Clinical Delivery of Overground Exoskeleton Gait Training in Persons With Spinal Cord Injury Across the Continuum of Care: A Retrospective Analysis

Abstract

Background:

After spinal cord injury (SCI), inpatient rehabilitation begins and continues through outpatient therapy. Overground exoskeleton gait training (OEGT) has been shown to be feasible in both settings, yet its use as an intervention across the continuum has not yet been reported.

Objectives:

This study describes OEGT for patients with SCI across the continuum and its effects on clinical outcomes.

Methods:

Medical records of patients with SCI who completed at least one OEGT session during inpatient and outpatient rehabilitation from 2018 to 2021 were retrospectively reviewed. Demographic data, Walking Index for Spinal Cord Injury-II (WISCI-II) scores, and OEGT session details (frequency, “walk” time, “up” time, and step count) were extracted.

Results:

Eighteen patients [male (83%), White (61%), aged 37.4 ± 15 years, with tetraplegia (50%), American Spinal Injury Association Impairment Scale A (28%), B (22%), C (39%), D (11%)] completed OEGT sessions (motor complete, 18.2 ± 10.3; motor incomplete, 16.7 ± 7.7) over approximately 18 weeks (motor complete, 15.1 ± 6.4; motor incomplete, 19.0 ± 8.2). Patients demonstrated improved OEGT session tolerance on device metrics including “walk” time (motor complete, 7:51 ± 4:42 to 24:50 ± 9:35 minutes; motor incomplete, 12:16 ± 6:01 to 20:01 ± 08:05 minutes), “up” time (motor complete, 16:03 ± 7:41 to 29:49 ± 12:44 minutes; motor incomplete, 16:38 ± 4:51 to 23:06 ± 08:50 minutes), and step count (motor complete, 340 ± 295.9 to 840.2 ± 379.4; motor incomplete, 372.3 ± 225.2 to 713.2 ± 272). Across therapy settings, patients with motor complete SCI experienced improvement in WISCI-II scores from 0 ± 0 at inpatient admission to 3 ± 4.6 by outpatient discharge, whereas the motor incomplete group demonstrated a change of 0.2 ± 0.4 to 9.0 ± 6.4.

Conclusion:

Patients completed OEGT across the therapy continuum. Patients with motor incomplete SCI experienced clinically meaningful improvements in walking function.

Introduction

Following spinal cord injury (SCI), intensive rehabilitative efforts begin in the inpatient rehabilitation setting and often continue with outpatient therapy services. During the first year following SCI, 44% of total therapy hours occur in inpatient rehabilitation,¹ with an average length of stay of 32 days.² The remaining 56% of therapy hours over the first year are completed in the outpatient setting¹ with goals to promote neurologic recovery and improved function through continued activity-based therapy.³ As walking and mobility-related goals are a high priority to patients with SCI,^4–6 gait training is a common activity-based intervention for persons with SCI during both inpatient rehabilitation and outpatient therapy settings.^3,7

Contemporary usual care gait training efforts vary for people with SCI, with a lack of clear recommendations to guide best practice.⁸ The use of body weight–supported treadmill training and conventional overground gait training using assistive devices has demonstrated limited and equivocal effectiveness to improve walking ability following SCI.^3,7,8 However, advancing technology has developed wearable powered robotic exoskeleton devices capable of walking overground.⁹ These devices are becoming increasingly prevalent and may provide an alternative to usual care gait training approaches.^9,10 Overground exoskeleton gait training (OEGT) has been shown to be safe and feasible with both the subacute^11,12 and chronic⁶ SCI populations and across varying injury levels and severity of SCI.^{6,7,10,13–15} Existing literature suggests a variety of potential benefits including psychological benefits; improved posture, strength, and bowel and bladder sensation and function; decreased pain and spasticity; and reduced complications affecting cardiovascular and respiratory systems, in addition to functional and gait improvements.^7,16,17

Despite documented feasibility and tolerability of OEGT in both inpatient and outpatient settings,^6,11,12 the impact of OEGT on clinical outcomes as a therapeutic intervention carried out across the therapy continuum has yet to be illustrated. With the dramatic decrease in inpatient rehabilitation length of stay after SCI,¹⁸ the risk of postrehabilitation health and functional decline after SCI has increased.¹⁹ Recent findings suggest that higher function²⁰ and a protracted gait intervention after SCI²¹ can reduce health care utilization and maximize outcomes for people with SCI. The purpose of this retrospective study is to describe the use of OEGT beginning early in inpatient rehabilitation, a critical window for neurological recovery,^22,23 and continuing into outpatient therapy settings and its effects on clinical outcomes in persons with motor complete versus motor incomplete SCI. Specifically, we anticipated patients with motor complete and motor incomplete SCI would experience improvements in device metrics, illustrating the ability to be progressively challenged across the therapy continuum using OEGT. Additionally, we hypothesized that patients with motor incomplete SCI, in contrast to motor complete, would demonstrate improvements in walking function over time.

