Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Walking once again : use of a robotic exoskeleton during subacute stroke rehabilitation to promote functional… Louie, Dennis Riley 2021

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata


24-ubc_2021_may_louie_dennis.pdf [ 2.59MB ]
JSON: 24-1.0396078.json
JSON-LD: 24-1.0396078-ld.json
RDF/XML (Pretty): 24-1.0396078-rdf.xml
RDF/JSON: 24-1.0396078-rdf.json
Turtle: 24-1.0396078-turtle.txt
N-Triples: 24-1.0396078-rdf-ntriples.txt
Original Record: 24-1.0396078-source.json
Full Text

Full Text

WALKING ONCE AGAIN: USE OF A ROBOTIC EXOSKELETON DURING SUBACUTE STROKE REHABILITATION TO PROMOTE FUNCTIONAL RECOVERY by  Dennis Riley Louie  B.Sc., The University of Toronto, 2009 M.Sc.PT., The University of Toronto, 2011  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Rehabilitation Sciences)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  March 2021  © Dennis Riley Louie, 2021   ii The following individuals certify that they have read, and recommend to the Faculty of Graduate and Postdoctoral Studies for acceptance, the dissertation entitled: Walking once again: Use of a robotic exoskeleton during subacute stroke rehabilitation to promote functional recovery  submitted by Dennis Riley Louie in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Rehabilitation Sciences  Examining Committee: Dr. Janice Eng, Department of Physical Therapy Supervisor  Dr. W. Ben Mortenson, Department of Occupational Science & Occupational Therapy Supervisory Committee Member  Dr. Jennifer Yao, Department of Medicine – Physical Medicine & Rehabilitation Supervisory Committee Member Dr. Kristin Campbell, Department of Physical Therapy University Examiner Dr. Leanne Currie, School of Nursing University Examiner     iii Abstract Stroke is a leading cause of adult disability, typically resulting in mental and physical impairments that can affect functional abilities such as walking. Recovering the ability to walk is an important goal for stroke survivors, having implications for long-term health and functional outcomes. Population trends and new medical interventions have resulted in changes in mortality and disablement, warranting an update of the prevalence of leg and walking impairment after stroke; interventions to optimize walking recovery after stroke also remain a key priority for research. Powered robotic exoskeletons, a new generation of electromechanical devices that can support and move the lower limbs to walk overground, may be a novel intervention to achieve higher-intensity walking practice and rehabilitation gains for patients after stroke. The aims of this dissertation were to characterize the current state of leg and walking impairment after stroke, as well as to investigate the role of exoskeletons, both their efficacy and acceptability, in stroke rehabilitation. A cohort study first established that approximately half of patients surviving a first-ever stroke are unable to walk, affecting their discharge disposition after acute hospitalization. The second study, a scoping review, mapped the existing research surrounding the use of powered exoskeletons for overground gait retraining and highlighted a need for more rigorous trials in the subacute stroke population. The third and fourth studies were conducted concurrently as a mixed methods trial to investigate the use of powered exoskeletons for gait retraining with non-ambulatory patients during subacute stroke rehabilitation. In the randomized controlled trial, no greater benefit was achieved by using a powered exoskeleton for walking recovery compared to standard physical therapy care. However, the qualitative component, an interview-based study exploring how patients and therapists react to using an exoskeleton for intensive rehabilitation, revealed that the technology is viewed highly favorably.    iv In summary, these studies, which focused on non-ambulatory patients, indicate an ongoing need to target walking recovery after stroke and suggest that powered exoskeletons are a welcome option to achieve repetitive walking practice. Further work is needed to clarify which patients will truly benefit, and to what extent, from this intervention.       v Lay Summary Stroke is a leading cause of adult disability in Canada, often resulting in muscle weakness and walking impairments. Robotic exoskeletons, motorized bracing systems that automate walking, may allow greater walking practice and rehabilitation gains after stroke. The first study of this thesis provided an update of leg weakness and walking difficulty immediately after stroke in a large Canadian hospital, highlighting the importance of walking ability for returning home. The second study summarized the available research on using exoskeletons for stroke, finding a scarcity of studies addressing early rehabilitation. The third study was a national, multi-site randomized clinical trial, which found no difference between exoskeleton-based and conventional physical therapy during early stroke rehabilitation. However, interviews conducted for the fourth study found that patients and therapists appreciated the opportunity to practice walking using an exoskeleton. Overall, this thesis provides greater understanding of the potential use of robotics for walking rehabilitation after stroke.    vi Preface This statement confirms that this thesis was written and compiled by Dennis Riley Louie (DRL). The co-authors of the manuscripts, including Dr. Janice J. Eng (JJE), Dr. W. Ben Mortenson (WBM), Dr. Jennifer Yao (JY), Dr. Robert Teasell (RT), Dr.Thalia Fields (TF), Dr. Amy Schneeberg (AS), Melanie Durocher (MD), Lisa Simpson (LS), and Michelle Lui (ML) made contributions commensurate with their supervisory committee or co-author duties.   The research in Chapter 2 was conducted using data collected with the approval of the University of British Columbia Clinical Research Ethics Board (H10-03180 / H18-01105). DRL and JJE developed the research question for the study presented in Chapter 2. DRL and LS were involved in collecting the data. DRL conducted the data analysis and wrote the manuscript. TF and JY provided scientific and clinical input throughout the process. All listed authors provided input and critical review before approving the final manuscript. A version of Chapter 2 has been submitted for publication: Louie DR, Simpson LA, Field TS, Mortenson WB, Yao J, Eng JJ. Prevalence of walking impairments after acute stroke and its impact on home discharge (submitted).  Chapter 3 is based on work conducted in the Rehabilitation Research Program and Laboratory (GF Strong Rehabilitation Centre, Vancouver, Canada). DRL and JJE formulated the research question together. DRL was responsible for designing and performing the search strategy, reviewing the articles, synthesizing the data, and drafting the manuscript. Christina Cassady (CC) acted as a second reviewer (acknowledged in manuscript). JJE edited and provided critical review of the manuscript. A version of Chapter 3 has been published: Louie DR, Eng JJ. Powered robotic exoskeletons in post-stroke rehabilitation of gait: a scoping review. J Neuroeng Rehabil. 2016;13(1):53 (doi: 10.1186/s12984-016-0162-5).    vii The research in Chapter 4 and 5 was conducted at three rehabilitation facilities in Canada, under the University of British Columbia Clinical Research Ethics Board (H15-01339), the University of Alberta Health Research Ethics Board (Pro00071806), and the Western University Health Sciences Research Ethics Board (108618). JJE (principal investigator), WBM, and JY conceived the original research questions and study design in order to secure funding from the Heart and Stroke Foundation to carry out the two studies (mixed methods trial). DRL subsequently refined the study design and developed the study protocols for the multi-site trial . For Chapter 4, DRL was responsible for overseeing the overall multi-site trial ( NCT02995265), including obtaining research ethics and operational approval for each site, as well training of all study assessors, coordinators, and intervention therapists. DRL coordinated the study in British Columbia, MD coordinated the study in Alberta, and RT coordinated the study in Ontario. DRL performed all statistical analyses, in consultation with AS, and wrote the first draft of the manuscript. All listed authors provided input and critical review before approving the final manuscript. A version of Chapter 4 has been submitted for publication: Louie DR, Mortenson WB, Durocher M, Schneeberg A, Teasell R, Yao J, Eng JJ. Efficacy of an exoskeleton-based physical therapy program for non-ambulatory patients during subacute stroke rehabilitation: a randomized controlled multi-site trial (submitted).  The research in Chapter 5 was conducted under the same research ethics approvals as Chapter 4. As mentioned, JJE, WBM, and JY conceived the research question. DRL further refined the research question and developed the final interview guides for data collection. DRL was responsible for overseeing the overall study and recruited participants in British Columbia. MD and RT provided participants from their respective sites. DRL conducted all interviews.   viii DRL, WBM, and ML analyzed the transcribed data together. DRL wrote the first draft of the manuscript. All listed authors provided input and critical review before approving the final manuscript. A version of Chapter 5 has been submitted for publication, and is currently under review: Louie DR, Mortenson WB, Lui M, Durocher M, Teasell R, Yao J, Eng JJ. Patients’ and therapists’ experience and perception of exoskeleton-based physiotherapy during subacute stroke rehabilitation: a qualitative analysis (under review).      ix Table of Contents Abstract ......................................................................................................................................... iii Lay Summary .................................................................................................................................v Preface ........................................................................................................................................... vi Table of Contents ......................................................................................................................... ix List of Tables ................................................................................................................................xv List of Figures ............................................................................................................................. xvi List of Abbreviations ................................................................................................................ xvii Acknowledgements .................................................................................................................... xix Dedication ................................................................................................................................... xxi Chapter 1: Introduction ................................................................................................................1 1.1 What is a stroke? ............................................................................................................. 1 1.1.1 Neurophysiology of stroke ...................................................................................... 1 1.1.2 Presentation of stroke .............................................................................................. 3 1.1.3 Stroke epidemiology ............................................................................................... 5 1.2 Recovery after stroke ...................................................................................................... 7 1.2.1 Phases of stroke recovery........................................................................................ 8 1.2.2 Organization of care ................................................................................................ 9 1.3 Stroke Rehabilitation .................................................................................................... 11 1.3.1 Rehabilitation of walking ...................................................................................... 