Inter-rater Reliability: In a review of 11 studies, Ottenbacher et al. (1996) (varying etiologies) reported a mean inter-observer reliability value of 0.95, a median test-retest reliability of 0.95, and a median equivalence reliability (across versions) of 0.92. Reliability was higher for items in the motor domain than for those in the social/cognitive domain with ICC=0.98 for total FIM, 0.95 for motor FIM, and 0.89 for cognitive FIM (Hobart et al. 2001) (varying etiologies). Donaghy & Wass (1998)(TBI) found ICC=0.85 for total FIM, 0.92 for motor FIM, and 0.69 for cognitive FIM.

Internal Consistency: Cronbach α of 0.93-0.95 was reported for admission versus discharge (Dodds et al., 1993) (varying etiologies) and α=0.88 to 0.91 (Hsueh et al., 2002) (stroke). Hobart et al. (2001) reported item-to-total correlations ranging from 0.53 to 0.87 for FIM total, 0.60 for FIM motor, and 0.63 for FIM cognitive. FIM-mean inter-item correlations were 0.51 for FIM, 0.56-0.91 for motor FIM, and 0.72-0.80 for cognitive FIM, with Cronbach α=0.95, 0.95 and 0.89 for FIM, motor FIM and cognitive FIM respectively.

Construct Validity: Linacre et al. (1994) reported 2 distinct aspects of disability within FIM-motor and cognitive function. However, Cavanagh et al. (2000) (stroke) suggested that the simple 2-factor model of the FIM not be sufficient to describe disability following stroke (66% of variance) and may not adequately measure within patient change whereas a 3-factor model (self-care, cognition and elimination) accounted for more variance (74.2%). Use of Rasch transformed scores for comparison of level of ability at the end of treatment show the motor FIM to be a discriminative, ordinal, outcome measure of disability (Brock et al., 2002; Linacre et al., 1994).

Construct Validity (Known Groups): FIM scores discriminated between groups based on spinal cord injury severity (p<0.05), presence of comorbid illness (p<0.005), as well as right or left –sided involvement in stroke patients both at admission (p<0.005) and discharge (p<0.05). Most of this score difference occurred on the communication domain (Dodds et al., 1993). On admission and discharge, FIM scores discriminated between groups with or without neglect (p<0.001, p<0.02) and with or without aphasia (p<0.01, p<0.09, stroke) (Ring et al., 1997).

Concurrent Validity: Motor-FIM showed strong concurrent validity in association with BI and Spearman’s correlation coefficient ranging from 0.74 (admission) to 0.92 (discharge) (Hsueh et al., 2002). Kwon et al. (2004) (stroke) reported r=0.95 between motor-FIM and BI scores, and 0.89 between motor-FIM and Modified Rankin Scores. FIM motor scores and cognition scores were significantly correlated with DRS ratings (r=0.641 and 0.728 respectively, p<0.05) and FIM cognition scores correlated with LCFS scores (r=0.645, p<0.05) (Hall et al., 1993).

Convergent/Discriminant Validity: FIM total and motor FIM scores correlated more strongly with Office of Population Census and Survey (OPCS) disability scores, London Handicap Scale (LHS) scores, MOS 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). However, cognitive FIM correlated most strongly with OPCS Disability scores and WAIS-verbal IQ scores and weakly with LHS, SF36 physical and mental components, and the General Health Questionnaire (Hobart et al., 2001).

Predictive Validity: FIM admission score was predictive of placement after discharge (Oczkowski & Barreca, 1993) (stroke); (Dodds et al., 1993), while FIM scores and length of stay was predictive of functional gain (p<0.0002) (Ring et al., 1997). Granger et al (1993) (stroke) reported FIM predictive of burden of care assessed in help in minutes/day (p=0.01). Singh et al. (2000) (stroke) reported FIM scores at 1 month post stroke predictive of depression at 3 months post stroke as part of a predictive model that also included “living at home” and “damage to inferior frontal region”. FIM scores at admission to rehabilitation were significantly associated with employment status one year post head injury (Cifu et al., 1997) (TBI). Admission motor FIM accounted for 52% of variance in discharge motor function among TBI patients, and admission cognitive FIM scores accounted for 46% of variance in discharge cognitive function. Admission motor FIM was the most significant predictor of length of stay (Heinemann et al., 1994).