Methods

Participants

Based on review of the medical record, patients were eligible and included if they completed a minimum of one OEGT session during both inpatient and outpatient therapy at our facility following SCI. Eligibility aligned with clinical and exoskeleton manufacturer inclusion criteria and frame limitations: age greater than 18 years; bowel and bladder continence or Foley catheter in place; involved in standing program; having weight of 100 kilograms or less, height between 1.5 and 1.9 meters tall, and standing hip width of 0.45 meters or less; having near normal range of motion in hips, knees, and ankles (able to attain a neutral ankle dorsiflexion with <12 degrees of knee flexion, no more than 12 degrees of knee flexion contracture, and no more than 17 degrees of hip flexion contracture); and having no upper leg length discrepancy greater than 1.3 cm or lower leg discrepancy greater than 1.9 cm. Patients were excluded from OEGT for spinal instability, untreated deep vein thrombosis (DVT), decreased standing tolerance due to orthostatic hypotension, known history of osteoporosis that prevented safe standing or increased the risk of fracture caused by standing or walking, uncontrolled spasticity, uncontrolled autonomic dysreflexia (AD), skin integrity issues on contact surfaces of the device or on surfaces that would prohibit sitting, and pregnancy.¹⁵

Device

The exoskeleton device used was the Ekso GT²⁴ (Ekso Bionics, San Rafael, CA), which is a wearable, battery-operated exoskeleton that is powered by motors at the hip and knee joints. The device allows stepping motion to be initiated by specific patient weight shift; the amount of motor assistance provided by the device can be varied as the patient progresses across therapy sessions. The Ekso GT provides bilateral assistance to the lower extremities and allows for programmatic settings to be adjusted throughout the session to provide maximal, adaptive, fixed, or no assistance (free mode) throughout the gait cycle.²⁴ This device has class II FDA approval for use with people with SCI.²⁵

A typical OEGT session in both inpatient rehabilitation and outpatient therapy settings was 45 minutes during which the participant completed standing and walking tasks in the exoskeleton device. Device parameters and treatment progression were selected by the treating physical therapist and individualized for each patient according to their tolerance and treatment goals.

Procedure

Hospital institutional review board approval and waiver of consent due to retrospective analysis were obtained prior to initiating data collection procedures. Based on eligibility criteria, medical records were retrospectively reviewed at one rehabilitation hospital comprised of inpatient and outpatient departments from December 2018 to December 2021 to identify patients who participated in OEGT for a minimum of one therapy session in each setting. Research staff were trained to review the demographic and clinical variables extracted from the medical record, and extracted data were audited by the research team for accuracy.

OEGT session data were first recorded onto paper case report forms by the treating therapist and then entered into the electronic medical record. A member of the research team then retrieved OEGT session data from the medical record and cross-referenced data with the case report forms for confirmation. Clinical outcome measures were captured at admission and discharge from inpatient and outpatient rehabilitation by the primary therapist.

Outcomes

Demographic data included age, gender, race, ethnicity, diagnosis, mechanism of injury, time since injury, and inpatient rehabilitation and outpatient therapy lengths of stay. Severity of injury was obtained using the American Spinal Injury Association Impairment Scale (AIS) for patients with traumatic injury and an AIS equivalent for nontraumatic injuries.²⁶ The primary clinical outcome collected during both inpatient and outpatient admissions was the Walking Index for Spinal Cord Injury-II (WISCI-II).^27,28 The WISCI-II defines the physical limitation for walking secondary to impairment at the person level and is indicative of the ability of a person to walk after SCI.²⁷ A change of one WISCI level can be considered clinically meaningful.²⁸ OEGT session data included frequency of robotic exoskeleton use, up time (i.e., amount of time spent in the device in upright standing), walk time (i.e., amount of time spent walking in the device), and number of steps.