11 Epidemiology of lower extremity impairment .................................................. 12 Epidemiology of walking limitations................................................................ 12 1.3.2 The sensorimotor integration theory of motor recovery ....................................... 13   x 1.3.3 Repetitive task-specific practice for optimal recovery ......................................... 15 1.4 Role of technology in stroke rehabilitation ................................................................... 17 1.4.1 Electromechanical interventions for gait rehabilitation ........................................ 18 1.4.2 Powered robotic exoskeletons............................................................................... 20 Potential benefit of exoskeletons in stroke rehabilitation ................................. 21 1.4.3 Adoption of technology ........................................................................................ 24 Perceptions of technology in stroke rehabilitation............................................ 25 1.5 Dissertation objectives .................................................................................................. 25 Chapter 2: Prevalence of walking impairments after acute stroke and its impact on home discharge .......................................................................................................................................27 2.1 Introduction ................................................................................................................... 27 2.2 Methods......................................................................................................................... 29 2.2.1 Participants ............................................................................................................ 29 2.2.2 Data collection and variables ................................................................................ 30 Primary objective: leg and walking impairment after stroke ............................ 30 Secondary objective: predicting home discharge ............................................. 31 2.2.3 Statistical analysis ................................................................................................. 31 2.3 Results ........................................................................................................................... 32 2.3.1 Leg and walking impairment after stroke ............................................................. 33 2.3.2 Predicting home discharge .................................................................................... 35 2.4 Discussion ..................................................................................................................... 38 2.4.1 Limitations ............................................................................................................ 41 2.4.2 Conclusion ............................................................................................................ 42   xi Bridging Statement I....................................................................................................................43 Chapter 3: Powered robotic exoskeletons in post-stroke rehabilitation of gait: a scoping review ............................................................................................................................................44 3.1 Introduction ................................................................................................................... 44 3.2 Methods......................................................................................................................... 47 3.2.1 Summary of search strategy .................................................................................. 47 3.2.2 Inclusion and exclusion criteria ............................................................................ 48 3.2.3 Study selection and data extraction....................................................................... 48 3.3 Results ........................................................................................................................... 49 3.3.1 Study design .......................................................................................................... 50 3.3.2 Participants ............................................................................................................ 51 3.3.3 Exoskeletons ......................................................................................................... 57 3.3.4 Training Period ..................................................................................................... 59 3.3.5 Training Protocol .................................................................................................. 59 3.3.6 Walking measures ................................................................................................. 60 3.3.7 Effectiveness of exoskeleton-based gait training .................................................. 60 3.3.8 Adverse effects...................................................................................................... 62 3.4 Discussion ..................................................................................................................... 62 3.4.1 Future directions for research and suggestions for clinical practice ..................... 64 3.4.2 Limitations ............................................................................................................ 65 3.4.3 Conclusion ............................................................................................................ 66 Bridging Statement II ..................................................................................................................67   xii Chapter 4: Efficacy of an exoskeleton-based physical therapy program for non-ambulatory patients during subacute stroke rehabilitation: a randomized controlled multi-site trial ....69 4.1 Introduction ................................................................................................................... 69 4.2 Methods......................................................................................................................... 71 4.2.1 Participants ............................................................................................................ 72 4.2.2 Exoskeleton device ............................................................................................... 73 4.2.3 Interventions ......................................................................................................... 73 4.2.4 Outcomes .............................................................................................................. 75 4.2.5 Statistical analysis ................................................................................................. 76 4.3 Results ........................................................................................................................... 77 4.4 Discussion ..................................................................................................................... 83 4.4.1 Limitations ............................................................................................................ 88 4.4.2 Conclusion ............................................................................................................ 89 Bridging Statement III ................................................................................................................90 Chapter 5: Patients’ and therapists’ experience and perception of exoskeleton-based physical therapy during subacute rehabilitation: a qualitative analysis ................................91 5.1 Introduction ................................................................................................................... 91 5.2 Methods......................................................................................................................... 93 5.2.1 Design ................................................................................................................... 93 5.2.2 Participants ............................................................................................................ 93 5.2.3 Data collection ...................................................................................................... 94 5.2.4 Data analysis ......................................................................................................... 95 5.2.5 Research characteristics ........................................................................................ 96   xiii 5.2.6 Techniques to enhance trustworthiness ................................................................ 96 5.3 Findings......................................................................................................................... 97 5.3.1 Theme 1: A matter of getting into the swing of things ......................................... 99 5.3.2 Theme 2: More of a positive experience than anything else .............................. 101 5.3.3 Theme 3: The best step forward.......................................................................... 104 5.4 Discussion ................................................................................................................... 107 5.4.1 Limitations .......................................................................................................... 109 5.4.2 Conclusion .......................................................................................................... 110 Chapter 6: Overall discussion, future directions, and conclusions .......................................111 6.1 Summary of study findings ......................................................................................... 112 6.2 Integration and contribution of dissertation to current research ................................. 113 6.3 Future research directions and recommendations ....................................................... 124 6.4 Personal reflection statement ...................................................................................... 130 6.5 Conclusion .................................................................................................................. 132 References ...................................................................................................................................133 Appendices ..................................................................................................................................172 Appendix A Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) Statement.............................................................................................................. 172 Appendix B Sample R script of analysis from Chapter 2 ....................................................... 174 Appendix C Reporting guidelines for Chapter 4 .................................................................... 180 C.1 CONSORT 2010 checklist of information to include when reporting a randomized trial 180 C.2 The TIDieR (Template for Intervention Description and Replication) Checklist .. 182   xiv Appendix D Guidelines for progressing participant in Exoskeleton Group ........................... 184 D.1 Definition of EksoGT settings: ............................................................................... 184 D.2 Progression of settings to create best quality, fully-assisted gait in exoskeleton: .. 185 D.3 Algorithm to increase active participation and challenge while in exoskeleton ..... 186 D.4 Progression of time spent walking while standing in exoskeleton ......................... 187 D.5 Algorithm to continue or discontinue daily exoskeleton training ........................... 188 Appendix E Data collection schedule and outcome measures used in Chapter 4 .................. 189 E.1 Schedule of data collection ..................................................................................... 189 E.2 Outcome measures .................................................................................................. 189 Appendix F Data analysis script, and results from per-protocol and sensitivity analyses ...... 