Characterizing Practice Effects in Performance-Based Tests Administered to Users of Unilateral Lower Limb Prostheses: A Preliminary Study.

Abstract

Performance-based tests are viewed as a gold standard for measuring physical capability. Practice effects, however, may threaten their predictive, discriminative, and evaluative applications. Despite these potential consequences, practice effects have received limited attention in users of lower limb prostheses (LLP). To perform an initial characterization of the occurrence, time-course, and magnitude of practice effects in three performance-based tests administered to users of LLP. Secondary analysis of data from a multisite repeated-measures study. Outpatient clinic and research laboratory. Convenience sample of established ambulatory users of unilateral transtibial and transfemoral prostheses (n = 60). Not applicable. Practice effects were identified as significant changes in slope of participants' cumulative trial-by-trial records. The occurrence, time-course, and magnitude of practice effects were computed for the Timed Up and Go (TUG), Four Square Step Test (FSST), and the 10-m Walk Test (10mWT). Across tests, practice effects were observed in 45% to 76% of participants. The proportion of participants with practice effects (ie, occurrence) was significantly greater for the FSST than the 10mWT (P = .008). The median number of trials (ie, time-course) required for participants to reach a consistent level of performance was not significantly different between tests (FSST: 4 trials, TUG: 4 trials; 10mWT: 3.5 trials; P = .481). Practice effect magnitude (ie, difference between the mean of trials during the plateau and best performance over the first two trials) was significantly greater than zero for the FSST (1.6s; 16%) and TUG (1.4s; 13%) (P < .05). Results indicate that the FSST, TUG, and 10mWT were susceptible to practice effects in this sample of users of LLP. Practice effects may obscure significant differences in walking and balance ability, and thereby, in the absence of modifications, limit the use of these tests for making individual patient clinical decisions and analyzing group-level data.