Functional Independence Measure

Stroke:

(Beninato et al, 2006; n = 113; mean age = 63.9 (14.3) years; mean FIM score at admission = 63.4 (24.4) points, Acute Stroke)

• FIM Total Score = 22 points
• FIM Motor Subscale = 17 points
• FIM Cognitive Subscale = 3 points

Stroke:

(Inouye et al, 2001; n = 243; mean age = 64 (11) years; assessed at admission and discharge, Acute Stroke) 

• FIM scores of > 73 at admission were significantly younger (58 + 11 [SD] yr) than patients with FIM scores of 37 to 72 (64 + 11 yr) or scores < 36 (66 + 12 yr)
• FIM total scores of 37 to 72 at admission showed higher gains (37 + 15) than patients who scored > 73 (20 + 10) or < to 36 (29 + 23) 

(Tur et al, 2003; n = 102; mean age = 61.6 (10.9) yeas; 45-60 minutes of daily physical and occupational therapy, speech therapy daily as needed; Turkish sample, Acute Stroke)

Stroke:

(Hsueh et al, 2002; n = 118; mean age = 67.5 (10.9) years; measured at inpatient rehab admission and discharge, Acute Stroke) 

• Excellent internal consistency (FIM Motor Subscale) (Cronbach's alpha = 0.88 admission; 0.91 discharge)

Predictive Validity Evidence:

Stroke:

(Inouye et al, 2001; n = 243; mean age = 64 (11) years; assessed at admission and discharge, Acute Stroke) 

• Patients with FIM total scores of 37 to 72 at admission showed higher gains (37 + 15) than patients who scored > 73 (20 + 10) or < to 36 (29 + 23) 

(Denti et al. 2004; n = 359; mean age = 80.8 (4.7) years; time between stroke onset and admission = 22.3 (14.6) days, Acute Stroke) 

• FIM total scores at admission were found to be the most powerful predictor of Montebello Rehabilitation Factor Scores (Beta coefficient = 0.42)

Concurrent Validity Evidence:

Stroke:

(Hsueh et al, 2002; Acute Stroke) 

• Excellent correlation between the FIM Motor Subscale and the 10-item version of the Barthel Index (BI) (r = 0.92 (at admission) - 0.94 (at discharge))
• Excellent agreement between the FIM Motor Subscale and 5-item version of BI (r = 0.74 (at admission) - 0.94 (at discharge))

(Salter et al, 2010) 134 patients, a mean age of 68.64 (± 14.2) years old, and an average of 31.84 (± 59.2) days post-stroke, receiving care in an inpatient rehabilitation setting, were tested with the FIM at admission and discharge.

• There was excellent, positive and significant correlations with performance at admission and discharge on the FIM (total and motor) with the Clinical Outcome Variables Scale [COVS] (0.823 and 0.771 respectively).
• The COVS and FIM had excellent correlation (-0.61,-0.69)) with length of stay (P<0.01), such that lower scores at admission meant shorter length of stay.

(Ward et al, 2011) Thirty inpatients with first ischaemic stroke were evaluated with the FIM, the SIS-16 and the STREAM at admission:

• The FIM score was significantly (P<0.001) and highly correlated (excellent) with the predicted length of stay (-0.9438 ) and the actual length of stay (-0.6846)
• The validity of the FIM for predicting the LOS was higher (-0.9438) than the SIS-16 (-0.6743) and the STREAM (-0.8011)
• The validity of the FIM associated with the actual LOS was lower (-0.6846) compared to the SIS-16 (-0.7953) and the STREAM Total (-0.7972).

(Yang et al, 2013). In a prospective observational study of 122 patients with a first time stroke admitted to a rehabilitation center over a 12 month period:

• The FIM score on admission and discharge significantly predicted the Pittsburgh Rehabilitation Participation Scale [PRPS] (0.53; P<0.0001 and 0.40; P<0.001 respectively)
• The level of participation on discharge (PRPS score) was predicted by functional status on admission (FIM; 0.309), cognitive impairment (Elderly Cognitive Assessment Questionnaire-ECAQ; 0.249) and fatigue (Fatigue Severity Scale-FSS; -0.304) .
• Patients with lower levels of participation were more likely to be functionally dependent, cognitively impaired and have more fatigue.