192 F.1 Sample R script of analysis ..................................................................................... 192 F.2 Sensitivity analysis.................................................................................................. 196 F.3 Per-protocol analysis ............................................................................................... 198 Appendix G Standards for Reporting Qualitative Research (SRQR) ..................................... 200 Appendix H Interview guides ................................................................................................. 202 H.1 Patient Interview Guide .......................................................................................... 202 H.2 Physical Therapist Interview Guide ........................................................................ 203 Appendix I Study findings from the ExStRA interviews (1-page summary for participants) 205 Appendix J Sub-themes within each theme ............................................................................ 206    xv List of Tables Table 2.1 Patient characteristics in relation to binary ability to walk in the total cohort ............. 34 Table 2.2 Walking ability of whole sample (n = 487) .................................................................. 35 Table 2.3 Association of patient characteristics with home discharge ......................................... 36 Table 2.4 Final multivariate regression models for discharge home (vs. other institutions) ........ 37 Table 3.1 Meaningful change values for functional walking outcomes in stroke ........................ 49 Table 3.2 Summary of studies included in the review .................................................................. 52 Table 3.3 Details of powered exoskeletons in this review ............................................................ 58 Table 4.1 Demographic characteristics ......................................................................................... 79 Table 4.2 Primary outcome analysis ............................................................................................. 80 Table 4.3 Secondary walking outcomes ....................................................................................... 81 Table 4.4 Secondary outcomes of impairment, balance, mood, cognition, and quality of life .... 81 Table 4.5 Per-protocol analysis of primary and secondary walking outcomes ............................ 83 Table 5.1 Participant characteristics ............................................................................................. 98 Table 5.2 A list of device critiques ............................................................................................. 103 Table 5.3 Barriers and solutions to implementing an exoskeleton-based training program in stroke rehabilitation .................................................................................................................... 105 Table 5.4 Characteristics to guide patient selection for exoskeleton-based rehabilitation ......... 106    xvi List of Figures Figure 1.1 Illustration of the sensorimotor integration theory of motor recovery ........................ 15 Figure 1.2 Powered exoskeletons and the sensorimotor integration theory ................................. 22 Figure 2.1 Receiver operating characteristic curves for Model A and Model B .......................... 38 Figure 3.1 Flow diagram of study selection process..................................................................... 50 Figure 4.1 CONSORT flow diagram of study participants .......................................................... 78    xvii List of Abbreviations 5MWT 5-Metre Walk Test 6MWT 6-Minute Walk Test 95% CI 95% confidence interval ANCOVA analysis of covariance AUC  area under the curve BBS  Berg Balance Scale BWSTT body weight-supported treadmill training CONSORT Consolidated Standards of Reporting Trials ExStRA Exoskeleton for post-Stroke Recovery of Ambulation (trial) FAC  Functional Ambulation Category FMA-LE Fugl-Meyer Assessment, Lower Extremity motor score FIM  Functional Independence Measure HAL  Hybrid Assistive Limb (exoskeleton) ICH  intracerebral hemorrhage IQR  interquartile range MCID  minimal clinically important difference MoCA  Montreal Cognitive Assessment NIHSS  National Institutes of Health Stroke Scale OR  odds ratio PHQ-9  Patient Health Questionnaire “Physio” Physical therapy/therapist RCT  randomized controlled trial   xviii ROC  receiver operating characteristic (curve) SD  standard deviation SF-36-MCS 36-Item Short Form Survey (mental component summary) SF-36-PCS 36-Item Short Form Survey (physical component summary) SRQR  Standards for Reporting Qualitative Research STROBE Strengthening the Reporting of Observational studies in Epidemiology (guideline) TIA  transient ischemic attack  TIDieR  Template for Intervention Description and Replication (guideline)  tPA  tissue plasminogen activator    xix Acknowledgements To my supervisor, Dr. Janice Eng, thank you for all of the support and guidance you have provided over the past six years. Your mentorship has and will have a lasting impact on me. Between your wisdom, patience, demeanour, and drive, I could not have asked for a better supervisor to work with and learn from. Throughout my studies, you have presented me with numerous opportunities to develop, not just as a researcher, but as a leader and educator. I have felt truly supported throughout this process, and I am sincerely grateful.  To my committee members, Dr. Ben Mortenson and Dr. Jennifer Yao, thank you for all of the support and expertise you have provided in my journey as a doctoral student and growth as a clinician-scientist. You have both been so approachable and knowledgeable, offering much-needed insight into both the rigour and applicability of this research. I am very appreciative to have had you both on my supervisory committee. This work would not have been possible without the many individuals with stroke who participated in this research. I offer my deepest respect and gratitude to these survivors, who were committed to research participation in a time of radical life changes. I also wish to thank the physical therapists who gave their time and support, either as participants themselves, or as assessors and interventionists. I am also thankful for the additional research and hospital staff, who worked hard to support this research at each site, as well as the co-authors who each contributed their knowledge, efforts, and expertise.   I am also grateful to everyone in the Rehabilitation Research Laboratory, for it is a bustling community of bright minds and hard workers. From passing conversations to team projects, the dynamic in the lab was a necessary atmosphere to keep me both focused and involved. I cannot imagine this experience without the opportunities to learn from so many   xx principal investigators, the support from research coordinators and assistants, and the banter with fellow trainees. I want to specifically thank a number of fellow trainees in the Eng Lab who have left a mark on my doctoral journey. To fellow doctoral students, Tara Klassen and Lisa Simpson, whose camaraderie and way-paving have meant so much to me. To post-doctoral fellows, Brodie Sakakibara, Sue Peters, and Chieh-Ling Yang, who have provided mentorship and wisdom beyond their years. And finally, to Shannon Lim, whose academic and professional life has overlapped almost entirely with mine, for the personal friendship beyond what I ever anticipated from my doctoral studies.  I am fortunate to have received financial funding throughout my doctoral studies. The generous financial support from the Vanier Canada Graduate Scholarship program, Canadian Partnership of Stroke Recovery, University of British Columbia (Graduate Program in Rehabilitation Sciences, Graduate and Postdoctoral Studies), Vancouver Coastal Health Research Institute, and Canadian Institutes of Health Research (CIHR). Furthermore, the ExStRA trial was funded by the Heart and Stroke Foundation of Canada and CIHR, which enabled the work to be completed.   Finally, I owe the utmost gratitude to my friends and family, for enduring this journey with me. To all of my friends who have patiently let me vent and for the kind words of moral support. To my loving mother, Ruby Louie, who has given so much to see me through life so far. To my father, Russell Louie, who sought out the educational opportunities for me as a child which undoubtedly led me to this point. To my brother and sister-in-law, Daryl and Katie, for our adventures and experiences as a familial pod. And to Josh, who has seen the best and worst of me throughout this degree, thank you for constantly reminding me of what’s important and for always keeping me grounded.   xxi Dedication To my grandmother, Pearl Louie, whose love persisted beyond her stroke. And to my grandfather, Tim Louie, who instilled in me a reverence for learning and education.   1 Chapter 1: Introduction 1.1 What is a stroke? A stroke is characterized as an acute compromise of the blood supply to the brain, presenting as a sudden loss of neurological function which persists longer than 24 hours (Hankey, 2017). This loss of function can range from a subtle change in speech or movement to complete disablement and death. The term ‘stroke’ only gained medical prominence in the 20th century and originated from the Greco-Roman ‘apoplexia’, meaning ‘to strike suddenly’ or ‘to be struck down with violence’ (Engelhardt, 2017). This semantic derivative is very fitting, capturing both the abruptness of onset and the grave consequences of stroke.  1.1.1 Neurophysiology of stroke The brain has a complex anatomy that is organized into different regions, divided by both structure and function. The brain is the human central processing unit, responsible for shaping thoughts, behaviours, and personality, as well as acting as the control centre for voluntary and involuntary bodily functions. It is not surprising that the brain receives 15 – 20% of the body’s blood supply, and thus oxygen and nutrients (Williams and Leggett, 1989; Xing et al., 2017). As such, disruption of normal blood flow to the brain can have devastating consequences. The mechanism of this disruption can be broadly classified as an ischemic or hemorrhagic event (Donnan et al., 2008). An ischemic stroke occurs when an artery is occluded, either by a thrombotic (e.g., atherosclerotic plaque accumulation) or embolic (e.g., travelling blood clot or air bubble) event, preventing oxygenated blood from reaching downstream brain tissue (Markus and Bevan, 2014). The lack of oxygen causes cell death in the affected brain region and disrupts function in the surrounding tissue, respectively known as the ischemic core and penumbra (Donnan et al., 2008).   2 The magnitude of tissue damage depends on a number of factors, including the location and duration of the blockage (Markus and Bevan, 2014; Prabhakaran et al., 2015). An occlusion of a large blood vessel which feeds many small arteries will result in a much larger ischemic core than the occlusion of a terminal branch of a small artery; similarly, a longer-lasting occlusion will result in a larger area of cell death. Acute medical treatment of ischemic stroke is thus focussed on removing the thromboembolism as early as possible to allow reperfusion of the downstream tissue (Prabhakaran et al., 2015). This can be achieved chemically through administering anticoagulants such as tissue plasminogen activator (tPA), or mechanically by surgical procedures such as endovascular thrombectomy (Hankey, 2017; Prabhakaran et al., 2015). Ischemic strokes are most common and represent approximately 70 – 90% of all strokes (Campbell et al., 2019; Donnan et al., 2008; Jørgensen et al., 1999). A hemorrhagic stroke occurs when a vessel ruptures, thus bleeding into the surrounding brain tissue, and is often referred to as an intracerebral hemorrhage (ICH). The free floating blood coagulates into a hematoma, a physical structure hardened by interlinking chemical bonds, which causes direct mechanical and compressive injury to the brain parenchyma (Caceres and Goldstein, 2012). The extent of injury is dependent on the amount of bleeding, which again depends on the involved blood vessel and duration of bleeding (Wilkinson et al., 2018). The brain can naturally resorb a hematoma and so acute surgical intervention to remove the hematoma is not always performed; instead, management through intensive blood pressure control is the primary treatment for acute ICH (Hankey, 2017; Wilkinson et al., 2018). However, when significant bleeding has occurred, the risk of aggressive and invasive surgical intervention, such as craniectomy, is outweighed by the potential to salvage brain tissue or save a life. Intracranial hemorrhages account for approximately 10-15% of all strokes (Paolucci et al., 2003).    