198
R. Maritz et al.
Table III. Functional Independence Measure (FIM™) total score
transformation table: original scores to interval scores
Original
FIM™
score Rasch
Transformed
estimate interval score
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72 –6.279
–5.686
–5.281
–5.005
–4.784
–4.594
–4.423
–4.263
–4.112
–3.966
–3.824
–3.686
–3.552
–3.420
–3.291
–3.166
–3.043
–2.923
–2.807
–2.693
–2.582
–2.473
–2.367
–2.264
–2.163
–2.064
–1.968
–1.874
–1.783
–1.693
–1.605
–1.519
–1.435
–1.353
–1.273
–1.194
–1.116
–1.041
–0.966
–0.893
–0.821
–0.750
–0.680
–0.611
–0.543
–0.476
–0.409
–0.343
–0.278
–0.213
–0.149
–0.085
–0.021
0.043
0.106
18.0
23.3
26.9
29.4
31.4
33.1
34.6
36.0
37.4
38.7
40.0
41.2
42.4
43.6
44.8
45.9
47.0
48.0
49.1
50.1
51.1
52.1
53.0
53.9
54.9
55.7
56.6
57.4
58.3
59.1
59.8
60.6
61.4
62.1
62.8
63.5
64.2
64.9
65.6
66.2
66.9
67.5
68.1
68.7
69.4
70.0
70.6
71.1
71.7
72.3
72.9
73.5
74.0
74.6
75.2
Original
FIM™
score
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
Rasch
Transformed
estimate interval score
0.169
0.232
0.295
0.358
0.422
0.485
0.548
0.612
0.677
0.741
0.806
0.871
0.937
1.004
1.071
1.139
1.208
1.277
1.347
1.418
1.491
1.564
1.638
1.714
1.791
1.869
1.949
2.030
2.113
2.197
2.283
2.371
2.461
2.553
2.647
2.742
2.840
2.941
3.043
3.148
3.256
3.365
3.478
3.593
3.712
3.834
3.963
4.098
4.245
4.410
4.605
4.853
5.225
5.784
75.7
76.3
76.9
77.4
78.0
78.6
79.1
79.7
80.3
80.9
81.4
82.0
82.6
83.2
83.8
84.4
85.0
85.6
86.3
86.9
87.6
88.2
88.9
89.6
90.3
90.9
91.7
92.4
93.1
93.9
94.7
95.4
96.2
97.1
97.9
98.8
99.6
100.5
101.5
102.4
103.4
104.3
105.4
106.4
107.4
108.5
109.7
110.9
112.2
113.7
115.4
117.7
121.0
126.0
ordinal-scaled FIM™ raw scores into interval-scaled
FIM™ scores (see Table III).
DISCUSSION
This is the first study to provide evidence of the unidi-
mensionality of the FIM™ 18-item version when admi-
nistered to neurological and musculoskeletal patients
in an inpatient rehabilitation setting. Although the
www.medicaljournals.se/jrm
baseline analyses and the traditional testlet approaches
did not result in Rasch model fit, an alternative 2-testlet
approach, emphasizing the sameness of the FIM™
items with 2 equally sized super-items, achieved mo-
del fit. The robustness of these results was confirmed
in repeating the same strategy for all subsets of the
calibration sample. These results provide evidence for
the internal construct validity of the FIM™ total scores
based on 18 items, and thus support its reporting as
a total score in clinical practice. Based on the current
results, an interval-scale transformation table of the
FIM™ total scores for use in national quality moni-
toring for neurological and musculoskeletal patients
could be provided.
The traditional testlet approach applied in this study
builds on the successful Rasch analysis strategy for
FIM™ motor items (17). Nevertheless, when adding
the cognitive item set to the motor items, the 4 testlet
solution in the original paper could not be confirmed in
this study. Highlighting the sameness of all the items
through the 2-testlet approach attained model fit. The
assumption of multidimensionality is often pursued
in FIM™-related Rasch analyses (36). The extent of
local dependency among the 18 items, clustered into
the underlying structures raises questions as to whether
the FIM™ should a priori be divided into multidimen-
sional concepts, as proposed by Linacre and colleagues
(16). The successful summation of the 18 FIM™ items
in the current study places emphasis on the higher order
construct of functional independence, incorporating
both motor and cognitive aspects. Likewise, the Rasch
analysis performed supports the theory that, when
activities of daily living are observed, motor activities
reflect some cognitive aspects and vice versa (18, 37).
From a clinical perspective, the FIM™ offers different
levels of granularity for reporting. Scores can be repor-
ted at the level of the single items, the item headings,
e.g. sphincter control, on the level of the motor and
cognitive subscales, or the level of the overall 18-item
summary of functional independence. Thus, different
granular levels of reporting are available, depending
upon the use required.
The study can be also be seen as initial evidence
that the German, French and Italian translations of
the FIM™ do not substantially differ from each other,
given the absence of substantial DIF by language.
Furthermore, this study provides first evidence for
the internal construct validity of the FIM™ 18-item
version for musculoskeletal patients, given that no sub-
stantial DIF was present between the musculoskeletal
and the neurological rehabilitation group. Neverthe-
less, the use of cognitive items for a musculoskeletal
patient population within a national outcome report,
as in the ANQ, remains debatable, and care should be