Reproducibility of the physiological cost index among individuals with a lower-limb amputation and healthy adults

Abstract

Background and purpose: The physiological cost index (PCI) is a clinical measurement used to estimate the energy cost of walking. The reproducibility of the PCI has been questioned and no study has investigated the measurement error among individuals with a lower-limb amputation. The aim was to investigate the test-retest reproducibility of the PCI in individuals with a lower-limb amputation and healthy adults.

Methods: The study comprised 28 individuals (20 males, eight females, mean age 49 years) with a unilateral amputation due to reasons other than vascular disease and 31 healthy volunteers (20 males, 11 females, mean age 47 years). PCI values were obtained by registering heart rate at rest and during level indoor walking for 5 minutes at a comfortable speed. A within-day test-retest assessment was performed. Reproducibility analyses included intra-class correlation, analyses of systematic differences between measurements, calculation of the smallest detectable change (SDC) and coefficient of variation (CV), and the results were finally visualized using Bland-Altman plots.

Results: The reliability in terms of intra-class correlation was excellent for both groups (0.966 and 0.948). In the amputee group, the PCI revealed a mean difference of 0.026 (p = 0.016) between tests (PCI = 0.555, standard deviation [SD] = 0.214 and PCI = 0.581, SD = 0.236, respectively). In the healthy group, there was no systematic difference between tests (PCI = 0.329, SD = 0.114 and PCI = 0.331, SD = 0.110, respectively). The SDC was 0.116 in the amputee group and 0.070 in the healthy group, giving a CV of 20.4% and 21.0%,respectively.

Conclusions: The within-day test-retest reproducibility of the PCI was excellent among individuals with lower-limb amputations and healthy adults in terms of intra-class correlation and acceptable in terms of agreement. The SDC, which was calculated for each group, should be considered when demonstrating an individual difference after an intervention.