3 There are several other classifications of stroke to be acknowledged. A transient ischemic attack (TIA) has the same mechanism as an ischemic stroke, in which there is a blockage of a brain artery, but self-resolves within 24 hours of onset (Hankey, 2017). Though short-lived, TIAs are regarded as a warning sign of an impending ischemic stroke, and 10 – 15% of those individuals who experience a TIA subsequently experience a stroke within three months (Easton et al., 2009). Recent research has shown permanent changes in brain connectivity and function after a TIA, even after apparent symptom resolution (Ferris et al., 2017). Cerebral venous thrombosis similarly results in blockage of blood flow, except the clot occurs within the venous system of the brain. This type of stroke is rare, accounting for 0.5 – 1% of all strokes (Saposnik et al., 2011), and presents considerably differently from an arterial ischemic stroke due to the dissimilarity in pathophysiological mechanisms of injury (Silvis et al., 2017). Strokes caused by a subarachnoid hemorrhage are also rare, accounting for 5% of all strokes (Donnan et al., 2008); because the bleeding has occurred in the protective space surrounding the brain, the presentation of a subarachnoid hemorrhage does not mirror a hemorrhage directly within the brain tissue. These other types of stroke are worth mentioning but will not be the focus of this dissertation. 1.1.2 Presentation of stroke The brain has an extensive circulatory system, and disruption of normal blood flow within the brain can result in a wide array of potential symptoms. The primary deficits of a stroke will be reflected as loss of the functions associated with the affected brain structures (Nudo et al., 2001), such as arm and leg weakness if the motor cortex is injured or visual loss after a stroke of the occipital lobe. However, due to the interconnectivity of the brain and axonal projections throughout the central nervous system, a stroke can have widespread effects on cognitive processes and body systems beyond the functional area of the stroke (Crofts et al., 2011; Nudo et   4 al., 2001). Strokes can also vary in severity, manifesting as barely observable changes or resulting in severe impairment, disability, or death. Regardless of severity, stroke results in permanent damage to the affected tissue and is considered an acquired brain injury. Upon stroke onset, commonly observed impairments include one-sided weakness (hemiparesis), difficulty walking, numbness, visual disturbances, incoordination (ataxia), and slurred speech (Hankey, 2017). Cognitive and behavioural changes can also occur alongside physical impairments. For example, language comprehension, critical reasoning, memory, self-awareness, and personality can be affected as a result of stroke (Balami et al., 2013; Pare and Kahn, 2012). Automatic processes such as breathing, swallowing, and blinking can also be disturbed (Pare and Kahn, 2012). The specific signs and symptoms of stroke that present acutely after a stroke are in direct relation to the brain areas affected. There are significant long-term effects of stroke, in addition to the immediate physical and mental consequences of the brain injury. Behavioural responses to physical impairments, such as favouring the use of a less affected limb or restricting physical activity altogether, can drive a cyclical pattern of learned non-use of an impaired limb and general physical deconditioning (Billinger et al., 2014; Taub et al., 2006). Emotional responses to the life-changing event can take its toll as depression or anxiety. Ultimately, the physical and mental changes post-stroke can drastically affect an individual’s identity, mood, social participation, and quality of life (Lapadatu and Morris, 2019; De Wit et al., 2017). Understandably, stroke is a leading cause of adult disability (Lopez et al., 2006; World Health Organization, 2020) and also a major focus of research and healthcare spending (Muka et al., 2015). Given that the long-term impact of stroke largely hinges on the initial presentation and response to the event, it emphasizes the importance of early intervention and management.   5 1.1.3 Stroke epidemiology Stroke is considered a global health crisis, with increasing incidence worldwide and mortality second only to ischemic heart disease (Hankey, 2017; World Health Organization, 2020). Currently, the global incidence of stroke is 240 to 600 per 100,000, amounting to an estimated 13.7 million new strokes each year (Campbell and Khatri, 2020; Donnan et al., 2008). With advances in detection and medical care, survival rates after a first incident stroke is improving worldwide (Benjamin et al., 2017; Feigin et al., 2016; Kamal et al., 2015). However, this means that the prevalence of individuals living with disability after stroke is also increasing (Feigin et al., 2016; Hankey, 2017). Due to the resultant morbidity post-stroke, 5-year mortality is also alarmingly high (50%) (Luengo-Fernandez et al., 2013), further emphasizing the global stroke burden (Feigin et al., 2016). Within Canada, there are an estimated 405,000 individuals living with the effects of stroke (Heart and Stroke Foundation, 2016; Krueger et al., 2015); however, it is also suggested that this number is underestimated and may actually be higher after consideration of pediatric stroke and those survivors residing in long-term care facilities (Krueger et al., 2015). Approximately 62,000 Canadians experience a stroke each year, adding to the growing population of stroke survivors (Heart and Stroke Foundation, 2016). The prevalence of stroke survivors is also increasing as a result of population growth, aging, and decreasing mortality rates; by 2038, it is projected that there will be 654,000 to 726,000 Canadians living with the effects of stroke (Krueger et al., 2015). This trend is mirrored in other nations, such as the United States of America, Australia, and Spain (Benjamin et al., 2017; Fisher et al., 2014; Rodríguez-Castro et al., 2018).   6 There are many risk factors for stroke, with smoking, physical inactivity, alcohol use, and obesity contributing to 60% of all cases (Krueger et al., 2015). As mentioned previously, physical inactivity, which in itself is a contributor to other risk factors for stroke such as high blood pressure and diabetes (Virani et al., 2020), is also a long-term behavioural consequence of stroke (Tieges et al., 2015). The majority of people living with stroke do not reach physical activity guidelines (Fini et al., 2017), spending 9.3 – 19.3 hours per day in sedentary postures or behaviours (Tieges et al., 2015; Wondergem et al., 2019). It is unfortunate, yet not surprising, that there is a high risk of stroke recurrence. The risk of a second stroke rises from 6.5 – 9.8% in the first year after stroke to 16 – 37% by the 5-year mark (Benjamin et al., 2017; Jørgensen et al., 1997; Virani et al., 2020). An alarming trend over the past 20 years in Canada and the United States is the rise in stroke incidence in adults less than 60 years of age (Heart and Stroke Foundation, 2017; Virani et al., 2020). This is related to the rising prevalence of traditional risk factors, such as hypertension, hypercholesterolemia, diabetes mellitus, and obesity, in younger adults (Maaijwee et al., 2014). They can likely be attributed to a generally sedentary lifestyle, afforded by the reduction in labour-intensive jobs and technological innovation over the past half-century. A rising incidence among younger adults, coupled with advancing medical and rehabilitation efforts, entails that an increasing number of stroke survivors will live longer, with disability, than previous generations. This has many implications, ranging from economic burden at the national level (Canadian Institute for Health Information (CIHI), 2009; Mittmann et al., 2012), to local hospital and institutional overcrowding (Simmons, 2005; Sutherland and Trafford Crump, 2013), to management and care at the individual level.   7 1.2 Recovery after stroke Improvement after stroke stems from several factors, including spontaneous recovery, neurological recovery, and functional recovery. Spontaneous recovery is the return of brain function due to natural physiological responses of the brain to the initial insult, such as reperfusion of blood flow to the occluded area, resolution of edema, and reversal of hypometabolism in remote brain regions (Bowden et al., 2013; Kwakkel et al., 2006; Paolucci et al., 2003). Spontaneous recovery begins within the first days after stroke and is maximal in the first 4 – 10 weeks, tapering off by 6 months (Kwakkel et al. 2006). It is important to note that spontaneous recovery leads to functional improvement after stroke regardless of rehabilitative efforts and, in the case of lower limb impairment, accounts for 33 – 39% of observed improvement (Kwakkel et al. 2006). Neurological recovery is very complex and largely dependent on neuroplasticity, the brain’s ability to change. Neural reorganization after stroke encompasses growth of new neurons, sprouting of intracortical axons and dendrites, and formation of new connections between surviving neurons (Jones and Adkins, 2015; Nudo et al., 2001). Importantly, neurological recovery is stimulated by behavioural experiences that place demand on the brain, such as training of the paretic limbs. The majority of neurological recovery occurs early after stroke, with most survivors achieving their best neurological scores within 11 weeks (Jørgensen et al. 1999). It is important to remember, however, that neuroplasticity is a feature of the brain regardless of the timeline of neurological recovery after stroke, and individuals with stroke demonstrate neuroplastic changes well into the chronic phase (>6 months post-stroke). Furthermore, since neurological and spontaneous recovery are occurring concurrently in the   8 early months after stroke at the cellular and histological level, both resulting in improvements in brain connectivity and physical function, they are difficult to fully distinguish in practice.  Functional recovery is the improvement that individuals with stroke experience in their ability to carry out meaningful tasks. Functional recovery is correlated with neurological recovery, with the majority occurring within the first 11 – 13 weeks after stroke (Jørgensen et al. 1999), though many studies have demonstrated functional improvement into the chronic phase of stroke (van Duijnhoven et al., 2016; Ferrarello et al., 2011). Functional recovery can be attributed to improvement in physical impairments, which in turn impacts activity limitations, and participation restrictions (WHO, 2001). It can constitute changes in movement (e.g., reaching, walking), communication (if aphasic), and cognition, which ultimately determines the functional abilities of a stroke survivor. Furthermore, using assistive devices, new movement strategies, or modified equipment can also allow successful completion of meaningful tasks. To this effect, functional recovery is the summation of neurological and physiological improvements, as well as behavioural compensations and environmental adaptations. 1.2.1 Phases of stroke recovery The timeline of stroke and recovery is commonly divided into three phases: acute, subacute, and chronic. Like injuries to other tissue, the acute phase of stroke refers to the initial 48 – 72 hours after stroke onset (Prabhakaran et al., 2015), during which time numerous physiological responses are taking place. Oxygen deprivation and mechanical injury disrupt the normal cellular and molecular processes of functioning brain cells, triggering cell death, inflammatory responses, and widespread edema (Campbell et al., 2019; Wilkinson et al., 2018). Medical intervention to address the cause of stroke takes place during the acute phase of stroke, to lessen the extent of tissue damage.    