Examination of the Validity of the FIM in Patients with Neurological Conditions in Early Post-acute Rehabilitation Facilities

Purpose: The most frequently used assessment instrument in the early stage of rehabilitation is the Functional Independence Measure (FIM), which quantifies patient's level of independence in basic daily life activities. The FIM is a multidimensional measure. Postulated dimensions within the FIM are the motor and the cognitive dimension. Until now the psychometric properties of the FIM in patients with more severe functioning limitations are not sufficiently known. The objective of this study is therefore to study the psychometric properties of the FIM in patients with more severe functioning limitations in the early stages of rehabilitation. The specific aims are to examine (1) whether there are floor effects and hence whether the FIM discriminates functioning across patients with neurological conditions in early post-acute rehabilitation facilities and (2) whether the two dimensional structure also applies to our study population.

Methods: The study population consisted of a convenience sample of patients with neurological conditions treated in early post-acute facilities. The FIM measures patients' performance of basic activities of daily living using 18 items. The FIM is a Likert-like-summated rating scale, with response categories for each item ranging in value from 1 to 7. Scores 1 and 2 indicates stages of dependence, and scores 3 to 7 stages of independence in that item.

To assess the discriminative capability of the FIM in patients with neurological conditions in early rehabilitation facilities floor effects for the FIM items were examined. A floor effect is defined as the percentage of the sample scoring the minimum possible item score. Floor effects exceeding 50% are considered to be substantial. We expected floor effects in FIM items, particularly in patients with impaired consciousness. Thus, results are reported stratified by the state of consciousness. Impairment in consciousness was documented with the dichotomous ICF category consciousness functions (b110) (0='no impairment', 1='impairment').

To examine the dimensional structure of the FIM a Latent Class Factor Analysis (LCFA) was carried out. Due to the significant floor effects in FIM items of patients with impaired consciousness, LCFA was carried out only with data of patients without impaired consciousness.

Result: 292 neurological patients were included in the study. Of the 292 patients, 261 (89.4%) had complete information on all FIM items. Analyses were carried out on those 261 patients. 54 patients (20.7%) had impairment in the ICF category consciousness functions (b110). Six FIM items had substantial floor effects. In patients without impaired consciousness the two postulated dimensions could be identified. Model fit, however, indicated a more complex dimensional structure.

Discussion: The FIM is appropriate for the assessment of patients without impaired consciousness, but shows substantial floor effects in patients with impaired consciousness. The dimensional structure of the FIM might be more complex in neurological patients in early stages of rehabilitation.