An estimated 750,000 Americans experience a stroke annually. Most stroke survivors require rehabilitation. Limited access to rehabilitation facilities has a pronounced burden on functional outcomes and quality of life. Robotic devices deliver reproducible therapy without the need for real-time human oversight. This study examined the efficacy of using home-based, telerobotic-assisted devices (Hand and Foot Mentor: HM and FM) to improve functional ability and reduce depression symptoms, while improving access and cost savings associated with rehabilitation.
Twenty stroke survivors performed three months of home-based rehabilitation using a robotic device, while a therapist remotely monitored progress. Baseline and end of treatment function and depression symptoms were assessed. Satisfaction with the device and access to therapy were determined using qualitative surveys. Cost analysis was performed to compare home-based, robotic-assisted therapy to clinic-based physical therapy.
Compared to baseline, significant improvement in upper extremity function (30.06%, p= 0.046), clinically significant benefits in gait speed (29.03%), moderate improvement in depressive symptoms (28.44%) and modest improvement in distance walked (30.2%) were observed. Participants indicated satisfaction with the device. Home-based robot therapy expanded access to post-stroke rehabilitation for 35% of the people no longer receiving formal services and increased daily access for the remaining 65%, with a cost savings of $2,352 (64.97%) compared to clinic-based therapy.
Stroke survivors made significant clinically meaningful improvements in the use of their impaired extremities using a robotic device in the home. Home-based, robotic therapy reduced costs, while expanding access to a rehabilitation modality for people who would not otherwise have received care.
Stroke is one of the leading causes of long-term disability  with an estimated 795,000 incidences of stroke in the United States annually . Of the approximate 665,000 survivors , 80% experience moderate to severe upper extremity (UE) impairments  and two-thirds experience lower extremity (LE) impairments . Most require long-term rehabilitation to regain functional capacities required to perform activities of daily living and ambulation. This represents an approximately $34 billion cost to the healthcare system with an estimated average yearly rehabilitation cost of $11,689 per stroke survivor following acute and subacute rehabilitation discharge . This substantial burden to the healthcare system has emphasized the need to investigate opportunities to improve care for stroke survivors while reducing mounting costs.
To date, best practice for successful rehabilitation often involves intensive, repetitive practice that actively engages the participant in goal-oriented and task-specific activities to regain functional capacities in upper and lower extremities . Unfortunately, the quality of stroke services for rural patients is suboptimal and limited access to rehabilitation facilities has a pronounced burden on functional outcomes and quality of life. A recent study demonstrated, using logistic modeling, that rural stroke survivors were less likely to receive stroke rehabilitation therapy than their urban counterparts . Moreover, with the prevalence of stroke being predicted to increase by almost 25% by 2030  and rural populations being identified as being particularly vulnerable to stroke , there is a great need to develop accessible, cost effective therapy to minimize functional disability and optimize functional motor recovery for rural stroke survivors.
Robot-assisted therapy is a promising option for improving voluntary upper extremity (UE) movement in stroke survivors with finite access to conventional therapy [11–13]. Additionally, several recent studies have concluded that robotic assisted therapy improves lower extremity (LE) strength and locomotor function [14,15]. Recent advances in robot-assisted therapy have greatly increased the level of function patients can achieve. Successful rehabilitation techniques involve highly intensive, repetitious practice that actively engages the participant in goal-oriented and task-specific activities. Many studies have observed that home-based, robotic-assisted therapy demonstrate equivalent outcomes compared to one-on-one therapeutic delivery [11,16,17]. The results of these studies indicate that robot-assisted therapy provides reliable, reproducible treatment while measuring performance without the need for real-time human oversight .
Although the goals of using robotic assistive devices are to improve active range of motion (AROM), strength, and function in the distal musculature of stroke survivors is promising, these modalities are underutilized in the home. Therefore, combining telemedicine with in-home robot-assisted therapy (telerehabilitation) for people with residual impairment following stroke has the potential to reduce barriers while proving cost-effective, consistently high-quality treatment to patients with limited access to rehabilitation clinics because of location or availability of treatment modalities . This relatively new idea of telerehabilitation is defined as the provision of rehabilitation services at distance using information and communication technologies . A recent systematic review examining studies published after 2000 cites positive outcomes for patients and caregivers who have utilized telerehabilitation . Additionally, caregivers and patients report high levels of overall satisfaction and acceptance of telerehabilitation interventions. Recently, a prospective, single blinded, multisite, randomized controlled trial successfully paired robot-assisted therapy and a telerehabilitation intervention . Equivalent outcomes were observed in the dose-equivalent robot-assisted therapy group and the usual and customary care group. Further, specific evaluation of LE robotic intervention found that 12 weeks of home-based rehabilitation elicited improvements in locomotor function and strength . Recently, preliminary data investigating robotic telerehabilitation in stroke survivor’s homes, reported improvements in residual upper and lower limb impairments, while reductions in the cost of care decreased the burden on the healthcare system . However, to date, a paucity of evidence regarding the efficacy and cost effectiveness of telerehabilitation interventions in rural stroke survivors persists and a knowledge gap exists as to what effect participation in telerehabilitation has on utilization of available therapy.