9  The subacute phase of stroke is accepted as the period between the acute and chronic phases of stroke, but the actual definition and timeframe are more nebulous. The subacute phase can refer to the period during which functional connectivity within the brain is disrupted (Stinear et al., 2015). It can also be regarded as the timeframe over which spontaneous recovery of the brain occurs, including histological injury mechanisms and resolution thereof, from days after the onset of stroke to 6 months post-injury (Bernheisel et al., 2011; Krakauer et al., 2012). As mentioned previously, it also coincides with the largest neurological and functional changes seen after stroke. As such, the subacute phase is a crucial period for rehabilitative efforts (Krakauer et al., 2012).   The chronic phase is typically regarded as beginning at the 6-month mark from stroke onset. The beginning of the chronic phase marks the end of spontaneous recovery processes (Krakauer et al., 2012), meaning that improvements in stroke-related impairments are unlikely without active intervention.  1.2.2 Organization of care Patients presenting with acute stroke are admitted to hospital for medical intervention and monitoring. Surgical intervention may be carried out depending on the time course and nature of the stroke (Campbell and Khatri, 2020; Hankey, 2017). Hospitals that have specialized stroke units, which are comprised of an interdisciplinary team of skilled healthcare staff and an intensive model of care, have been shown to yield better outcomes after stroke than general hospital care (Fuentes and Díez-Tejedor, 2009; Langhorne, 2013). The primary focus of acute hospitalization is medical stabilization and minimizing the risk of stroke evolution (worsening of stroke due to additional clotting or further bleeding). To this effect, overly early and intense   10 rehabilitation efforts during acute hospitalization has been shown to be detrimental on stroke outcomes (AVERT Trial Collaboration group et al., 2015). Once medically stable, many survivors are transferred to a post-acute care facility. Approximately 19% of all survivors are discharged to an inpatient rehabilitation facility to participate in intensive therapies to address the impairments caused by stroke (Canadian Stroke Network, 2011). Patients in intensive inpatient rehabilitation typically attend daily physical therapy, occupational therapy, speech-language therapy, and other services for which they are eligible. There are many clinical and nonclinical factors that dictate which patients with stroke are eligible for inpatient rehabilitation, (Hong et al., 2019) with stroke severity consistently playing a role (Schlegel et al., 2004; Treger et al., 2008). Though patients with moderate to severe stroke benefit the most from rehabilitation, only 37% of patients in this category are discharged to a rehabilitation facility (Canadian Stroke Network, 2011). Patients admitted to inpatient stroke rehabilitation have a median length of stay of 35 days and experience significant functional recovery during this time (Canadian Stroke Network, 2011; Chan et al., 2013).  Individuals with stroke who are not admitted to inpatient rehabilitation are of two camps. Those with minimal or manageable impairments are sometimes discharged directly home and recommended to attend rehabilitation as outpatients; approximately 40 – 60% of patients return home after acute hospitalization (Canadian Stroke Network, 2011; Dutrieux et al., 2016; Schlegel et al., 2004). Research has shown that earlier reintegration into a home setting has positive effects on psychological well-being (Dewilde et al., 2020; Yu et al., 2019). Those individuals with more severe strokes, deemed unfit for intensive rehabilitation based on activity tolerance, intellectual capacity, or prognosis, are transferred to chronic care facilities (Hong et al., 2019). Some chronic care facilities provide slow stream rehabilitation, with lower intensity   11 and duration of daily rehabilitation therapies, with the goal of returning patients to independent or semi-independent living (Tourangeau et al., 2011). Other long-term care facilities, such as nursing homes, admit patients with very severe stroke who are not expected to recover (Tourangeau et al., 2011).  1.3 Stroke Rehabilitation The main goal of stroke rehabilitation is functional recovery, whether achieved through promoting neurological improvements or by facilitating behavioural adaptations. Intensive inpatient rehabilitation coincides with the period in which recovery and improvement after stroke are greatest, and thus affords an opportunity to make large gains in function (Jørgensen et al., 1999). Compared to other discharge destinations such as home or skilled nursing facilities, intensive inpatient rehabilitation has been associated with larger improvements in functional outcomes (Chan et al., 2013). A significant focus of rehabilitation research is the development and refinement of interventions to optimize functional recovery after stroke; with the mounting burden of stroke in Canada and worldwide, continued research efforts are imperative. 1.3.1 Rehabilitation of walking A major focus during inpatient rehabilitation is the recovery of walking, a goal identified by the majority of stroke survivors (Bohannon et al., 1988; Harris and Eng, 2004). This is not surprising, given that walking is a form of mobility (the ability to move between meaningful places), and the loss of which can negatively impact an individual’s sense of identity and psychological well-being (Gibson and Teachman, 2012; Taylor et al., 2019). Additionally, the ability to walk has been shown to relate to improved physical and cognitive ability, cardiovascular health, and long-term outcome after stroke (Liu-Ambrose et al., 2007; Pang et al., 2005; Tang et al., 2014). Furthermore, it is associated with discharge destination, long-term   12 mobility, and community reintegration (Bijleveld-Uitman et al., 2013; Mayo et al., 1999; Pereira et al., 2014; Portelli et al., 2005). Epidemiology of lower extremity impairment The most obvious contributor to the ability to walk is lower extremity functioning. Motor loss is a common sequela of stroke, with subjective muscle weakness initially reported in 54 – 63% of acute strokes (Yew and Cheng, 2015). According to one of the most frequently cited population studies describing the effects of stroke and timeline of recovery, the Copenhagen Stroke Study (Jørgensen, 1996), approximately 65% of patients initially demonstrate some degree of lower extremity weakness (Jørgensen et al., 1995). The degree of initial leg paresis was a determinant of downstream walking function at the end of rehabilitation (Jørgensen et al., 1995).  Advances in stroke detection and medical intervention warrant an update of these statistics. Acute treatments for stroke, such as thrombolysis and endovascular therapy, have seen a rise in the last two decades (Rabinstein, 2017). These treatments can reverse the cause of an ischemic stroke in order to mitigate the extent of tissue damage, which may in turn lessen the clinical effects of stroke (Rabinstein, 2017). Furthermore, the recent trends of lower mortality of stroke and increasing incidence in younger adults may in turn have an impact on the epidemiological presentation of stroke. Epidemiology of walking limitations  Approximately 63% of patients are not able to walk independently at the onset of stroke, similar to the incidence of lower extremity weakness (Jørgensen et al., 1995). The severity of initial leg paresis is a determinant of final walking function (Jørgensen et al. 1999), with those more severely impaired still requiring assistance to walk after rehabilitation. However, the ability   13 to walk is not just dependent on lower extremity strength. Walking is a complex task and depends on several other body systems, including sensation, coordination, balance, and cognition  (Kollen et al., 2005; Li et al., 2018). For example, greater initial scores on the Berg Balance Scale at admission to inpatient rehabilitation predicts walking speeds suitable for community ambulation at discharge (Louie and Eng, 2018). Given the multifactorial nature of walking, the rate of its recovery does not completely parallel lower extremity motor improvement early after stroke (Jørgensen et al., 1995). Even at the end of stroke rehabilitation, only 53% of inpatients regain the ability to walk independently (Shum et al., 2014); in total, 30% of stroke survivors still have limited walking ability even after rehabilitation (Jørgensen et al. 1999). The time course of walking recovery generally occurs within 11 weeks of the onset of stroke (Jørgensen et al., 1995). Only 1 in 4 patients who are initially unable to walk on their own regain independence despite intensive conventional physical therapy rehabilitation (Jørgensen et al., 1995); those who do regain independent walking ability often still demonstrate impairments, such as asymmetry, in their gait pattern (Patterson et al., 2015). However, it has been increasingly recognized that rehabilitation can influence neurological recovery and cortical reorganization (Teasell and Hussein, 2016), and so determining optimal methods to retrain walking is of utmost importance, especially in the first few months after stroke when survivors experience the most of their recovery. 1.3.2 The sensorimotor integration theory of motor recovery There are many theories which seek to explain how the brain and body learn and control movement; similarly, there are theories that focus on how the damaged brain recovers to regain motor control (Nudo, 2011). The sensorimotor integration theory, which states that motor output is informed and shaped by sensory information relating to one’s body and the environment, can   14 be applied to both motor control and motor recovery (Krakauer, 2006; Wolpert et al., 1998). For motor recovery, this theory posits that afferent input occurring during movement, whether visuospatial, cutaneous, or proprioceptive, recruits and trains non-damaged areas in the brain to plan more effective movement patterns (Cauraugh and Kim, 2003; Hallett, 2001). A simple visual schematic of the sensorimotor integration theory is presented in Figure 1.1. This theory highlights the importance of the sensory system and interconnectivity between different brain regions.  The sensorimotor integration theory has been widely applied in stroke. It has been used to  justify different treatment approaches, from manual techniques, such as the Bobath approach (Levin and Panturin, 2011), to technology-based interventions using neuromuscular stimulation (Cauraugh and Kim, 2003). The majority of interventions in physical rehabilitation can be framed in terms of the sensorimotor integration theory, as it simply links the sensory experience of movement to its recovery. In other words, physical therapy interventions in stroke rehabilitation generally focus on improving functional abilities, whether through hands-on assistance or verbal guidance to perform a new movement pattern; the experience of successfully completing the new task will then shape and strengthen the neurological pathways that allow that very task to be repeated voluntarily. Figure 1.1 was designed to illustrate the feedback loop of sensory experience and motor recovery that rehabilitation efforts seek to achieve.   