Data analysis

Demographic and injury-related characteristics were summarized using means and standard deviations for continuous variables and counts and percentages for categorical variables. Demographic variables were summarized overall and stratified by injury impairment (motor complete vs. motor incomplete). OEGT session metrics were summarized by impairment and across both inpatient and outpatient settings. Graphs were used to visualize the trends and change in session metrics over time. Analysis was performed using SAS 9.4 (SAS Institute, Cary, NC).

Results

Demographic and injury characteristics

Overall, 569 patients with a spinal cord injury/disorder (SCI/D) diagnosis were admitted to our inpatient rehabilitation hospital during the 3-year study period. Of these, 74 participated in OEGT during inpatient rehabilitation and 18 continued the intervention into outpatient therapy. These 18 patients met eligibility criteria and were included in this analysis (Table 1). AIS scores indicated that nine patients experienced motor complete injuries (AIS A and B) and nine experienced motor incomplete injuries (AIS C and D). Patients were 37.4 ± 15 years of age, and were majority male (83%), and White (61%). Injuries were primarily due to a traumatic event, with motor vehicle incident (39%) and gunshot wound (17%) being the most common mechanisms. Over half (56%) of the patients had a cervical level injury, and there were no lumbar injuries.

Table 1.

Summary of demographic data

	All (N = 18)	Motor complete (n = 9)	Motor incomplete (n = 9)
Age at injury	37.4 ± 15	30.1 ± 11.3	44.7 ± 15.1
Sex, male	15 (83.3%)	8 (88.9%)	7 (77.8%)
Race
White	11 (61.1%)	6 (66.7%)	5 (55.6%)
Asian	3 (16.7%)	1 (11.1%)	2 (22.2%)
African American/Black	2 (11.1%)	1 (11.1%)	1 (11.1%)
Unknown	2 (11.1%)	1 (11.1%)	1 (11.1%)
Insurance - Inpatient
Self-pay or uninsured	2 (11.1%)	0 (0%)	2 (22.2%)
Unknown/missing	16 (88.9%)	9 (100%)	7 (77.8%)
Insurance - Outpatient
Private	13 (72.2%)	7 (77.8%)	6 (66.7%)
Self-pay or uninsured	2 (11.1%)	2 (22.2%)	0 (0%)
Medicare	1 (5.6%)	0 (0%)	1 (11.1%)
Unknown/missing	2 (11.1%)	0 (0%)	2 (22.2%)
Mechanism of injury
Motor vehicle	7 (38.9%)	3 (33.3%)	4 (44.4%)
Motorcycle	1 (5.6%)	1 (11.1%)	0 (0%)
Gunshot wound	3 (16.7%)	2 (22.2%)	1 (11.1%)
Water sports	2 (11.1%)	1 (11.1%)	1 (11.1%)
Air sports	1 (5.6%)	1 (11.1%)	0 (0%)
Other unclassified	1 (5.6%)	0 (0%)	1 (11.1%)
N/A, nontraumatic	3 (16.7%)	1 (11.1%)	2 (22.2%)
Paraplegia or tetraplegia
Paraplegia	9 (50.0%)	5 (55.6%)	4 (44.4%)
Tetraplegia	9 (50.0%)	4 (44.4%)	5 (55.6%)
SCI level
Cervical	10 (55.6%)	4 (44.4%)	6 (66.7%)
Thoracic	7 (38.9%)	4 (44.4%)	3 (33.3%)
Lumbar	0 (0%)	0 (0%)	0 (0%)
Missing/unknown	1 (5.6%)	1 (11.1%)	0 (0%)
ASIA Impairment Scale
A	5 (27.8%)
B	4 (22.2%)
C	7 (38.9%)
D	2 (11.1%)

Note: Values are given as n (%) except for age, which is given as mean ± SD. Motor complete = ASIA Impairment Scale A and B; motor incomplete = ASIA Impairment Scale C and D; SCI = spinal cord injury; WISCI-II = Walking Index for Spinal Cord Injury-II.

OEGT session frequency

The average number of OEGT sessions across inpatient and outpatient settings was approximately 19 for both motor complete and motor incomplete SCI groups spanning over an average of 17 to 18 weeks. Figure 1 illustrates the number of patients at each session count stratified by impairment group. Both groups completed a greater number of sessions in outpatient therapy than in inpatient rehabilitation (Table 2). Interestingly, following inpatient discharge, 22.6 ± 20.2 days (motor complete) and 52.1 ± 40.2 days (motor incomplete) elapsed before the first outpatient OEGT session.