(O’Brien et al, 2013). A sample of 371,211 Medicare beneficiaries who were receiving services in an inpatient rehabilitation facility (IRF) within 60 days post stroke (> 65 years of age, 43.7% male, 41.7% right sided impairment, 796% white) were evaluated with the FIM at admission and discharge. In addition, the change in LOS at the IRF and community discharge was compared over time with the implementation of a prospective payment system (PPS) for individuals on Medicare.

• Average LOS decreased a total of 3.8 days (from 17.9 in 2002 to 16.1 days in 2007)
• Mean admission FIM scores decreased a total of 4.4 points ( from 57.2 to 53.8 points)
• The mean discharge FIM sores decreased a total of 3.6 points ( from 80.1 to 76.5 points) in 4 of 5 years with no significant decline in 2004.
• Frequency of community discharges declined steadily with an average overall decrease of 5.4 % (from 6.6% to 61.2%) over the 5.5 years of study
• Controlling for study year and covariates, each day in IRF was associated with an increase of 0.50 discharge points (95% CI = 0.48, 0.52)
• The association between LOS and discharge destination was excellent, averaging 0.997 (95% CI = 0.994, 0.999) based on the co-variates of admission FIM, age, gender, ethnicity, side of lesion, complications and year.

(Van Heugten et al, 2015) Systematic review of studies (51) investigating convergent, criterion and predictive validity of cognitive dysfunction in patients in the acute phase (4 weeks) post stroke using multi-domain instruments.

• The Conistat, Montreal Cognitive Assessment [MOCA] and Functional Independence Measure-Cognitive showed adequate predictive validity

(Bates, 2015-Part 1) A retrospective analysis of 4020 veterans receiving consultative or comprehensive rehabilitation care post-stroke. The study examined initial characteristics of veterans predictive of grade IV achievement on the FIM.

• The final model contained the following variables: age, initial physical grade, initial cognitive stage, renal failure, nutritional compromise, type of rehabilitation services, and recovery time between admission and discharge assessments. A point system was assigned to each of the above variables, such that the clinician could enter in the above information and determine the likelihood of a patient achieving a grade IV. The area under the ROC curve was adequate of the derivation and validation cohorts (0.84 and 0.83, respectively). The Hosmer-Lemeshow statistic was not significant (ρ = 0.93).

(Bates, 2015-Part 2)

• The above model (Bates, 2015-Part1) was enhanced to become a prognostic index, predicting likelihood of recovery to or above the grade VI benchmark (Modified Independent). There was adequate fit with a nonsignificant Hosmer-Lemeshow statistic of P = 0.38 and Adequate area under the curve of 0.83 in the derivation cohort and 0.82 in the validation cohort. A similar predictive equation was derived with the sum score quartiles slightly modified.

   

• (Huang, 2010) Fifty-eight participants an average of 17.85 (range, 7-88) months post-stroke participated in distributed constraint induced therapy two hours per day, five days a week for three weeks. Assessments were administered prior and after therapy, and a Chi-squared Automatic Interaction Detector method was used to identify the strongest predictors of change on the Stroke Impact Scale. Participants with an initial Total FIM score ≤ 109 at admission, improved significantly more (P = 0.006) on the Stroke Impact Scale and on measures of activities of daily living and instrumental activities of daily living at completion of the intervention.

   

  (Lin, 2010) Seventy-four participants an average age of 54.11 (± 11.44) years old and 17.46 (± 17.67) months post-stroke were seen for upper extremity intervention. Participants received constraint-induced movement therapy, bilateral arm training, or conventional rehabilitation for two hour sessions, five times per week for three weeks. Assessments were done at baseline and post-intervention.

• Poor to excellent predictive validity was found between the domains of the Stroke Impact Scale and the FIM (0.26-0.70, p < 0.05)
• Poor to excellent predictive validity was found between the domains of the Stroke Specific Quality of Life Scale and the FIM (0.22-0.63, p < 0.01). The language, personality, thinking, and vision domains were not significant.

