Asymmetry measures
Walking on the rehabilitative shoe may benefit gait symmetry. All participants improved step length symmetry after training, and the average change in step length symmetry found in our study is similar to that shown in a study focused on gait symmetry during split-belt treadmill (SBT) training [
37]. Four participants improved DLS symmetry; the two that did not improve were the severely impaired participant (initial gait velocity of 9.0 cm/s) and the highly functional participant (initial gait velocity of 113.5 cm/s). Although these two did not respond with a DLS asymmetry change, participant 3 improved step length asymmetry and had a substantial decrease on the TUG, and participant 6 improved on the 6MWT and had a substantial increase in gait velocity. For comparison, related studies show no change in DLS symmetry following SBT training [
37,
38]. Our results suggest that over-ground gait training using the rehabilitative shoe could provide an additional benefit to the recovery of DLS symmetry for some individuals after stroke.
The literature does not provide estimates of the clinical relevance of gait asymmetry measures while walking over ground, but does provide some spatiotemporal measures for treadmill walking [
39]. However, gait asymmetry has been associated with balance [
40,
41] and is considered a major cause of future degenerative issues with hips, knees, and backs for stroke survivors with gait hemiparesis [
42,
43].
Functional measures
Walking on the rehabilitative shoe may help individuals with hemiparetic stroke improve their functional walking. Two of the participants’ gait velocity increased beyond a substantial meaningful change (≥ 14 cm/s), two other participants’ gait velocity increased beyond a small meaningful change (≥ 6 cm/s), and the remaining two improved less than these ranges. These ranges are based on people 30 to 150 days post stroke [
35,
44]. One participant improved gait velocity beyond the clinically meaningful change of ≥ 16 cm/s that another study reported for people less than 60 days post stroke [
45]. All the participants in our study were more than one year post stroke, which is much greater than the groups reported in these studies. Another important measure is that two out of three participants who were initially categorized as household ambulators (i.e., gait velocity of < 40 cm/s) became limited community ambulators (i.e., gait velocity of 40–80 cm/s) after training [
5]; these two participants were 5 and 10 years post stroke
All participants who were assessed improved on the TUG, and two of them improved beyond the MDC of − 3.5 sec [
36]. Although all five participants assessed improved the distance walked in the 6MWT following training, none of them surpassed the smallest minimal clinically important difference (MCID) of 34 m reported by Fulk and He [
46]. The four-week training with only 6 hours total walk time may not be long enough for each participant to show a meaningful change in aerobic capacity. Continued daily use of the device for a longer time coupled with concomitant exercise may help them further increase aerobic capacity over time.
Therapeutic mechanisms
The device presented is unlike any known existing rehabilitation therapies and is thought to function through a combination of mechanisms. These mechanisms likely benefit each individual uniquely since stroke presents in different ways. For example, all the participants showed a shorter stance phase with the paretic side compared to the nonparetic side, but three of the participants (1, 2, & 6) had a shorter step length with the paretic side. Although these three participants showed a smaller improvement in step length asymmetry, participant 1 showed the largest double limb support asymmetry improvement and participants 2 and 6 showed the largest gait velocity improvements. Encouraging more use of the paretic foot likely had a larger benefit to these participants. These unique benefits suggest that our device may have a heterogenous set of mechanisms that can benefit a wide set of stroke patients’ specific gait impairments. Further, all of our participants benefitted from this treatment, which is different than some of the SBT studies that show no gait symmetry improvements, especially step length symmetry, in approximately 40% of participants [
37,
38]. Below are details on some the mechanisms we believe cause our device to help correct gait.
Asymmetric Motion: Both the presented device and the SBT cause one foot to move backward faster than the other. In SBT training, the gait asymmetry of the patient is increased by having two treads move at different speeds so that the patient must compensate to stay moving on the treadmill. When the belts are returned to the same speed, the patient will retain the “adjusted”, now more symmetric, gait on the treadmill [
37,
48]. Our presented device moves the foot backward relative to the paretic foot, much like the motion of the fast tread of a SBT. Both the SBT and our device beneficially change step length symmetry, but only our device shows improvements in double limb support symmetry. This additional gait benefit is likely due the device attaching to the foot, which allows training in an over-ground context.
Context Awareness: The corrected walking patterns from existing therapeutic methods, such as treadmills, do not completely transfer to over-ground walking because the dynamic and sensorimotor aspects of walking over ground are distinctly different than walking on a treadmill [
22,
28,
49,
50]. Research has indicated that only about 60% of the gait correction from walking on a split-belt treadmill transfers to walking over ground in individuals with stroke [
51].
When walking over ground, an individual has complete control over velocity, whereas the treadmill speed limits one’s ability to change velocity. Another important difference is the amount of visual flow: on a treadmill, the scene is not moving, so there are no visual cues reinforcing the forward motion that would be present when walking over ground. Since walking is highly context dependent [
25,
51‐
54], these visual cues indicating a different context may prevent the learned patterns on the treadmill from being expressed during over-ground walking. Our device allows over-ground walking in the environment of daily activities. A user of our device experiences a congruent dynamic optical/visual flow as opposed to an individual on a SBT, who typically views a static scene that is incongruent to training movements.
Cueing: The benefits of this device may also arise from the multiple cues produced by the device that guide the user through their gait. The first cue is that the nonparetic foot height is decreasing after first contact in stance; a second cue is that the nonparetic foot begins moving backward during the transition to stance. These cues start before the paretic leg transitions from stance, which provides a set of cues that possibly indicate the type of step to take with the paretic leg. For example, the first cue may induce more weight bearing on the paretic leg at mid-stance, while the second cue may foster earlier toe off of the paretic leg at terminal stance.
Encouraging Paretic Leg: The device can also increase the relearning of the paretic leg by reducing the effective output of the nonparetic leg by generating a backward motion. The motion induced by the device encourages the wearer to increase the use of their paretic leg. This effect is similar to the idea of Constraint Induced Movement Therapy [
15,
16]. By slightly destabilizing their nonparetic leg, the user will naturally start to spend more time on their paretic side, which may help to foster those abilities and confidence in using that side of their body.
Home Rehabilitation: The literature has continued to show that patients are dissatisfied with their options for training after they are discharged from the rehabilitation hospital/clinic [
55‐
59]. Moreover, most individuals with stroke prefer a home-based approach for their initial rehabilitation [
60]. The ability to train at home enables individuals to more frequently rehabilitate themselves, which leads to better results in motor relearning [
61] and can maintain individuals’ ability to perform activities of daily living [
62,
63]. Our device has the potential to be used in the home setting, which could reduce the costs and increase the access as well as the amount of rehabilitation.
There are open questions related to the frequency of training, the length of each session, and how many weeks the training should continue. The intensity of the training during each session can also be customized by adjusting the spiral wheel to make the generated backward motion longer and/or faster. This customization could also be adjusted at regular intervals to keep a constant intensity level. Future studies will evaluate how to optimize the therapy further.
Safety is vital, particularly during home care. Using the device independently in a safe way is being evaluated in a separate home-based trial. In the study presented here, we found that participants became comfortable with the device within the first three sessions and needed little or no assistance after that. Out of the over 400 bouts of walking in our study, the attending PT only provided physical assistance twice due a perceived need for patient support. As such, we expect that home-based therapy could be provided for many patients after they complete a few sessions in the clinic and become qualified for home-use. The specific requirements of being eligible for home-use are being evaluated and will be discussed further once the larger home-based study is complete.