Figure 1.

Overground exoskeleton gait training (OEGT) session count by setting and stratified by motor complete and motor incomplete spinal cord injury (SCI). For both SCI groups, a greater number of sessions were completed in the outpatient (OP) therapy setting than while in inpatient (IP) rehabilitation.

Table 2.

WISCI-II scores across inpatient and outpatient therapy

	n	Time since injury, weeks	LOS, weeks	No. of OEGT sessions	Admit WISCI-II	Discharge WISCI-II	Gain WISCI-II	Effect size (d) for gain
Inpatient
Complete	9	9.6 ± 20.2	6.8 ± 3.6	8.3 ± 5.2	0 ± 0	0.3 ± 1.0	0.3 ± 1.0	0.30
Incomplete	9	9.5 ± 10.2	6.3 ± 3.2	5.8 ± 3.4	0.2 ± 0.4	1.9 ± 3	1.7 ± 3.1	0.55
Outpatient
Complete	6	13.0 ± 7.5	9.9 ± 6.0	12.2 ± 9.3	0 ± 0	3.0 ± 4.6	3.0 ± 4.6	0.65
Incomplete	7	18.8 ± 10.8	13.0 ± 8.6	12.7 ± 7.3	4.0 ± 2.8	9.0 ± 6.4	5.0 ± 6.0	0.83

Note: Values are given as mean ± SD. Complete = motor complete injuries (ASIA Impairment Scale A and B); Incomplete = motor incomplete injuries (ASIA Impairment Scale C and D); LOS = length of stay; OEGT = overground exoskeleton gait training; WISCI-II = Walking Index for Spinal Cord Injury-II.

Up time, walk time, and step count

OEGT session walk time, up time, and step counts are summarized in Figure 2A for patients with motor complete SCI and Figure 2B for patients with motor incomplete SCI. Patients averaged 16:03 ± 7:41 minutes (motor complete) and 16:38 ± 4:51 minutes (motor incomplete) of up time and 7:51 ± 4:42 minutes (motor complete) and 12:16 ± 6:01 minutes (motor incomplete) of walk time during the first OEGT session. Walk time increased steadily through session 13 for both groups (motor complete, up time 29:49 ± 12:44 minutes, walk time 24:50 ± 9:35 minutes; motor incomplete, up time 23:06 ± 08:50, walk time 20:01 ± 08:05 minutes). Few patients had more than 19 sessions, resulting in highly variable estimates at higher session counts.

Figure 2.

Overground exoskeleton gait training (OEGT) session metrics for patients with (A) motor complete and (B) motor incomplete SCI. Across sessions, patients with spinal cord injury demonstrate upward trends in number of steps, walk time, and up time irrespective of severity of injury.

Walking function

WISCI-II scores were available for nine patients with motor complete SCI and nine patients with motor incomplete SCI at inpatient admission and discharge, and six patients with motor complete SCI and seven patients with motor incomplete SCI at outpatient admission and discharge (Table 2). At inpatient admission, patients were unable to stand or walk, scoring an average of 0 ± 0 (motor complete) and 0.2 ± 0.4 (motor incomplete). Discharge scores represent limited gains in walking function by the end of inpatient rehabilitation. Following outpatient therapy, patients with motor complete SCI were able to walk in parallel bars with bracing and assistance of one person (WISCI-II level 3), and patients with motor incomplete SCI were able to walk with a walker and braces with no physical assistance (WISCI-II level 9). Cohen's d effect sizes for gain in WISCI-II scores during inpatient rehabilitation shows a small effect for complete and medium effect for incomplete injuries. During outpatient therapy, medium and large effects were seen for complete and incomplete injuries, respectively. Figures 3A and 3B highlight the walking function over time as expressed by the WISCI-II scores for motor complete (A) and motor incomplete (B) SCI.

Figure 3.

Trajectory of individual Walking Index for Spinal Cord Injury-II (WISCI-II) scores by patient from time since injury at four timepoints: admission/discharge from inpatient rehabilitation and admission/discharge from outpatient therapy. (A) Patients with motor complete spinal cord injury (SCI) (n = 9) remained at a WISCI-II = 0 throughout. Exceptions were observed in two patients who experienced a conversion from motor complete to motor incomplete by outpatient discharge. (B) Patients with motor incomplete SCI (n = 9) demonstrate varying gains in WISCI-II scores across the inpatient to outpatient continuum.

