Table 17.19 Characteristics of the Functional Independence Measure+Functional Assessment Measure
| Criterion | Evidence |
| Reliability |
Test-Retest: ICC= 0.98 ((Hobart et al., 2001); stroke). Interobserver Reliability: Interrater (untrained raters) agreement reported to be 67% for FAM items and 55% for patients at admission to rehabilitation (Hall et al., 1993). ). Agreement between raters was less than 70% for 7 items and κ values for FIM+FAM items ranged from 0.35-0.95, while FAM items ranged from 0.35 (adjustment to limits) to 0.92 (swallowing). Possible observer bias was identified for 4 items: employability, writing, comprehension and problem-solving (McPherson et al., 1996). At the item level, interrater reliability ranged from ICC= 0.36 (social interaction) to 0.97 (transfer-toilet, transfer-bed/chair/wheelchair & stairs) and the average ICC for motor FIM+FAM was 0.91 and cognitive FIM+FAM was 0.74, while ICC values for total FIM+FAM was 0.83 (Donaghy & Wass, 1998) (TBI). Internal Consistency: Cronbach’s α=0.96 for FIM+FAM total, 0.96 for FIM+FAM motor and 0.91 for FIM+FAM cognitive/social item-to-total correlations ranged from 0.40-0.81 for FIM+FAM with mean inter-item correlation of 0.46 (Hobart et al., 2001). Values included α=0.99 for motor scale, 0.98 for cognitive scale and 0.99 for total FIM+FAM ((Hawley et al., 1999); ABI). |
| Validity |
Construct Validity: FIM+FAM not unidimensional, as factor analysis demonstrated 2 principal components, with eigenvalues >1-16 items reflecting physical functioning and 14 items reflecting cognition, language and psychosocial functioning (Hawley et al., 1999). Linear regression analysis revealed that FIM+FAM cognition scores at 6 months explained 33% of variance in CIQ scores at 6 months post-discharge while FIM+FAM motor scores accounted for 22% of variance (this was compared to FIM cognition and motor scores that accounted for 31% & 21% of the variance, respectively (Gurka et al., 1999). Concurrent Validity: Hall et al. (1993) reported FIM+FAM motor scores correlated with FIM motor (r=0.992) and FIM cognition scores (r=0.645) as well as with DRS ratings (r=0.680). FIM+FAM cognition scores correlated with FIM motor (r=0.652), FIM cognition (r=0.952), DRS ratings (r=0.746), and LCFS scores (r=0.626). FIM+FAM correlated with BI (r=0.525, p<0.001), OPCS Index (r=0.824, p<0.001) and with the original FIM (r=0.962, p<0.001) (McPherson & Pentland, 1997). Concurrent Validity (Convergent/Discriminant): FIM+FAM total and motor FIM+FAM 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 or psychological distress (SF36 mental component, General Health Questionnaire). However, cognitive FIM+FAM 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) (stroke). |
| Responsiveness | When ceiling effect is defined as scoring ≥180 on the FIM+FAM, 34% of patients scored in the ceiling range at discharge from rehabilitation and 79% at one year post discharge. This represented an improvement over the FIM (49%, 84%). There was no advantage in terms of ceiling effect seen with regard to cog-FIM and the cognitive items of the FIM+FAM (K. Hall et al., 1996b). Rasch analysis revealed FAM items cover a wider range of difficulty than the FIM items and, therefore, expand the range of scale difficulty beyond the FIM alone. However, both FIM and FAM items tend to cluster in the mid-range (Hall et al., 1993), with 2% of patients reportedly obtaining maximum scores on FIM+FAM (McPherson & Pentland, 1997), and 80-90% of patients obtained “near maximum” scores on the FAM (Gurka et al. 1999). In terms of SRM means for FIM+FAM total, motor FIM+FAM and cognitive FIM+FAM were reported to be 0.42, 0.52 and 0.19 respectively-there were no significant floor or ceiling effects reported for FIM+FAM (Hobart et al., 2001). |
| Tested for ABI/ TBI patients? | Developed specifically for ABI/TBI population. |
| Other Formats | UK FIM+FAM: A version of the FIM+FAM adapted for use in the United Kingdom resulting in revised manuals and decision trees for items identified as particularly difficult to score. Accuracy of scoring by individuals (when compared to a vignette with previously determined “correct” scores) was reported to be 77.1%. Accuracy of team scoring was reported to be 86.5% for the total score. Revision of the manual & decision trees increased accuracy of scoring for items perceived as difficult to score (Turner-Stokes et al., 1999). In seven studies, in which the majority of the population had an ABI, the UK version of the FIM+FAM demonstrated acceptable utility, concurrent validity, inter-rater reliability, and responsiveness (Turner-Stokes & Siegert, 2013). |
| Use by proxy? | N/A |