15 Figure 1.1 Illustration of the sensorimotor integration theory of motor recovery   1.3.3 Repetitive task-specific practice for optimal recovery Current research guidelines recommend that patients with stroke should engage in meaningful and goal-oriented training that is intensive, repetitive, and task-specific to improve ambulation (Hebert et al., 2016; Hornby et al., 2011). Though not explicitly stated in the guidelines, the recommendations for intensity, repetition, and task-specificity are key principles in the sensorimotor integration theory of motor learning and recovery (Kitago and Krakauer, 2013). However, it is not always feasible to achieve all of these requirements, especially for walking practice. Patient factors, such as physical impairment, motivation, and fatigue, can often dictate the pace of therapy. Therapist factors can also impact the efficiency of rehabilitation,   16 relating to confidence, diligence, or expertise (Cahill et al., 2021). In a study involving over 1000 hours of therapy observation, Clarke et al. identified a lack of understanding of intensity-related research amongst therapists as a primary factor influencing the amount and frequency of therapy provided (Clarke et al., 2018). Striking a balance between safe and high-intensity walking practice is imperative, as therapists in neurological rehabilitation are amongst the highest risk for work-related injuries compared to other practice settings, resulting in higher attrition rates from this field (Cromie, 2000; Vieira et al., 2016).     Animal models in the stroke literature show that neural changes in the brain, indicating recovery, is a result of behavioural experience (Krakauer et al., 2012). These behavioural experiences include forced use of the paretic limb, such as stepping or reaching, and lead to sprouting of intracortical axons and dendritic spine morphogenesis (new growth or change in brain cells) (Jones and Adkins, 2015). This neuroanatomical reorganization is particularly dynamic soon after injury (Jones and Adkins, 2015), with the mice and rats in the stroke models receiving thousands of repetitions during their “rehabilitation” (Krakauer et al. 2012). This is substantially higher than the repetitions achieved in stroke rehabilitation, which is especially low for those who have greater impairment; movement repetitions during inpatient therapy have been shown to be as low as 6 – 32 repetitions in a single session (Lang et al., 2009; Rand and Eng, 2012). Despite the difference from animal models in repetitions achieved, clinical stroke research has found that repetitive task-specific training, the practice of real-life tasks with the intention of (re)acquiring a skill, does improving walking endurance and speed (English and Hillier, 2011; Van Peppen et al., 2004; van de Port et al., 2012). Other than practicing the act of walking itself, research has shown that practicing part of a desired task can still lead to beneficial   17 outcomes; for example, standing practice and simply loading the paretic lower limb is correlated with better walking outcomes after 6 months (Kollen et al., 2005; Mercer et al., 2014). As such, it is important to consider alternative methods to allow stroke survivors who are unable to participate independently in gait-related task-practice to still receive that behavioural experience.  1.4 Role of technology in stroke rehabilitation The use of technology can potentially enhance outcomes for patients after stroke. Generally speaking, technology is the application of scientific knowledge for practical purpose. In terms of rehabilitation, technology can take on many forms and purposes. It can vary from simple devices to complex equipment and systems. It can be used by patients themselves to improve day-to-day functioning, as well as by clinicians to monitor or provide therapy. Using technology for rehabilitation can increase therapeutic engagement, whether through improving task-specificity, amount of therapy (repetition and duration), or patient motivation (Clark et al., 2019; Fager and Burnfield, 2014).  Rehabilitation technology may also offer an opportunity to provide therapy which would otherwise not be feasible. For example, using a knee brace or mechanical lift might be the only way a therapist is able to safely support a physically dependent patient to stand and practice taking a few steps. The dawn and rapid development of electronic and computerized technology has and continues to revolutionize the landscape of rehabilitation (Boehlen and Sample, 2020; Sidler, 1986). Rehabilitative technology to improve walking has evolved from simple equipment, such as orthotics and canes, to complex body weight-supported treadmills and electromechanical robotics. Now, these latter devices can do all of the work of helping a patient to be upright and taking hundreds of steps in therapy.    18 1.4.1 Electromechanical interventions for gait rehabilitation The use of electromechanical devices has received much attention as a potential intervention to achieve repetitive task-specific training for patients with stroke (Hesse et al., 2013; Mehrholz et al., 2013; Moucheboeuf et al., 2020). One reason is that electromechanical devices, such as body weight-supported treadmills and robotic gait orthoses, can help therapists overcome the challenge of safely ambulating a patient who requires more physical assistance. Furthermore, because of the taxing nature of assisting walking, patients may achieve greater duration and repetition of walking practice using these devices since they are not limited by therapist fatigue or lifting constraints (Schwartz and Meiner, 2015). Without such technology, patients with stroke requiring more assistance from therapists are often trapped in a position where it is not feasible to provide high-repetition walking therapy over the course of standard inpatient rehabilitation (Rand and Eng, 2015). The electromechanical devices used for walking practice range from a treadmill-and-harness system to robotic devices that brace the legs and assist the stepping motion partially or fully. In its simplest form, body weight-supported treadmill training (BWSTT) suspends the patient over a treadmill and offers an opportunity for repetitive task-based practice without needing sufficient muscle activation or strength to stand independently against gravity. With the harness and treadmill speed control, it also affords therapists an opportunity to challenge their patients at greater walking intensities than would be safe to do so manually. It has been shown that BWSTT leads to greater gains than conventional overground training for patients with stroke (Ada et al., 2010; Charalambous et al., 2013; Eich et al., 2004; Hornby et al., 2016). However, one drawback of BWSTT is the significant demand it places on therapists, who are often still required to assist with moving the paretic lower limb or stabilizing the trunk during walking   19 practice. Therapist fatigue may potentially affect the patient’s success in training, due to the variability and degradation of consistent assistance provided (Hidler et al., 2009). Technological progression has led to the development of driven gait orthoses, such as the Lokomat (Hocoma, Zurich, Switzerland) and G-EO System (Reha Technology, Olten, Switzerland), which are robotic systems that provide the trunk and lower limb assistance during suspended-in-place walking practice over a treadmill. These devices reduce the strain on the therapist and allow for even higher repetitions of walking than BWSTT, with a consistent swing on each step. One systematic review of electromechanical gait trainers has found that using such a device for walking practice increases the likelihood of independent walking after stroke (Mehrholz et al., 2015); another found improved gait velocity and maximum walking distance (Mehrholz et al., 2018). Some drawbacks of these robotic treadmill devices include a restriction in side-to-side weight shift, a lack of visuospatial flow, and limited arm use or swing during gait (Hidler et al., 2009) — all crucial elements of walking. Another review also found that electromechanical gait training (including BWSTT) for non-ambulatory patients with subacute stroke (<3 months post-stroke) leads to better short-term and long-term walking improvements than overground walking (Ada et al., 2010). There are also several studies and reviews that have shown that BWSTT and robotics-assisted gait trainers are not superior to conventional overground walking retraining. In the LEAPS trial, Duncan et al. found no significant difference between home exercise and BWSTT for improvement in walking recovery (Duncan et al., 2011). Hidler et al. found that conventional overground walking practice leads to gains in speed twice as large as those who trained in the Lokomat, for ambulatory patients with stroke (Hidler et al., 2009). A Cochrane review showed no statistically significant difference between BWSTT and any other walking intervention for   20 walking ability or speed (Moseley et al., 2005). It has been suggested that these devices might not be effective because the task components that are practiced do not necessarily replicate the biomechanics of overground walking (Dobkin and Duncan, 2012). The conflicting evidence regarding electromechanically-assisted gait training is in part due to the variation of the participants included in each study (Mehrholz et al., 2018; Moucheboeuf et al., 2020). The abovementioned research has included a combination of either subacute or chronic stroke and either ambulatory or non-ambulatory participants. The current consensus, according to best practice guidelines, is that non-ambulatory subacute patients make the most gains in walking recovery with electromechanically-assisted gait training (Hebert et al., 2016). 1.4.2 Powered robotic exoskeletons  Powered robotic exoskeletons are a more recently developed electromechanical technology to enable walking for anyone with lower extremity weakness, improving upon the constraints of prior devices (Chen et al., 2013). These wearable robots strap to the legs and have electrically actuated motors to control joint motion to automate walking and can be used independent of a treadmill or overhead harness system. Powered exoskeletons were initially designed to be used as an assistive device to allow individuals with neurologically complete spinal cord injury to walk without adverse effects (Kolakowsky-Hayner et al., 2013; Kressler et al., 2014), but have been posited to be a clinical tool for rehabilitation in other neurological conditions. In addition to the physical support provided to brace the limbs, the control mechanism of stepping can vary from fully automatic and complete assistance, to patient-initiated and partial assistance, or no assistance at all.   21 These new powered robotic exoskeletons have been minimally researched in the clinical setting, given their recent development. The first iterations of powered exoskeletons were first developed in 2010 – 2011, undergoing both software and hardware changes in the ensuing years (Chen et al., 2013). In addition, there are several competing commercial developers that have each produced their own exoskeleton. The majority of early research has focused on the safe and functional usage of these devices for assisted walking in complete spinal cord injuries (Louie et al., 2015), rather than as a rehabilitation intervention in other neurological populations that retain some walking function. Only in April of 2016 did the American Food and Drug Administration first approve the use of robotic exoskeleton in the treatment of individuals with stroke hemiparesis; they are now being commercially advertised as useful tools in neurological rehabilitation. Potential benefit of exoskeletons in stroke rehabilitation There are good reasons that powered robotic exoskeletons are being touted as clinically useful interventions for gait retraining after stroke. Using a powered exoskeleton to enable successful overground walking allows the patient to experience repetitive task-specific practice and the associated sensory information (Morone et al., 2017). The sensory experience of weight-bearing, balance control, and overground movement are an improvement to suspended and stationary electromechanical devices. Additionally, robotics also enhance repeatability of a treatment, increase training motivation, as well as enable objective and quantifiable measures of performance (Belda-Lois et al., 2011). The amount of assistance provided by powered exoskeletons can be varied by the operating therapist, allowing for the provision of patient-specific task practice to reinforce voluntary movement and challenge muscular strength. Figure   22 1.2 displays how exoskeleton-based gait training interfaces with the original illustration of the sensorimotor integration theory.  