(Ward et al 2011) On admission to the acute rehabilitation ward, the FIM and the STREAM were found to be highly correlated in thirty patients acute post ischemic stroke.(ρ=0.7766; P<0.0001)

(Shindo et al, 2015) To explore the concurrent validity of the FIM scale with the Simple Test of Evaluation Hand Function [STEF], 34 inpatients (33-86 years of age) sub acute post stroke (less than 60 days post episode) were evaluated at admission. The STEF had statistically significant, adequate correlations with the FIM^TM: FIM Total score (0.444;P<0.009), FIM motor (0.411;P<0.016) and FIM self care (0.402; P<0.019) .

1. et al, 2014) The aim of this study was to explore the validity of the Cognitive Behavioral Rating Sale ( CBRS) with the FIM discharge data on 100 patients, mean age of 72.2 (± 10.9) years old and 61.0 (±61.2) days post-stroke. The Spearman Rank Correlation Coefficient was excellent between the CBRS and the FIM total Score (-0.70; p<0.01), the Cognitive FIM (-0.72; P<0.01), and the Motor FIM (-0.63; p<0.01) for patients post stroke.

(Caglar, 2014) A retrospective analysis on 142 patients post-stroke that went to an IRF. A linear regression was run to determine which factors contributed to Motor-FIM (M-FIM) gain and Cognitive-FIM (C-FIM) gain.

• The adjusted R² was 0.173 (p = 0.000) for M-FIM gain and the significant factors were the admission M-FIM (B = 0.809, SE = 0.199, β = -0.446, p = 0.000) and if the patient had diabetes Mellitus (B = 14.269, SE = 6.775, β = -0.177, p = 0.037).
• The adjusted R² was 0.146 (p = 0.001) for C-FIM gain and the significant factors were the admission C-FIM (B = -4.068, SE = 1.048, β = -0.369, p = 0.000) and if the patient had diabetes Mellitus (B = 36.226, SE = 17.904, β = -0.175, p = 0.045).

(Cooke, 2010) One hundred and ninty-seven, first stroke participants were included an average of 45.4 ± 67.6 days post-stroke to examine the relationship of clock drawing post-stroke.

• A significant relationship was found between the FIM-Motor and the Clock Drawing Test (Exp (B) = 0.984, p = 0.030).

Convergent Validity Evidence:

Stroke:

(Tur et al, 2003; Acute Stroke)

• Adequate correlation with length of hospital stay (r = -0.39)
• Adequate to Excellent correlation with Brunnstrom’s motor recovery stages in upper extremity, lower extremity, and hand at admission and discharge (r = 0.51 - 0.68)

Discriminate Validity Evidence:

Stroke:

(Brock et al, 2002; Rasch analysis; n = 106; mean age = 68.7 (11.3) years; median time since onset = 11 days, Acute Stroke) 

• Difficult items on motor portion of the scale discriminated better among higher functioning patients
• Raw FIM scores (as opposed to score subjected to Rasch analysis) may underestimate change

(Cavanagh et al, 2000; ischemic and hemorrhagic stroke patients, Stroke) 

• Simple 2-factor model of the FIM instrument may not be sufficient to describe disability following stroke (66% of variance) 
• May not adequately measure within patient change whereas a 3-factor model (self-care, cognition and elimination) accounted for more variance (74.2%)

(Van Heugten et al, 2015) Systematic review of studies (51) investigating convergent, criterion and predictive validity of cognitive dysfunction in patients in the acute phase (4 weeks) post stroke using multi-domain instruments .

• No instrument (including the FIM) assessed all of the commonly affected cognitive domains after a stroke
• Strong significant intercorrelations were found between the Occupational Therapy Cognitive Assessment (LOTCA), the MMSE and the FIM-Cognitive subscale

(Canbek, 2013) Fifty-five participants who experienced their first-ever stroke and went to an IRF an average of 8± 5 days post-stroke.

