparin lock36.035.0.81
Morse 5: Gait34.037.0.50
Morse 6: Mental status35.535.51.00
Morse Total Score33.337.7.37
Table 4. Difference in Mean (SD) for FIM scores between Fallers and Nonfallers
VariableFallersNonfallersp95% CI


  1. *CI, confidence intervals; w/c, wheelchair.

  2. *p < .05.

FIM A: Eating4.20 (1.80)4.94 (1.14).04*[−1.46, −0.02]
FIM B: Grooming3.46 (1.36)4.43 (1.04).001*[−1.55, −0.39]
FIM C: Bathing2.63 (1.00)3.17 (1.12).04*[−1.05, −0.04]
FIM D: Dressing – Upper3.03 (1.22)3.83 (1.30).01*[−1.40, −0.20]
FIM E: Dressing – Lower2.00 (1.11)2.51 (1.38).09[−1.11, 0.08]
FIM F: Toileting2.20 (1.26)3.00 (1.55).02*[−1.47, −0.13]
FIM G: Bladder management2.29 (1.45)2.34 (1.31).86[−0.71, 0.60]
FIM H: Bowel management2.86 (1.24)3.00 (1.09).61[−0.70, 0.41]
FIM I: Transfers – Bed, chair, w/c2.71 (1.13)3.31 (1.26).04*[−1.17, −0.03]
FIM J: Transfers – Toilet2.31 (1.30)3.11 (1.35).01*[−1.43, −0.17]
FIM K: Transfers – Tub, shower1.74 (1.27)2.09 (1.48).30[−1.00, 0.32]
FIM L: Walk/Wheelchair1.89 (1.23)2.17 (1.45).38[−0.93, 0.36]
FIM M: Stairs1.11 (0.58)1.69 (1.35).03*[−1.07, −0.07]
FIM N: Comprehension5.14 (1.50)5.91 (1.25).02*[−1.43, −0.11]
FIM O: Expression5.09 (1.62)5.86 (1.31).03*[−1.47, −0.07]
FIM P: Social interaction4.63 (1.78)5.77 (1.50).005*[−1.93, −0.36]
FIM Q: Problem solving4.29 (1.76)5.11 (1.51).04*[−1.61, −0.05]
FIM R: Memory4.43 (1.82)5.14 (1.50).08[−1.51, 0.08]
FIM Motor total31.46 (11.87)37.51 (11.38).03*[−11.60, −0.51]
FIM Cognitive total23.43 (7.81)27.49 (6.68).02*[−7.52, −0.59]
FIM Cumulative total54.89 (15.55)65.00 (13.31).005*[−17.02, −3.21]


In this study, significant differences in admission FIM scores between fallers and nonfallers in an IRF setting were identified. Interestingly, no significant differences were found between fallers and nonfallers on the Morse Fall Scale. These findings are supported by other work, which suggests a strong correlation between admission FIM scores and risk for falling. In their prospective study, Saverino, Benevolo, Ottonello, Zsirai and Sessarego (2006) reported that fallers had lower motor, cognitive, and total FIM scores at admission when compared with nonfallers. Gilewski et al. (2007), Maeda, Kato and Shimada (2009), and Teasell, McRae, Foley and Bhardwaj (2002) also found that fallers in IRFs had lower total FIM scores at admission. However, another study cited that tests of executive functioning, and not cognitive FIM scores, accounted for the variance in predicting falls in an IRF (Rapport, Hanks, Millis & Deshpande, 1998). One study indicated a direct, inverse relationship between number of falls and admission FIM score (Teasell et al., 2002).

Similar to this study, the Morse Fall Scale alone has been reported to identify 75%–90% patients as high fall risk (Gilewski et al., 2007). Morse (1996) suggested that a facility determine its cut-off score for high fall risk, ranging from 25 to 45 in her original text. This study used 25 as the cut-off score. However, if the cut-off score was increased to 45, 19 (54.3%) fallers and 24 (68.6%) nonfallers were rated as high fall risk. Again, no significant differences in the number of patients rated as high fall risk were found between the two groups (p = .22).

Like the Morse Fall Scale, other fall assessment tools include items such as mental status, mobility, incontinence, age, use of sedatives, and history of falls (Hendrich, Bender & Nyhuis, 2003; Krauss et al., 2005; Oliver, Britton, Seed, Martin & Hopper, 1997; Papaioannou et al., 2004). However, these results suggest that what may be more important in the IRF setting are specific cognitive deficits and assistance needed for toileting and activities of daily living using the FIM items. In this study, gait FIM scores were not different between groups. Most patients are admitted to an IRF, because they have deficits in mobility, such as ambulation, justifying the admission. On the other hand, patients who struggle with the motor planning to perform a transfer, the fine motor coordination to groom or eat, or demonstrate difficulty with sitting balance to dress the upper body may be at a higher risk for falls. Those with communication and social cognitive deficits can also be at a high fall risk.