Figure 1.2 Powered exoskeletons and the sensorimotor integration theory   It is important to consider the impairment level, time post-stroke, and other key factors that strongly influence recovery after stroke when providing a specific intervention for gait retraining (Hidler et al., 2009). Given the previous conflicting findings of electromechanically-assisted walking interventions, it has been suggested that experimental treatment in neurorehabilitation should be reserved for the most impaired and disabled patients who still   23 retain some level of motor control (Dobkin and Duncan, 2012). Research in neurophysiology has shown that the greatest neuroplastic changes are associated with the generation and repetition of novel movements (Bowden et al. 2013), and so exoskeleton-based gait retraining may better be suited for those individuals who are unable to walk on their own. The early and unfamiliar behavioural experience of walking may take advantage of the heightened neuroplasticity and neural changes occurring in the recovering stroke brain (Bowden et al., 2013; Charalambous et al., 2016). Powered robotic exoskeletons meet many of the requirements for motor learning and recovery where previous electromechanical devices have fallen short. Major criticisms of electromechanical devices, such as the Lokomat, are the lack of variability in the assisted movement and a tendency for patients towards passive participation (Dobkin and Duncan, 2012). It is argued that robotic control algorithms should be able to provide an error signal in a physiologically meaningful movement path (Belda-Lois et al., 2011; Dobkin and Duncan, 2012). In other words, robotics-assisted gait training should accommodate the variability of walking while being sufficiently challenging, such that successful stepping is not always achieved. It is the specificity of the task and the discrepancy between actual and desired outcome which stimulate active motor control and learning (Belda-Lois et al., 2011). Current powered exoskeletons have accommodated these recommendations, by incorporating programmable software to adapt the movement pattern and level of assistance to the individual. The lack of an overhead harness allows more side-to-side movement, requiring greater patient participation to stay balanced and introducing more variability during gait training. Using a powered exoskeleton also allows patients to accomplish higher duration and repetition of walking practice, which is   24 needed for motor learning (Krakauer et al. 2012), without the limitation of insufficient strength or therapist fatigue. 1.4.3 Adoption of technology As should be the case with any new rehabilitation intervention, adoption of therapeutic technology into practice requires evidence of efficacy (Dobkin and Duncan, 2012). This is best proven through rigorous randomized controlled trials (RCTs), wherein a well-defined experimental intervention using the new technology is found to be at least equal to, if not better than, conventional physical therapy. For example, the principles of task-specificity, intensity, and repetition, which are now standard practice, were adopted after numerous trials demonstrated greater efficacy than previous iterations of conventional stroke rehabilitation (Kwakkel et al., 1997; Langhorne et al., 1996; Page, 2003).  In addition to efficacy and of equal importance, acceptance is another requirement for the adoption of new technology or interventions into clinical practice (Turchetti et al., 2014). A discrepancy between the efficacy and acceptability of a device can prevent its adoption; for example, video game-based therapy has been shown to improve outcomes after stroke (Bonnechère et al., 2016), but few therapists use video game devices regularly (Langan et al., 2017). Acceptability can be viewed as the users’ reaction to the technology itself (Holden and Karsh, 2010). Though simple in definition, acceptance and its influence on adoption are complex processes; many theoretical models have been proposed to better conceptualize these concepts (Straub, 2009). Factors such as ease of use, perceived usefulness, social influences, and personal characteristics are among many considerations that are included in models of technology acceptance (Davis, 1989; Venkatesh et al., 2003). In the example of video game-based therapy,    25 perceived barriers, such as learning to use the game systems, may in part explain the abandonment of this technology (Levac and Miller, 2013). Perceptions of technology in stroke rehabilitation The perceptions of patients and therapists are crucial in determining the actual use of technology in practice (Chen and Bode, 2011). Lack of information or access are amongst the main reasons for not using technology (Hughes et al., 2014), yet therapists and patients with stroke have high hopes for rehabilitation technology (Lam, 2015; Stephenson and Stephens, 2018). This optimism granted to new technology may reflect a hope to improve stroke outcomes beyond stereotypical limits of recovery (Krakauer and Marshall, 2015; Smith et al., 2017). For example, electromechanical movement retraining technology is generally viewed in a positive light (van Ommeren et al., 2018). With the rapid development of powered robotic exoskeletons and their commercialization, they are increasingly present in various clinical settings (Jayaraman et al., 2017). The diffusion of the technology so far into clinical settings before the availability of supporting evidence is likely a result of marketing efforts and expectations of efficacy. However, it is still crucial to determine their actual efficacy, as well as their acceptability, for the purposes they are being employed. No research, as of yet, has specifically investigated the use of powered overground exoskeletons in the recovery of walking after stroke. Furthermore, because they are to be used by a therapist for patient care, the views of both therapists and patients should be considered when investigating acceptability.  1.5 Dissertation objectives In summary, changes in population trends, medical intervention, and stroke mortality account for a growing number of Canadians living with the effects of stroke. Improving   26 functional outcomes, such as independence of walking, ultimately influences the downstream quality of life for stroke survivors, through enabling community reintegration, as well as by reducing caregiver and healthcare burden. Technological advances have seen the development of powered robotic exoskeletons, which are becoming commercially available and touted by developers as an unparalleled rehabilitation experience. However, there is minimal research evidence so far regarding the use of powered robotic exoskeletons in stroke rehabilitation, and previous research of electromechanical devices for walking recovery has been inconclusive. Research efforts surrounding novel robotic exoskeletons are needed to evaluate both the efficacy and acceptability of such interventions for walking recovery before they are implemented widely in the clinical setting.  This thesis will address the current gaps in the literature surrounding walking recovery after stroke, specifically focussed on using a powered robotic exoskeleton as a gait retraining intervention. The specific objectives of this research are: 1. To provide an update of the nature and prevalence of leg impairment and walking limitation early after stroke, and to characterize the importance of walking on discharge disposition after acute hospitalization (Chapter 2). 2. To map the current evidence regarding the use of powered robotic exoskeletons in post-stroke rehabilitation for walking recovery (Chapter 3). 3. To determine the efficacy of an exoskeleton-based gait intervention during inpatient subacute stroke rehabilitation to improve walking recovery for non-ambulatory patients (Chapter 4). 4. To explore the experience and acceptability of exoskeleton-based gait rehabilitation from the perspectives of patients with stroke and their therapists (Chapter 5).   27 Chapter 2: Prevalence of walking impairments after acute stroke and its impact on home discharge 2.1 Introduction Stroke is the second leading cause of adult mortality and disability-related disease burden worldwide (World Health Organization, 2020), often resulting in mental and physical impairments (Hankey, 2017). Of particular concern to stroke survivors is the presence of lower extremity weakness and walking limitation, often cited as the highest priority for rehabilitation and research efforts (Harris and Eng, 2004; Rudberg et al., 2020). Gait training and the ability to walk after stroke have been routinely linked to greater functional independence (Pundik et al., 2012), community mobility (Bijleveld-Uitman et al., 2013), and improved quality of life in the chronic stage of stroke (Muren et al., 2008). Furthermore, walking ability has been associated with lower healthcare utilization and higher exercise adherence after stroke (Caetano et al., 2020; Minet et al., 2020). Thus, it is clear that the ability to walk is a key determinant of long-term stroke outcomes.  It is important to estimate the prevalence of lower extremity impairment and walking limitation as early as possible after stroke, given their relation to individual prognosis and healthcare burden. Although conducted more than 20 years ago, the Copenhagen Stroke Study is one of the most frequently cited population studies, having documented the prevalence of various impairments from early on after stroke (Jørgensen, 1996). From consecutive hospital admissions for acute stroke between 1992 and 1993, the investigators found that 51% of patients were unable to walk immediately after stroke, 12% could walk with assistance, and 37% could walk independently (Jørgensen et al., 1995). However, recent global and national reports have shown   28 a downward trend in stroke mortality and disability (Johnson et al., 2019; Wafa et al., 2020), likely due to various factors such as earlier detection, improved medical care (thrombolysis and endovascular thrombectomy), and increasing incidence in younger adults (Benjamin et al., 2017; Hathidara et al., 2019; Kamal et al., 2015; Lackland et al., 2014). An update of the prevalence of lower extremity function and walking limitation after stroke is warranted to ensure healthcare is optimized for current trends and to facilitate resource planning.  With rising admissions for stroke and accompanying concerns surrounding healthcare and resource allocation, it is also critical to predict the trajectory of patients as early and accurately as possible. Numerous studies have established the predictive ability of different functional measures to predict walking outcomes and discharge disposition after inpatient rehabilitation (Bland et al., 2012; Brown et al., 2015; Stillman et al., 2009; Vluggen et al., 2020). However, studies predicting discharge disposition after acute hospitalization have largely focused on measures of stroke severity (Reznik et al., 2018; Rundek et al., 2000; Schlegel et al., 2003). Though walking ability is often considered a downstream outcome in stroke prediction models, it may also serve as a useful predictor of discharge disposition after acute care when assessed early after stroke. It can be quickly assessed and represents a functional summary of many factors, such as physical impairment, confidence, motivation, and balance.   Hence, the primary objective of this study was to provide contemporary estimates of the prevalence of leg and walking impairment after first-ever stroke, focussing on differences between patients who are able to walk immediately after stroke and those who are not. A secondary objective was to characterize the predictive nature of early walking ability for being discharged home after acute hospitalization. Based on previous literature and emerging trends in stroke incidence, we hypothesized that between 40 – 50% of patients with stroke would   29 experience walking limitations immediately after stroke, which would significantly impact their discharge outcome. 