• Poor to Excellent construct validity was seen between the FIM-Motor and the Tinetti POMA

(Kucukdeveci, 2013) One hundred and eighty-eight community dwelling participants (mean age 63.1 ±12 years), a median of 27 (range 3-240) months post-stroke were evaluated on the FIM and the World Health Organization Disability Assessment Schedule (WHODAS-II).

• Adequate to Excellent convergent validity was found. All correlations significant at p < 0.001.

(Ottiger et al A new multidisciplinary observation scale for inpatients post stroke based on the ICF model of activity and participation was created to document outcomes post stroke (LIMOS). This scale included four components of the ICF:1). interpersonal activities, [mobility and self-care,; 2}. Communication; 3} Knowledge and general tasks; 4) domestic life. The activities were rated as limitations or restriction in domains as: none, slight, moderate, severe or complete. This new scale was correlated with FIM scores.

• Excellent convergent validity was found between the LIMOS and the FIM (r=0.89; P<0.0001)
• An excellent association was reported between the FIM mobility subscale and the LIMOS mobility subscale (r=0.90; P<0.0001)
• Adequate to excellant associations were found between the subscales of the LIMOS (self care, general tasks, domestic life) and the subscales of the FIM (r=0.36-0.79)

Stroke:

(Brock et al, 2002; Acute Stroke) 

• Minimal ceiling effect: 16% achieved ceiling on FIM Motor Subscale during inpatient rehabilitation 

(Dromerick et al, 2003; n = 95, Acute Stroke) 

• No floor or ceiling effects at either time using the FIM instrument

(Hsueh et al, 2002; Acute Stroke) 

FIM Motor subscale: 

• Minimal floor effect at admission to inpatient rehab (5.8%) and at discharge from inpatient rehab (3.5%) 
• No ceiling effect at admission to inpatient rehab (0%) and at discharge from inpatient rehab (0%)

Stroke:

(Hsueh et al, 2002; Acute Stroke) 

• Motor subscale: 
  • Large effect size with standardized response mean = 1.3

(Ward et al, 2011) A prospective cohort study of 30 subjects newly diagnosed with ischemic stroke (mean days since stroke onset 7.8 days (± 3.5)) was designed to demonstrate sensitivity of the FIM to change in an acute rehabilitation setting.

• The FIM score on admission was significantly associated (adequate to excellent correlations) with discharge destination as well as predicted and actual length of stay.
• The SRM (admission to discharge change score) was 2.34 for the motor FIM (P<0.0001). This FIM SRM was greater than the SRM for the SIS-16 and SRM for the STREAM.

(Salter et al, 2010) Following admission and discharge of 292 patients post stroke (134 with complete data and 158 with incomplete data, respectively an average of 31.8 and 67.3 days post stroke), FIM^TM scores improved significantly (P<0001) from admission to discharge from a mean of 73.86 (24.13) to 95.70 (24.65) . The SRM was 1.36.

SCI:

(Hall et al, 1999; cross-sectional data from SCI Model Systems National Database; average of 8 days post injury [SD = 13 days]; sample size varying pending time post injury, Acute SCI)

Divide the score by 13 (i.e. 13 motor items) to obtain the average ratings on the 1 to 7 scale

Divide the score by 5 (i.e. 5 cognitive items) to obtain the average ratings on the 1 to 7 scale

*All cases with level and completeness data available; These are not all the same sample of individuals across admission and discharge

(Kay et al, 2010; n = 1780; discharged from one of 479 inpatient rehab facilities in US; age 65-74 years; diagnosed with incomplete paraplegia, Acute SCI)

SCI:

(Grey and Kennedy, 1993; n = 40; mean age at time of injury = 29.6 (9.57) years; mean time post-injury at discharge = 24.75 (8.57) weeks, Chronic SCI) 

• Excellent correlation between total FIM scores taken by clinician discharge report and self-report at one month (r = 0.828) 
• Poor to Excellent correlation between FIM subscales scores taken by clinician discharge reort and self-report at one month: 
  • Self care: r = 0.841 (Excellent)
  • Sphincter control: r = 0.710 (Adequate) 
  • Mobility: r = 0.733 (Adequate) 
  • Locomotion: r = 0.454 (Adequate) 
  • Communication: r = 0.029 (Poor) 
  • Social cognition: r = 0.085 (Poor) 