This study aimed to examine the efficacy of using a home-based, tele robotic-assisted device to: improve functional ability, reduce depression symptoms, and create a satisfactory experience, increase access to, and monitor participant utilization of cost efficient rehabilitation when compared to the cost of clinic-based therapy for rural stroke survivors.
Figure 1 shows the flow of the participants through each stage of the study. Twenty, mostly rural and highly rural Veteran stroke survivors (67.0±11.4) years old at enrollment) with UE hemiparesis resulting from unilateral stroke (mean time since stroke of 20.4±9.26) months) met inclusion criteria and were enrolled in this study. All 20 participants showed UE or LE impairment secondary to corticospinal tract infarcts. Overall the home-based, telerehabilitation therapy was safe and well tolerated. No adverse events occurred. Participant demographic and geographic location information is presented in Table 1. Nineteen participants completed the study, consisting of three months of home-based, telerehabilitation. One participant in the HM group dropped out after device deployment due to medical reasons unrelated to the study. This participant did not complete post-intervention assessments and the data was excluded from all analyses. Means and SD for clinical outcome measures for participants at baseline and postintervention are presented in Table 2.
|Robotic-Assisted Therapy||Upper extremity, n=10||Lower extremity, n=10||Total, n=20|
|Male, gender, n (%)||9 (90)||10 (100)||19 (95)|
|Mean time since stroke, months (range)||21.5 (1.8–136.7)||19.4 (3.5–67.7)||20.4 (1.8–136.7)|
|Mean age (years) at enrollment, mean (SD)||63.4 (9.1)||70.6 (12.7)||67 (11.4)|
|Dominant Right hand/foot, n (%)||10 (100)||10 (100)||20 (100)|
|Right side affected, n (%)||6 (60)||4 (40)||10 (50)|
|Withdrew after baseline visit||1||0||1|
|Distance from VAMC miles (SD)||76.08||(35.0)||53.64||(35.5)||63.32||(35.8)|
|Time device in home in days (SD)||119.6||(39.4)||105.4||(43.3)||112.9||(40.8)|
|Outcome Measure||Baseline||Post-intervention||Mean difference (%)||95% CI||Sig, p|
|FIM||97.67||96.13||−1.53 (1.57)||−8.45 to 5.38||0.642|
|CES-D||12.41||8.88||−3.53 (28.44)||−8.61 to 1.55||0.160|
|10 meter walk test (m/s)||0.31||0.40||0.09 (29.03) ŧ||−0.11 to 0.29||0.197|
|6 minute walk test (m)||59.70||77.73||18.03 (30.20)||−12.06 to 48.13||0.153|
|ARAT (unaffected)||50.22||55.78||5.56 (11.07)||−5.89 to 17.00||0.295|
|ARAT (affected)||30.67||39.89||9.22 (30.06) ŧ||0.38 to 18.07||0.046*|
Compared to baseline, affected UE ARAT scores improved by an average of 9.22 (26.13) points. The observed change represents a statistically (30.06%, p= 0.046) significant improvement in upper extremity function. Further, the average improvement surpassed the previously validated MCID of 5.7  indicating a clinically meaningful change in upper extremity function.
On average, participants using the FM device demonstrated a 29.03% increase in gait speed, from 0.31 to 0.40m/s. The observed change in gait speed represents a small but clinically significant improvement (0.09m/s, p=0.197). Based on previously validated stratification on gait speed and predicted functional walking capacity , the changes observed following the home-based telerehabilitation demonstrates the stroke survivors improved from home ambulators to limited community ambulators.
For participants who used the FM device, modest, non-significant (30.2%, p=0.153) improvements in total distance walked were observed from baseline (59.7±51.0) to post-intervention (77.7±32.9).
Self reported depressive symptoms over the past week, as assessed by the CES-D, demonstrated a moderate (28.44%, p=0.160) decrease from baseline (12.4±14.2) to post-intervention (8.9±9.3), indicating a lower prevalence of depressive symptoms at the final assessment.
Across all participants using either the FM or HM, no change (p=0.642) in FIM scores was observed from baseline (97.7) to post-intervention (96.1).