Changes in FIM scores from admission to discharge were in the expected direction (p<0.0005) (Dodds et al. 1993). Significant differences in FIM total, FIM motor and FIM cognition scores were reported between rehab discharge and follow-up one year post injury (p<0.0001 for all). Change between 1 and 2 years, as well as between 1 and 5 years was distributed across all items with most change in cognitive function (Hammond et al., 2001) (TBI).

When ceiling effect is defined as the top 29% of the scale (scoring ≥ 108), 49% of TBI patients scored in this range at rehabilitation discharge, and 84% by year one post injury (K. Hall et al., 1996b). 4% of patients obtained maximum FIM scores ((McPherson & Pentland, 1997) (TBI). Neither floor nor ceiling effects were reported at admission or discharge from rehabilitation post stroke, and a 16% ceiling effect was reported for motor FIM ((Brock et al., 2002) (stroke); (Dromerick et al., 2003) (stroke)). Van der Putten et. al. (1999) (multiple sclerosis and stroke) reported no significant floor or ceiling effects when administering the FIM to stroke patients. Effect sizes of 0.30, 0.34 and 0 were reported for the total-FIM, motor-FIM and cognitive-FIM respectively. Wallace et al. ((2002); stroke) reported ES=0.28 (0.46 in known changers), SRM=0.62 (0.94 among known changers) and AUC ROC curve=0.675. Dromerick et al. (2003) reported SRM=2.18 from admission to discharge from rehabilitation and the FIM detected change in 91/95 individuals including change in 18 patients in whom the BI detected no change (p<0.001). FIM motor was predictive of direct assistance required while FIM cognition scores was predictive of amount of supervision required ((Corrigan et al., 1997) (TBI). SRM=0.48 was reported for FIM total and 0.54 and 0.17 for motor and cognitive FIM respectively, with no significant floor/ceiling effects reported although there was a 16.1% ceiling effect noted for cognitive FIM (Hobart et al., 2001).

Standardized Interview: Daving et al. ((2001); stroke) examined the FIM home interview for intrarater stability,kappa values >0.40 on 17/18 items. Motor FIM reliability was reported higher than social/cognitive items (K= 0.46 to 0.61). On sequentially separate interviews, self-care items had K values of 0.4-0.6, while transfers, locomotion, and social/cognitive items were below 0.4 (poor).

Telephone Interview: Smith et al. ((1996); stroke) reported total-FIM ICC=0.97, motor-FIM ICC= 0.98, and cognitive-FIM ICC=0.57, comparing telephone interview with direct observation in the home. Item level agreement was superior for items in the motor domain (Kappa values exceeded 0.45). Petrella et al. (2002) (orthopedics) reported good predictive validity (discharge FIM vs phone FIM at 8 weeks; r=0.436, p=0.02) though not as good as observed FIM scores (r=0.699, p<0.0001). Phone FIM showed good concurrent validity with the observed FIM (r=0.741, p<0.0001) and was sensitive to change over time (t=-3.603, p=0.001). Duncan et al. ((2002); stroke) reported a 46% ceiling effect on the motor FIM when administered by telephone at 6-month follow-up of stroke patients.

Segal et al. (1996) (stroke) reported ICCs for patient assessment versus proxy assessment for both in-person and telephone interviews of 0.90 and 0.91, respectively. Agreement was much higher for motor-FIM than for cognitive-FIM. The authors speculate that, due to the more subjective nature of the cognitive dimension of the FIM, this portion of the scale may not be appropriate for proxy assessment.

Agreement between FONE-FIM scores provided by the patient versus a significant other was stronger for motor items (ICC=0.79) than for cognitive items (ICC=0.61) (Tepper et al., 1996) (TBI).

Agreement between patient and proxy FIM scores was reported to be ICC=0.88 for the motor FIM and 0.38 for cognitive FIM. Poorest agreements were noted for items that were most subjective and required opinion/judgement. In cognitive areas, proxies tended to score patients lower than the patients did themselves, while in activity areas, proxies tended to score patients higher than the patients themselves. When patients were grouped according to severity, it was noted that among patients with severe TBI, proxies rated patients as less disabled that the patients themselves, while for less severely injured patients, the opposite was true (Cusick et al., 2000) (TBI).