(Karamehmetoglu et al, 1997; n = 50; mean age = 33.94; 22% with tetraplegia and 78% with paraplegia, SCI) 

• Excellent intrarater correlation of FIM scores obtained by questioning the patient and by observation of patient performing the activity (r = 0.94) 

(Kucukdeveci et al, 2001; FIM in Turkey; n = 62; mean age = 32.7; mean time since injury = 16.4 months; with cervical injury 21%; with thoracic injury 42%; with lumbar 37%, Chronic SCI) 

• Excellent FIM Motor interrater reliability (ICC = 0.90) 
• Excellent FIM Cognitive interrater reliability (ICC = 0.98) 

(Segal et al, 1993, n = 57, discharging from acute care and admitting to rehab hospital; data collected within a max of 6 days, Subacute SCI) 

• Excellent interrater reliability for total FIM scores across two settings (r = 0.83) 
• Poor to Excellent interrater reliability for individual items (r = 0.02 - 0.77) 
• Excellent interrater reliability for patients with complete quadriplegia (n = 14, r = 0.87), complete paraplegia (n = 13, r = 0.74), and incomplete paraplegia (n = 9, r = 0.85) 
• Adequate interrater reliability for patients with incomplete quadriplegia (n = 17, r = 0.49)

SCI:

(Kucukdeveci et al, 2001; FIM instrument version in Turkey, Chronic SCI)

• Excellent internal consistency at admission and discharge for FIM Motor (Cronbach’s alpha = 0.934 - 0.953) and FIM Cognitive (Cronbach’s alpha = 0.930 - 0.983) 

(Stineman et al, 1996; with nontraumatic SCI, n = 2,609, mean age = 64.6 years; with traumatic SCI, n = 1,831, mean age = 43.0 years, sample from Uniformed Data System for Medical Rehabilitation [UDS_MR^SM], SCI) 

• Excellent internal consistency for nontraumatic spinal cord diagnosis (Cronbach’s alpha for total = 0.91; for FIM Motor = 0.91; for FIM Cognitive = 0.90) 
• Excellent internal consistency for traumatic spinal cord diagnosis (Cronbach’s alpha for FIM Total Score = 0.92; for FIM Motor = 0.94; for FIM Cognitive = 0.90) 

SCI:

(Ditunno, et al., 2007; n = 141, mean age = 32 years; Entered into study within 8 weeks of onset of SCI; data taken at entry, 3 and 6 and 12 months, subjects required to have score of < 4 on the Locomotor FIM (LFIM) at entry, Acute SCI) 

(Donnelly et al, 2004; n = 41; mean age = 49(118.1); mean time since injury = 52 (73.1) days; with paraplegia, n = 18; with tetraplegia, n = 20; Incomplete, n = 27; complete, n = 11, SCI) 

• Adequate correlation between admission and discharge scores of the FIM Total Score and the Canadian Occupational Performance Measure (COPM) Performance (r = 0.388 - 0.452) and COPM Satisfaction (r = 0.513 - 0.514) 
• Adequate correlation between change scores of the FIM Total Score and FIM motor with COPM Performance (r = 0.364, r = 0.351) and Satisfaction (r = 0.497, r = 0.497) from admission to discharge 

(Fujiwara et al, 1999; n = 14; C6 level of injury, mean age = 30.7 years; mean length of time from injury = 462.0 days, Chronic SCI) 

• Excellent correlation of FIM motor score and AIS motor score (Spearman’s rank correlation coefficient = 0.73) 
• Excellent correlation of shoulder strength (sum of MMT for serratus anterior, upper pectoralis major, and latissiums dorsi) and FIM motor score (Spearman’s rank correlation coefficient = 0.95) 
• Excellent correlation of AIS shoulder strength score (deltoid) and FIM transfer score (Spearman’s r = 0.93) 

