Short communicationMinimal detectable change for gait variables collected during treadmill walking in individuals post-stroke
Introduction
Treadmill training is commonly used for post-stroke gait rehabilitation [1], [2], [3]. To evaluate the efficacy of gait rehabilitation, it is critical to accurately measure post-stroke gait impairments before and after rehabilitation. Kinematic, ground reaction force (GRF), and spatiotemporal gait data provide a comprehensive assessment of gait deficits [4], [5], [6], [7]. An estimate of typical variability in gait data with repeated testing is helpful in determining if observed changes in gait are due to the intervention being investigated or simply measurement errors due to test–retest or day-to-day gait variability. The minimal detectable change (MDC) [8], [9] represents the amount of change in a variable necessary to conclude that the change is not attributable to error; it is the smallest change that falls outside the expected range of error and represents a “real” change [8], [9].
The reliability of kinematic and spatiotemporal data [5] and MDCs for GRF data [6] during overground walking has been previously reported for individuals post-stroke. A limitation in these studies was that the test and retest were conducted on the same day and therefore, did not account for day-to-day gait variability [5], [6]. Thus, application of these results when interpreting data from reports involving pre- and post-intervention testing is limited. Furthermore, despite the increasing use of treadmills for stroke rehabilitation [1], [2], [3] and for measuring gait [1], [2], [4], [10], MDCs for gait variables obtained during treadmill walking post-stroke have not been previously reported. The objective of the current study, therefore, was to determine MDCs for kinematic, GRF, temporal, and spatial gait variables during treadmill walking using data from repeated testing sessions in individuals with post-stroke gait impairments.
Section snippets
Methods
Nineteen individuals (12 males; age 47–75 years) with post-stroke hemiparesis (72.6 ± 63.4 months post-stroke) able to walk for >5 min were recruited (Table 1). Exclusion criteria included inability to follow commands and orthopedic or other neurologic conditions interfering with walking. All subjects signed consents approved by the Human Subject Review Board. Each subject participated in 2 sessions conducted on separate days (20.7 ± 26.8 days between sessions), with identical testing procedures.
Results
We reported MDC values (see Table 2) and demonstrated excellent between-session (ICCs range from 0.799 to 0.986) and within-session reliability (ICCs ≥ 0.9) for kinematic, spatio-temporal, and GRF data collected during treadmill walking in individuals post-stroke. Bland–Altman plots showed no systematic trends in the direction or distribution of test–retest errors for majority of variables (Fig. 2).
Discussion
We reported MDCs for post-stroke gait variables obtained via three-dimensional motion capture during treadmill walking. The between-session MDCs from our study account for test–retest errors in gait variables across 2 testing sessions caused by factors such as recalibrating the camera system, re-attaching reflective markers, and day-to-day gait variability. Between-session MDCs estimate the minimum change in gait that must be produced by an intervention for the change to be considered real;
Acknowledgements
Funding Sources: National Institutes of Nursing Research R01 grant NR010786 and Bioengineering Research partnership R01 grant HD038582 to Dr. Binder-Macleod; NIH K01 HD050582 to Dr. Reisman; NIH Shared Instrumentation Grant S10 RR022396-01 to Dr. Lynn Snyder-Mackler; DOD Grant W911NF-05-1-0097 to Dr. Irene Davis. The authors thank Ms. Margie Roos, PT, NCS for clinical testing and subject recruitment; Ms. Leigh Shrewsbury for scheduling and recruitment; Erin Helm for assistance with
References (14)
- et al.
Repeatability of lower limb three-dimensional kinematics in patients with stroke
Gait Post
(2008) - et al.
Reliability, smallest real difference and concurrent validity of indices computed from GRF components in gait of stroke patients
Gait Post
(2009) - et al.
A taxonomy for responsiveness
J Clin Epidemiol
(2001) - et al.
Relationship between step length asymmetry and walking performance in subjects with chronic hemiparesis
Arch Phys Med Rehabil
(2007) - et al.
Evaluation of gait symmetry after stroke: a comparison of current methods and recommendations for standardization
Gait Post
(2010) - et al.
Comparing methods of measurement: why plotting difference against standard method is misleading
Lancet
(1995) - et al.
A randomized controlled trial of functional neuromuscular stimulation in chronic stroke subjects
Stroke
(2006)
Cited by (93)
Biomechanical differences between self-paced and fixed-speed treadmill walking in persons after stroke
2022, Human Movement ScienceCitation Excerpt :Most variability and symmetry outcome variables were not different in both conditions, except for the standard deviation of the paretic step length and the step length symmetry. However, both differences are too small to be clinical relevant (Geiger et al., 2019; Kesar et al., 2011). Additionally, no clinical relevant differences could be detected for the other biomechanical parameters (kinematics, kinetics, joint angle variability).
Combined user-driven treadmill control and functional electrical stimulation increases walking speeds poststroke
2021, Journal of BiomechanicsCitation Excerpt :All comparisons were performed within subjects, and Bonferroni corrections were used to adjust for multiple comparisons, which gave an αcritical = 0.0083 for the walking speed comparisons and αcritical = 0.0167 for all other comparisons. Within-session minimum detectable change (MDC) values for AGRF, PGRF, and TLA refer to work from Kesar and colleagues in 2011 (Kesar et al., 2011a). Since AGRF and PGRF are in the same plane but opposite directions, we used the AGRF MDC threshold of 0.80% BW as the detectable change threshold for PGRF.