Usage and utilization
Summaries of device exposure, usage, and utilization are presented in Table 3. Throughout the three-month study period, participants were given access to either the HM or FM device for an average of 106 days (exposure). Participants completed an average of 30.6 training sessions throughout the study. However, relatively wide individual variability was observed, with participants using the device as little as two and as many as 75 sessions. For each daily usage, participants performed robotic telerehabilitation an average of 90.6 min (range 15–153.6 min). Given the above information, participants utilized either the HM or the FM devices an average of 29.3% of the days the device was in the home. Large heterogeneity in device utilization was observed, ranging from 2.75% to 76.53%.
|ID||Mean daily therapy minutes||Usage (sessions)||Exposure (days)||Utilization|
Following completion of the intervention, 17 participants responded to 15 statements to measure satisfaction with the HM or FM device (Table 4). Overall, participants indicated satisfaction with the device and their overall improvement. For most of the statements, a response of “6” is the most positive response; however for four of the statements a response of “0” is the most positive (strongly disagreeing with the statement). Subjects’ responses on these four questions indicated that they either slightly or strongly disagreed with negative statements about the device. For the positive statements, participants indicated that they either slightly or strongly agreed (mean≥4.0). Participants generally disagreed with the negative statements (mean ≤ 2.0), however the mean response to the statement concerning donning and doffing the peripheral hand or foot piece was 3.5, indicating that participants had difficulty with this aspect of using the HM or FM.
|1. The instructions for using the Hand or Foot Mentor were clear and easy to understand||5.94 (0.25)||99%|
|2. This therapy was relevant to my rehabilitation||5.19||(0.83)||91%|
|3. My function was improved||4.44||(1.79)||82%|
|4. The games were appropriate||5.16||(1.23)||86%|
|5. The games were hard to see||1.09||(2.03)||27%|
|6. This therapy challenged me||4.81||(1.68)||74%|
|7. This therapy was too difficult||1.14||(1.56)||9%|
|8. This therapy was too easy||2.00||(1.96)||31%|
|9. The device was user-friendly||4.86||(1.70)||85%|
|10. I got bored with the games||2.00||(2.42)||39%|
|11. I enjoyed playing the games||5.43||(1.02)||91%|
|12. The pace of therapy was just right||5.00||(1.57)||85%|
|13. I had trouble donning and doffing the Hand or Foot piece||3.53||(2.22)||61%|
|14. Overall, I am satisfied with the progress I made using the Hand/Foot Mentor||5.09||(1.21)||89%|
|15. Therapy with the Hand/Foot Mentor met my expectations||5.13||(1.31)||91%|
Following the study therapist phone or in-person survey, it was determined that 7 (35%) participants were no longer receiving formal physical or occupational therapy services. The remaining 13 (65%) participants were receiving some form of formal services. Eight (40%) participants were receiving outpatient therapy services, of which, three (15%) sought out private services outside of the VA health system. The remaining five (25%) participants received home health care physical or occupational therapy. Participants reported between utilizing formal therapy services between two times per month and two times per week. For all participants mean access to therapy services was estimated as 1.2 session per week or 14.7 sessions for the duration of the study (assuming four weeks per month). Therefore, when considering the device usage data (Table 4), the home-based, robotic therapy intervention increased or extended access for all participants. Exploratory analysis to assess the affect of the home-based, telerehabilitation on those participants who are already receiving formal services demonstrates that the present study increases access from 22.6 sessions to 53.2 sessions, representing a 135.4% increase in rehabilitation exposure.
Total home-based, telerehabilitation costs were calculated based on the cost of equipment; device maintenance and data connection; home delivery, support and pickup; including weekly clinician follow-up and monitoring. These costs were compared against clinical based outpatient therapy at the VA medical center defined as three one-hour sessions per week for 90 days. The estimated monthly equipment cost of $82.14 and $54.76 for the foot and hand robotic devices, respectively, was amortized over an expected five-year lifetime (Table 5). This was combined with the cost of an annual maintenance and hosting contract of $1,199 per device to cover device repairs and the data connection and online portal for therapist monitoring of patient progress. Corresponding therapist transit costs were calculated based on a study average round-trip distance of 164 (48.68) miles using the standard government mileage reimbursement rate of $0.415 per mile (43) and 144.8 (39.7) minutes. The therapist costs of $65.23 an hour consisted of the cost of time to deploy a two-person team to the participant homes for device installation and orientation, repairs or device reorientation due to user error, device pickup, and weekly therapist monitoring and telephone based follow-up calls.
|FM||HM||Average of HM and FM||Outpatient Therapy (Projected)||Average Savings of RAT|
|Cost of Robotic Device||$82.14||$54.76||$68.45||–|
|Device Maintenance and Hosting||$99.92||$99.92||$99.92||–|
Projected outpatient therapy transit and therapist costs for three one-hour sessions held weekly at the Atlanta VA medical center were based on the mean distance from a participating VA clinic (63.3±35.8 miles) and patient mileage reimbursements and the projected cost of a physical therapist at $38.56 taken from the State of Georgia Bureau of Labor Statistics .
Final cost analysis reveals three months of home-based, telerehabilitation costs the VA health care system an average of $1,268.07 per Veteran, compared to an average of $3,619.95 per Veteran for outpatient clinic based therapy. This analysis reveals an average of $2,352 (64.97%) in savings compared to clinic-based therapy per stroke survivor. Further, the inclusion of home-based telerehabilitation leads to a return of approximately $2.85 of therapy on every dollar spent by the VA health system.
Download Full PDF