(Saboe et al, 1997; n = 160; mean age = 30 (13) years; assessed at admission, discharge, and 2 years post injury; Length of stay at tertiary care hospital 144 (111) days Chronic SCI) 

• Excellent correlation of FIM score 2 years post injury with admission and discharge ASIA motor (Spearman’s r = 0.68 - 0.80), ASIA light touch (Spearman’s r = 0.75 - 0.76), ASIA pinprick (Spearman’s r= 0.73 - 0.76), and Computed Vibration (Spearman’s r = 0.64 - 0.67) 
• Adequate correlation of FIM score 2 years post injury with admission bony injury level (Spearman’s r = 0.53) and admission and discharge ASIA Impairment (Spearman’s r = 0.50 - 0.53) 
• 56% of the variance of FIM scores 2 years post injury is accounted for with ASIA admission light touch scores with age being the next largest contributing factor 

(Yavuz et al, 1998; n = 29; mean age = 37 years; mean time between onset and rehab admission = 20 weeks, mean length of stay in inpatient rehab = 18 weeks, Subacute SCI) 

• Excellent correlation of FIM score with ASIA motor (r = 0.91) 
• Adequate correlation of FIM score with ASIA light touch (r = 0.58) and ASIA pinprick (r = 0.55) 
• Excellent correlation of Quadriplegia Index of Function and FIM (r = 0.97)

The FIM instrument was based on the results of a literature review of published and unpublished measures as well as input provided by an expert panel. Face and content validity were determined using subject matter experts (Granger, Hamilton, Keith, Zielezny, & Sherwins, 1986). 

Content validity was established through a pilot study done at 11 centers (n = 110 patients evaluated; Keith & Granger, 1987).

SCI:

(Jackson et al, 2008; n = 54 expert raters assessed locomotion measures as: 1) valid or useful, 2) useful but requires validation or changes/improvements, or 3) not useful or valid for research in SCI, SCI) 

• FIM – Locomotion item was rated as Valid/Useful by 6%, Useful But Requires Validation or Changes by36% , and Not Useful or Valid for Research in SCI by 58%

SCI:

(Grey and Kennedy, 1993; Chronic SCI) 

• Face validity was evaluated by asking clinicians specific questions addressing: 
  • Difficulty of understanding (88% had no difficulty)
  • Unnecessary items (97% reported no unnecessary items
  • Items that should be added (83% felt no extra items needed)

SCI:

(Grey and Kennedy, 1993; Chronic SCI) 

• 92% of subjects and 88% of clinicians reported a max score on communication 
• 75% of subjects and 73% of clinicians reported a max score on social cognition 

(Hall et al, 1999; Acute SCI) 

SCI:

(Spooren et al, 2006; n = 60; mean age = 38.9 years old; first measurement taken when subjects were first able to sit up in a chair for 3 hours, Acute SCI) 

• Large effect size for all subjects regardless of AIS classification between initial measurement (T1) and 3 months later (T2) as well as between initial measurement (T1) and discharge from rehab (T3) 
• Small to moderate effect size for subjects between T2 and T3 (ES = 0.37-0.79)

TBI:

(Donaghy & Wass, 1998; TBI) 

• Excellent interrater reliability (ICC = 0.85 for total FIM Scores, 0.92 for FIM Motor, and 0.69 for FIM Cognitive) 

(Heinemann et al, 1994; Rehabilitation Patients)

• Admission FIM Motor Scores accounted for 52% of variance in discharge motor function among TBI patients, admission FIM Cognitive Scores accounted for 46% of variance in discharge cognitive function – admission motor FIM was the most significant predictor of length of stay

(Montecchi et al, 2013) In 59 patients with mean age of 48.90 (± 14.01) years old, admitted to the intensive care unit acutely post acquired brain injury (from trauma, hypoxia, haemorrhage or ischemia), a new Trunk Recovery Scale (TRS) was developed.

• Excellent correlation between the FIM-Motor and the TRS (0.849)

The FIM instrument was based on the results of a literature review of published and unpublished measures as well as input provided by an expert panel. Face and content validity were determined using subject matter experts (Granger, Hamilton, Keith, Zielezny, & Sherwins, 1986). 