Close inspection of the motor complete SCI group revealed that improvement in walking function occurred in only three individuals. One patient was diagnosed as T12 AIS A with a zone of partial preservation to L1 upon admission to inpatient rehabilitation. This individual was able to walk with bracing and therapist assistance at discharge from inpatient rehabilitation but did not progress beyond requiring this level of assistance during outpatient therapy. Two patients diagnosed as motor complete at inpatient rehabilitation admission (T11 AIS B and C7 AIS B) converted to motor incomplete (AIS C) as they transitioned to the outpatient therapy setting.

Discussion

This retrospective study describes OEGT initiated in inpatient rehabilitation and continued into outpatient therapy settings following SCI and its impact on clinical outcomes. Importantly, we are interested in the delivery of OEGT initiated early in the rehabilitation process, a critical window for recovery,^22,23 and continued as a single intervention into outpatient therapy. Given the brevity of inpatient rehabilitation lengths of stay,² identifying interventions delivered across the continuum may be necessary to maximize dosage and neuroplastic potential following SCI.^22,29 Our findings suggest OEGT can be completed across the continuum of care for patients with SCI. However, our retrospective review identified several factors that appear to impact the delivery of OEGT during inpatient and outpatient settings. In general, patients completed a greater number of OEGT sessions in the outpatient therapy setting than during inpatient rehabilitation. This finding is consistent with others indicating patients receive a greater number of hours of postdischarge therapy than in inpatient rehabilitation.¹ More specifically, as it relates to OEGT, fewer sessions logged during inpatient stay may be due to medical concerns limiting patient readiness to participate in OEGT acutely following SCI.¹⁵ In this study, the average time from inpatient admission to the first OEGT session ranged from 21 to 32 days for motor incomplete and motor complete SCI, respectively, suggesting those with motor incomplete injuries may experience fewer medical concerns limiting readiness to initiate OEGT.

Furthermore, the transition from inpatient rehabilitation to outpatient therapy services requires a coordinated effort to deliver a uniform OEGT intervention. Despite the OEGT intervention delivered within the same building, the unique settings (inpatient vs. outpatient) within our rehabilitation hospital have differing hospital leadership, electronic medical records, billing and insurance demands, and therapy staff delivering OEGT. At our facility, a clinical liaison communicates between the patient and the inpatient and outpatient clinical teams to promote a seamless transition of care across settings. Additionally, inpatient therapists communicate directly with outpatient therapists concerning patient measurements, device settings, and OEGT status at the outset of outpatient therapy.

Although these processes were in place, we observed a delay between inpatient rehabilitation discharge and the first outpatient OEGT session of about 7 weeks. While inpatient to outpatient transition timelines have not been published, van den Berg-Emons et al. found that persons with traumatic SCI experienced a significant decline in physical activity level 2 months after discharge from inpatient rehabilitation.³⁰ The availability of individual postdischarge resources such as transportation and insurance likely affects the timing of the inpatient-to-outpatient transition.¹ Others may experience a delay in therapy services due to poor adjustment to injury and/or a lack of social support as they transition home.³¹ Although determination of whether or not the patient received therapy services during this 7-week transition period was not possible within the review of the medical record, some patients may have received other therapy services excluding OEGT (e.g., home health services, traditional outpatient therapy). Nonetheless, we observed a continued progression of OEGT session walk time, up time, and step counts from inpatient rehabilitation to outpatient therapy settings. Importantly, the continued progression of OEGT session characteristics demonstrates the capability of delivering a uniform intervention across continuum settings even with a gap in therapy service.

Despite inclusion of patients with a range of neurological levels and severity of injury, none of the current overground robotic exoskeleton literature describe OEGT use and clinical outcomes between motor complete and motor incomplete SCI.^6,14 Though recovery of walking is a priority among persons with SCI irrespective of severity of injury, time since injury, and age at time of injury,⁵ recovery of walking is less likely if the injury is complete.³² Thus, we anticipated there would be observed variations in OEGT session characteristics and walking function outcomes between motor complete and motor incomplete injuries. Indeed, we observed those with motor incomplete SCI to have greater up time, walk time, and step count during the first session than those with complete SCI, though both groups demonstrated a slow progression in subsequent OEGT sessions. Others have suggested those with motor complete SCI experience higher medical acuity and may account for some of the variance in OEGT session tolerance.¹⁵