Key Practice Points

  • Based on previous studies, patients admitted to an inpatient rehabilitation facility (IRF) are at equal or greater risk of falling compared to those in the acute hospital setting.
  • Current measurement tools used to assess fall risk may be inadequate for a general inpatient rehabilitation facility, which admits multiple diagnoses.
  • Based on the results of this study, it may be more important to evaluate the patients physical and mental functioning at admission as a predictor of fall risk than age or co-existing medical conditions.
  • Fall risk assessment tools that can be administered quickly by nursing staff shortly after admission can be most effective in identifying those at high fall risk.

Nursing Implications

For fall risk assessment and subsequent prevention to be effective, the risk assessment needs to occur immediately upon admission. Nurses are key team members to evaluate the patient's fall risk. Nurses see the patients first and can evaluate and complete a full 24-hour analysis of the patient's fall risk and whether that risk changes over the course of the day. In addition to being familiar with the FIM, they are also the staff members who assist with dressing, feeding, and transfers and communication with patients throughout the day. This study suggests that it may be more important to evaluate the patient's physical and mental functioning as a predictor of fall risk than age or co-existing medical conditions. Tools that can be administered quickly by nursing staff shortly after admission can be most effective in identifying those at fall risk.

Furthermore, fall risk assessment tools lose their external validity when utilized in a setting that is different from where it was originally developed (Oliver et al., 2008) and are less effective when used outside the original facility even when the setting and patients are similar (Perell et al., 2001). Perell et al. (2001) suggested that different settings should use different assessment tools. As mentioned previously, the Morse was developed using data from a primarily acute hospital. However, in this study, the Morse Fall Scale was used as a fall assessment tool in an IRF. Future directions should include the development and validation of a fall assessment tool specifically for the IRF setting.

Limitations of this study include small sample size and a retrospective design in which one can only discriminate, but not predict fallers. In addition, a variety of staff completed the Morse Fall Scale and FIM scores. Inconsistency in discriminating among the different options for each Morse variable (i.e., difference between weak and impaired gait) may have also occurred. The age of the participants ranged from 13 to 93 years with a mean age of 65.0 years. Therefore, the results may not be applicable to all IRF patients. Despite these limitations, this work is important, as little research has been done to examine fall risk and assessment tools in general IRFs.


Identification of high fall risk characteristics is essential to a fall prevention program. Based on the findings of this study, the Morse Fall Scale may not be the most appropriate tool for assessing high fall risk in a general IRF, as it rated 91% of nonfallers in this study as high fall risk. In addition, decisions about potential fall risk should consider the level of assistance needed to perform daily activities and utilize admission FIM scores.

IRB approval for this study was granted on September 16, 2008.

Funding was provided by Casa Colina Centers for Rehabilitation.

Earn nursing contact hours

Rehabilitation Nursing is pleased to offer readers the opportunity to earn nursing contact hours for its continuing education articles by taking a posttest through the ARN website. The posttest consists of questions based on this article, plus several assessment questions (e.g., how long did it take you to read the article and complete the posttest?). A passing score on the posttest and completion of the assessment questions yield one nursing contact hour for each article.

To earn contact hours, go to and select the “Education” page. There you can read the article again, or go directly to the posttest assessment by selecting “RNJ online CE.” The cost for credit is $10 per article. You will be asked for a credit card or online payment service number.

Contact hours for this activity will not be available after August 31, 2013.

The Association of Rehabilitation Nurses is accredited as a provider of continuing nursing education by the American Nurses Credentialing Center's Commission on Accreditation (ANCC-COA).


  • Frances Kwan, PT DPT OCS, is a physical therapist, at the Casa Colina Centers for Rehabilitation, 255 East Bonita Avenue, Pomona, CA 91769. Address correspondence to

  • Stephanie Kaplan, PT DPT ATP, and Emily R. Rosario, BS MS PhD, are at the Casa Colina Centers for Rehabilitation, Pomona, CA.

  • Mary Hudson-McKinney, PT MS DPT NCS, and Donna Redman-Bentley, PT PhD, are at the Western University of Health Sciences, Pomona, CA.