Date Published: May 29, 2019
Publisher: Public Library of Science
Author(s): Gareth D. Jones, Darren C. James, Michael Thacker, Rhian Perry, David A. Green, Eric R. Anson.
Gait-initiation onset (GI-onset) during sit-to-walk (STW) is commonly defined by mediolateral ground-reaction-force (xGRF) rising and crossing a threshold pre-determined from sit-to-stand peak xGRF. However, after stroke this method [xGRFthresh] lacks validity due to impaired STW performance. Instead, methodologies based upon instance of swing-limb maximum-vertical-GRF [vGRFmaxSWING], maximum-xGRF [xGRFmax], and swing-limb heel-off [firstHEELoff] can be applied, although their validity is unclear. Therefore, we determined these methodologies’ validity by revealing the shortest transition-time (seat-off–GI-onset), their utility in routinely estimating GI-onset, and whether they exhibited satisfactory intra-subject reliability.
Twenty community-dwelling stroke (60 (SD 14) years), and twenty-one age-matched healthy volunteers (63 (13) years) performed 5 standardised STW trials with 2 force-plates and optical motion-tracking. Transition-time differences across-methods were assessed using Friedman tests with post-hoc pairwise-comparisons. Within-method single-measure intra-subject reliability was determined using ICC3,1 and standard errors of measurement (SEMs).
In the healthy group, median xGRFthresh transition-time was significantly shorter than xGRFmax (0.183s). In both the healthy and stroke groups, xGRFthresh transition-times (0.027s, 0.695s respectively) and vGRFmaxSWING (0.080s, 0.522s) were significantly shorter than firstHEELoff (0.293s, 1.085s) (p<0.001 in all cases). GI-onset failed to be estimated in 48% of stroke trials using xGRFthresh. Intra-subject variability was relatively high but was comparable across all estimation methods. The firstHEELoff method yielded significantly longer transition-times. The xGRFthresh method failed to routinely produce an estimation of GI-onset estimation. Thus, with all methods exhibiting low, yet comparable intra-subject repeatability, averaged xGRFmax or vGRFmaxSWING repeated-measures are recommended to estimate GI-onset for both healthy and community-dwelling stroke individuals.
Stroke incidence is high in the UK with ~150,000 cases per year . While inter-disciplinary stroke management has led to an 80% survival rate , nearly 40% of survivors require assistance with everyday activities . These activities include transitional movements that are executed when initiating or arresting movement, for instance gait initiation (GI) or sit-to-stand (STS) [4, 5]. Stroke survivors find transitional movements particularly challenging, which contributes to an increased fall risk . Therefore, effective post-stroke rehabilitation strategies that target transitional movements and reliable assessment metrics to track functionality are vital.
Our main findings are that estimating GI-onset using the xGRFthresh method results in short transition-phase durations in most healthy individuals and some community-dwelling stroke patients; therefore representing a valid method (criterion 1). However, its utility was poor (criterion 2) with GI-onset unable to be estimated in a high proportion of trials particularly in stroke. The firstHEELoff method lacked validity by generating significantly longer transition-phase times than xGRFthresh or vGRFmaxSWING in both groups. There was no significant difference in transition-phase time between the vGRFmaxSWING or xGRFmax methods, and intra-subject reliability (criterion 3) was poor-to-moderate for each estimation method in both groups.
Our aim was to determine an optimal approach to estimate STW GI-onset suitable in both healthy and community-dwelling ambulatory stroke individuals from 4 different methods based on validity, utility and reliability criteria. The firstHEELoff method was the least valid by yielding significantly longer transition-times, thereby placing GI-onset at the end of the anticipation-phase of GI. The utility of the xGRFthresh method was poor because it failed to routinely estimate GI-onset, particularly in the stroke patients. In contrast, both the xGRFmax and vGRFmaxSWING methods were valid and presented with favourable utility using one and two force plates respectively. However, because single measure repeatability is poor-to-moderate for all estimation methods, averaging transition-phase times from multiple trials is required to mitigate high intra-subject variability. In conclusion, average repeated-measures using the xGRFmax or vGRFmaxSWING methods appear able to estimate GI-onset across the continuum of STW performance.