As expected, we also observed a greater recovery of walking function for those with motor incomplete SCI compared to those with motor complete SCI. Given that a change of one WISCI-II level can be considered clinically meaningful,²⁸ our findings suggest patients with motor incomplete SCI achieved a meaningful improvement in walking function in both inpatient rehabilitation and outpatient therapy settings, with the greatest gain occurring in the outpatient setting. Specifically, our average patient with motor incomplete SCI was unable to walk at inpatient rehabilitation admission and progressed to walking in the parallel bars with bracing and the assistance of two therapists at inpatient rehabilitation discharge. By the time they discharged from outpatient therapy at approximately 9 months post-SCI, the patient could ambulate with bracing, using a walker, and without physical assistance from another person. Our findings are consistent with others who report motor incomplete patients with AIS C (and less than 50 years of age) to have a 75% rate of recovery of walking function in 1 year^33,34 and an excellent walking prognosis at 1 year postinjury for those with AIS D.³⁵ Furthermore, we observed improvement in walking function for those with motor complete SCI who converted to motor incomplete SCI over time. This finding is consistent with others who reported that some patients with motor complete SCI will convert to incomplete status within 1 year of injury.^33,36–38 For our study, two of eight patients (25%) with motor complete SCI converted to motor incomplete status by outpatient discharge at 23 weeks postinjury. Both patients were able to walk with a walker, with bracing, and with no assistance, which can be considered clinically and functionally meaningful.

Study Limitations

This study has several limitations. Notably, the retrospective observational design of this study reduces our ability to generalize our findings. Further, our review of OEGT sessions and outcomes was limited to the information found in the medical record. Aside from the WISCI-II, other outcomes of potential interest for recovery of mobility and function after SCI, including the 10-meter walk test (10MWT),³⁹ 6-minute walk test,^40,41 and Berg Balance Scale,⁴² were not consistently completed across inpatient rehabilitation and outpatient therapy settings, limiting the overall clarity in motor recovery after SCI. Others have suggested that a combination of the 10MWT and the WISCI-II provides the most valid measure of improvement in walking following SCI,⁴⁰ while consistent use of standardized outcomes across the continuum of care is recommended.^39–43 Also, the medical record did not contain information regarding the level of robotic assistance provided to the patients across sessions. Likewise, we did not have access to feedback provided to the patient regarding their performance or therapists' decisions regarding level of robotic assistance. As there may be an association between device-provided assistance, patient feedback, and level of recovery,⁴⁴ we recommend inclusion of these data in prospective research efforts. One constant across our rehabilitation hospital settings was the 45-minute duration for therapy sessions. As a result, there may be an artificially imposed ceiling effect in OEGT session metrics due to the therapy session time constraint.

Additionally, the small sample size and absence of a control group limit our ability to compare the effect of OEGT with other gait training interventions provided as usual care following SCI. Our aim was to generate a hypothesis regarding the capability of a large urban rehabilitation hospital to initiate OEGT early in rehabilitation and deliver a uniform intervention across a continuum of care. Nonetheless, our small sample size may limit our ability to draw firm conclusions concerning estimations of time since injury to the initiation of inpatient or outpatient therapy, incidence of patients participating in OEGT converting from complete to incomplete SCI, and the inability to capture reasons why others who completed OEGT in inpatient were unable to continue into outpatient. Lastly, our 3-year time period for data abstraction was December 2018 through December 2021. The timelines for admission, length of stay, and transition time from inpatient rehabilitation to outpatient therapy may have been impacted by the COVID-19 pandemic and related access to care issues.⁴⁵

Conclusion

The use of OEGT across the continuum of care was described at our large urban rehabilitation hospital. We observed our patients to demonstrate progressive tolerance for OEGT across rehabilitation settings despite anticipated differences in walking function outcomes between motor complete and motor incomplete SCI. Regardless of a time lapse between inpatient discharge and outpatient therapy, OEGT sessions progressively challenged patients across settings. Further, patients with motor incomplete SCI experienced clinically meaningful improvements in walking function over time with the greatest gains occurring in the outpatient setting. Future research may consider examining dosage parameters of an OEGT intervention beginning in inpatient rehabilitation and continuing in the outpatient therapy setting to drive optimal walking outcomes for persons with SCI.