Reliability, validity, and responsiveness of the locomotor capabilities index in adults with lower-limb amputation undergoing prosthetic training.

Abstract

To assess the reliability, validity, and responsiveness of both the standard and revised Locomotor Capabilities Index (LCI) in people with lower-limb amputation who undergo prosthetic training. Reliability and validity study. Two freestanding rehabilitation centers. Fifty inpatients with a recent unilateral lower-limb amputation. Not applicable. Main outcome measures The standard LCI and a new version with a 5-level ordinal scale (LCI-5) were tested for internal consistency, test-retest reliability, ceiling effect, and effect size. The construct validity of both versions was analyzed by correlation with the Rivermead Mobility Index, a timed walking test, and the FIM instrument. The Cronbach alpha of both LCI versions was.95. The item-to-total correlations (Spearman rho) ranged from.50 to.87 (P<.0001 for all). The percent agreement and kappa values for the item scores ranged, respectively, from 78.4% to 100% and.58 to 1.00 in the LCI, and from 75.7% to 97.3% and.54 to.96 in the LCI-5. The intraclass correlation coefficient (model 2,1) for the total scores was.98 for both versions; the Bland-Altman plot revealed no systematic trend for either version. Both the LCI and LCI-5 correlated with all criterion measures (rho range,.61-.76), with the LCI-5 showing a larger effect size during the rehabilitation period and a lower ceiling effect. Patients with transtibial amputation were more independent in performing activities than were those with transfemoral amputation; their locomotor capability negatively correlated with age. Both the LCI and LCI-5 captured the global locomotor ability of people with lower-limb amputation during prosthetic training. The new LCI-5 presents similar and sometimes better psychometric properties than the standard LCI.