Content validity was established through a pilot study done at 11 centers (n = 110 patients evaluated; Keith & Granger, 1987).

Traumatic Brain Injury:

(Hall et al, 2001; TBI)

• Although the FIM instrument is reliable and key validity characteristics have been established, it has only 5 items directly addressing cognitive, behavioral, and communication issues, which limits its content validity for TBI

(Ellis et al, 2008; n = 68; mean age - 74 (8) years; H&Y stages II - V, number in each stage: II - 1, III - 18, IV - 37, V - 2)

(Marciniak et al, 2011; n = 89; mean age = 74.26 (9.38) years)

Elderly Adults: 

(Pollak et al 1996; n = 49 elderly residents of a continuing care retirement community; mean age 89.7 years; assessed twice 3 to 8 days apart, Elderly Adults) 

• Excellent FIM Motor test-retest reliability (ICC = 0.90) 
• Excellent FIM Cognitive test-retest reliability (ICC = 0.80) scores

(Hobart et al, 2001; Elderly Adults) 

• Excellent test-retest reliability (ICC = 0.98 for total FIM, 0.95 and 0.89 for FIM  Motor and FIM Cognitive, respectively)

Orthopedic Diagnoses and Stroke:

(Kohler et al, 2009; n = 143 patients (63% orthopedic and 13% stroke); mean age = 76 years; transferred and assessed from one Rehab unit to another; 1 to 3 days between assessments, Orthopedic Diagnoses and Stroke) 

• Adequate to Poor item-level interrater reliability (ICC = 0.124 to 0.661) 
• Poor agreement on 4 items:
  • Stairs 
  • Dressing 
  • Walking 
  • Bowel management

Various Diagnoses (meta analytic findings):

(Ottenbacher et al, 1996; n = 11 studies published between 1993 and 1995; total sample size = 1,568 participants, Various Diagnoses) 

• Excellent overall consistency (median interrater reliability = 0.95) between raters across patients with different diagnosis and levels of impairment

Multiple Sclerosis:

(Sharrack et al, 1999; n = 64; mean age = 40 years, MS)

• Excellent internal consistency (Cronbach's alpha = 0.98)

General Rehab: 

(Dodds et al, 1993; n = 11,102 (52% Stroke, 10% Orthopedic; 10% Brain Injury); mean age = 65 years, General Rehab) 

Neurological Disorders:

(Hobart et al, 2001; Neurological Disorders) 

• Excellent internal consistency (Cronbach's alpha = 0.95 FIM Total Score; 0.95 FIM Motor; 0.89 FIM Cognitive)

Predictive Validity Evidence:

Neurologic Disorders: 

(Ng, et al., 2007; n= 1502; mean age of total = 61.3 ± 15.0 years; mean acute LOS = 14.5 ± 17.5 days; mean inpatient rehab LOS = 21.5 ±19.0 days, Neurological Disorders) 

Discriminate Validity Evidence:

Rehabilitative patients:

(Hobart et al, 2001; n = 169; neurological rehab patient: MS, stroke, TBI, other)

• FIM total and FIM motor scores correlated more strongly with OPCS disability scores, LHS scores, SF-36 physical component scores and WAIS – verbal IQ, than with measures of mental health status or psychological distress (SF36 mental component, General Health Questionnaire)
• FIM Cognitive Scores correlated most strongly with OPCS Disability scores and WAIS-verbal IQ scores and weakly with LHS, SF-36 physical and mental components, and the General Health Questionnaire (ABIEBR)

Rehabilitation Patients:

(Coster et al, 2006; n = 516 subjects with neurologic, orthopedic, or complex medical conditions; mean age = 68.3 (14.97) years; discharged from tertiary care or rehab hospital, Rehabilitation Patients)

• Ceiling effect on FIM motor scale after discharge ranging from 10% at 1 month to 15% at 12 months
• Ceiling effect on the FIM cognitive scale after discharge for 70% of subjects at 1 month, reducing to 53% at 12 months

Rehabilitation Patients:

(Coster et al, 2006; Rehabilitation Patients)