2.2 Methods This observational study used screening data from a consecutive sample longitudinal cohort study examining recovery post-stroke, for which ethics and operational approval was obtained from the local university and hospital review board. This analysis was reported in accordance with the Strengthening the Reporting of Observational studies in Epidemiology (STROBE) guidelines (von Elm et al., 2007), which is provided in Appendix A. 2.2.1 Participants All individuals admitted between February 28, 2016 and August 31, 2017 to the stroke unit of the Vancouver General Hospital in British Columbia, Canada were prospectively screened for the longitudinal cohort study. This quaternary hospital serves a population of approximately 2.8 million people and is one of only two comprehensive stroke centres in its province (Statistics Canada, 2017). As such, patients with suspected stroke are frequently transferred from other areas to this centre for higher level of care. To answer the proposed research questions, individuals over the age of 18 who experienced a first-ever ischemic (including lacunar stroke) or hemorrhagic stroke, confirmed by CT or MRI, were included in this analysis. Individuals hospitalized for a transient ischemic attack, subarachnoid hemorrhage, or cerebral venous thrombosis were excluded, given their unique pathophysiology, prognosis, and management when compared to arterial stroke (Dmytriw et al., 2018; van Gijn et al., 2007; Pace et al., 2018; Zuurbier and Coutinho, 2017). Individuals already in hospital at the time of their stroke, admitted more than 48 hours after their stroke, or who died during the acute hospitalization period were also excluded.    30 2.2.2 Data collection and variables A trained research assistant extracted demographic information (age, sex), stroke characteristics (side, type, location, interventions, severity), and impairments (lower extremity, walking function) from medical records while prospectively screening patients on the stroke unit. Discharge details (length of stay and discharge location) were abstracted from the medical records by a second research assistant. Primary objective: leg and walking impairment after stroke Leg motor impairment was assessed on hospital admission by the consulting neurologist, using the National Institutes of Health Stroke Scale (NIHSS) lower extremity motor score (Brott et al., 1989; Lyden, 2017; Lyden et al., 1994). The 5-point motor score is rated from 0 (no weakness, indicated by no drift when held up against gravity) to 4 (no movement in the limb at all). The score for the paretic leg was recorded in addition to the total NIHSS score. Overall, the NIHSS measures stroke severity across 15 domains potentially affected by stroke, including level of consciousness, vision, speech and language, motor function, sensation, and coordination. Higher scores on the NIHSS, scored out of 42, indicate greater severity of stroke.  Early walking ability was extracted from the AlphaFIM (Uniform Data System for Medical Rehabilitation, 2009) outcome. The AlphaFIM is a condensed version of the Functional Independence Measure (FIM) (Uniform Data System for Medical Rehabilitation, 1997), and has been shown to be reliable and valid in the stroke population (Stillman et al., 2009). For the primary objective, the AlphaFIM provided two variables relating to walking ability: a binary variable of being able to walk in any capacity (Can Walk: Yes/No), and an ordinal variable of walking ability (ranging from total dependence to complete independence). The ordinal variable was revised from a 7-point to a 6-point scale, such that those requiring total or maximal   31 assistance were collapsed into a single (non-functional) value, given difficulties in discerning the two categories; this only affected six participants. Those deemed non-ambulatory (Can Walk: No) were assigned the non-functional ordinal value as well. The AlphaFIM was completed by the treating team (physical therapists, occupational therapists, nurses) within 3 days of ischemic stroke or 5 days of hemorrhagic stroke. Secondary objective: predicting home discharge The dependent variable for the secondary objective was the binary outcome of being discharged home or elsewhere after the acute hospitalization period. A home discharge without the need for further institutionalization is considered the optimal trajectory after stroke, and so for this analysis, discharge to another hospital, inpatient rehabilitation, or long-term care were grouped together. The independent variables were comprised of the aforementioned demographic information, relevant stroke characteristics, leg impairment, and walking ability.  2.2.3 Statistical analysis Descriptive statistics were used to summarize the sample. Mean and standard deviation (SD) were reported for continuous data; median and interquartile range (IQR) were reported for ordinal data and continuous data that were not normally distributed. To address the primary objective, counts and percentages were used to report prevalence of lower extremity and walking impairment. To further characterize the population relative to walking ability, demographics and stroke characteristics were compared between ambulatory and non-ambulatory participants using independent t-test, Mann Whitney U, and χ2 analyses. A complete case analysis was performed. For the secondary objective, logistic regression was performed to investigate the role of early walking ability in predicting discharge home from acute care, in relation to patient demographics and stroke characteristics. Univariate regression was first performed for each   32 independent variable and the binary outcome of being discharged home or elsewhere. Multivariate logistic regression was then performed by purposeful selection (Zhang, 2016), including all independent variables that were associated with the outcome variable at the p < 0.10 level. Since leg impairment and walking score are broadly captured within the total NIHSS score and binary ambulatory status, respectively, parallel multivariate models were built to avoid variable collinearity and conceptual overlap. Model A was built around the 6-point walking score and the 5-point leg motor impairment score, while Model B was built around the binary ambulatory status (yes/no) and total NIHSS score. Assumptions of logistic regression and interactions between significant independent variables were assessed. The Hosmer-Lemeshow test, Nagelkerke R2, and area under the receiver operating characteristic (ROC) curve were used to assess model fit and predictive capacity of the logistic regression model (Bewick et al., 2005; Smith and McKenna, 2013; Zhang, 2016). An area under the curve (AUC) of 0.9 or greater indicates outstanding discriminative ability, 0.8 – 0.9 indicates excellent discriminative ability, 0.7 – 0.8 indicates acceptable discriminative ability (Hosmer and Lemeshow, 2000; Mandrekar, 2010). All analyses were performed using RStudio version 1.3.959 (RStudio, Boston, MA, USA) running on R version 4.0.1 (R Foundation for Statistical Computing, Vienna, Austria). A p-value < 0.05 was considered statistically significant. A sample of the analysis script is provided in Appendix B.  2.3 Results Of 819 admissions to the stroke unit, 669 patients were admitted to the hospital for a first-time stroke within 48 hours of onset, 42 of whom (6.3%) died during acute hospitalization. Of those who survived, 487 patients had a complete initial NIHSS score and had been assessed   33 for early walking ability, forming the primary population. There were 140 patients that were excluded from the analysis because of missing initial NIHSS score or early walking ability (22.3%), but they did not differ from the primary population by age, sex, stroke characteristics (side, type, location), or discharge disposition. 2.3.1 Leg and walking impairment after stroke  Patient demographics, stroke characteristics, and discharge disposition for the sample (n = 487) are displayed in Table 2.1, grouped into those unable to walk (including total or maximal dependence) and those with some ambulatory capacity (moderate assist to complete independence). Less than half (44.1%) of patients admitted for stroke presented initially with some level of lower extremity weakness. The distribution of walking scores is shown in Table 2.2, indicating that 46.0% of patients were completely dependent, while up to 57.9% of patients were unable to walk without any form of physical assistance within the first 3 – 5 days of stroke. A significantly greater proportion of non-ambulators had a hemorrhagic stroke (15.2 vs. 9.1%); non-ambulators also had significantly greater stroke severity and lower extremity impairment scores than ambulatory participants. Those who retained or regained some walking function early after stroke had a shorter length of stay than those who did not (p < 0.001), and a greater proportion returned home after acute hospitalization. Early medical intervention for ischemic stroke, either thrombolysis or thrombectomy, was not significantly associated with early walking ability.   34 Table 2.1 Patient characteristics in relation to binary ability to walk in the total cohort Variable Total n = 487 Can walk: Yes n = 263 Can walk: No n = 224 p-value Age, in years, mean (SD) 68.5 (14.9) 67.4 (13.9) 69.7 (15.9) 0.09a Sex, female, n (%) 206 (42.3%) 106 (40.3%) 100 (44.6%) 0.33b Side affected, n (%)     0.31b Left 240 (49.3%) 124 (47.1%) 116 (51.8%)  Right 214 (42.1%) 112 (42.6%) 93 (41.5%)  Both 45 (8.6%) 27 (10.3%) 15 (6.7%)  Type of stroke    0.04b Ischemic, n (%) 429 (88.1%) 239 (90.9%) 190 (84.8%)  Hemorrhagic, n (%) 58 (11.9%) 24 (9.1%) 34 (15.2%)  Location of stroke, n (%)    0.08b Cortical 175 (35.9%) 106 (40.3%) 69 (30.8%)  Subcortical† 248 (50.9%) 127 (48.3%) 121 (54.0%)  Both 64 (13.1%) 30 (11.4%) 34 (15.2%)  Received thrombolysis, n (% of ischemic) 137 (31.9%) 72 (30.1%) 65 (34.2%) 0.37b* Received endovascular thrombectomy, n (% of ischemic) 91 (21.2%) 50 (20.9%) 41 (21.6%) 0.87b* Initial NIHSS, median (IQR) 6 (3 – 13) 4 (2 – 8) 9 (5 – 16) <0.001c Leg impairment (NIHSS motor score), n (%)    <0.001c 0 272 (55.9%) 189 (71.9%) 83 (37.0%)  1 64 (13.1%) 34 (12.9%) 30 (13.4%)  2 41 (8.4%) 13 (4.9%) 28 (12.5%)  3 57 (11.7%) 17 (6.5%) 40 (17.9%)  4 53 (10.9%) 10 (3.8%) 43 (19.2%)  Discharge destination, n (%)    <0.001b Home 248 (50.9%) 201 (76.4%) 47 (21.0%)  Rehab 113 (23.2%) 32 (12.2%) 81 (36.2%)  Long Term Care 31 (6.4%) 5 (1.9%) 26 (11.6%)  Repatriated to another hospital 95 (19.5%) 25 (9.5%) 70 (31.2%)   Length of stay, in days, median (IQR) 8 (4 – 17.5) 6 (3 – 10) 14.5 (7.75 – 31.25) <0.001c Abbreviations: IQR, interquartile range; NIHSS, National Institutes of Health Stroke Scale; SD, standard deviation aIndependent t-test; bχ2 test; cMann-Whitney U test †Includes midbrain and brainstem strokes *Comparison made only between those with ischemic stroke Bold indicates a significant p-value <0.05   35 Table 2.2 Walking ability of whole sample (n = 487) Walking Score n (%) 0 – Non-functional (Unable / Total / Maximal Assist) Unable / patient performs <25% of the effort / 25 – 49% of the effort 224a (46.0%) 1 – Moderate Assist Patient performs 50 – 74% of the effort 14 (2.9%) 2 – Minimal Assist Patient performs 75% or more of the effort 44 (9.0%) 3 – Supervision No hands on, patient performs 100% of the effort 61 (12.5%) 4 – Modified Independence Requires device, or not timely or safely 32 (6.6%) 5 – Complete independence No device required, timely, safely 112 (23.0%) aUnable: 218, Total assistance: 3, Maximal assistance: 3  2.3.2 Predicting home discharge In this cohort, 50.9% of patients admitted to hospital for acute first-ever stroke were discharged home. Only 21.0% of non-ambulators were discharged home, compared to 76.4% of those with some ambulatory function. Table 2.3 shows the association of each independent variable, on univariate logistic regression, with being discharged home after stroke for the whole sample. Median length of stay was shorter for those discharged home than for those discharged elsewhere (5.5 days (IQR 3 – 10) vs. 14 days (7.5 – 26.5), p < 0.001).    36 Table 2.3 Association of patient characteristics with home discharge Variable Discharged home n = 248 Discharged elsewhere n = 239 OR (95% CI) p-value Age, in years, mean (SD) 67.0 (14.4) 70.0 (15.2) 0.99 (0.97 – 1.00) 0.03 Sex, female (reference), n (%) 103 (41.5%) 103 (43.1%) 1.07 (9.74 – 1.52) 0.73 Side affected, n (%)      Left (reference) 117 (47.2%) 123 (51.5%)   Right  108 (43.5%) 97 (40.6%) 1.17 (0.81 – 1.70) 0.41 Both 23 (9.3%) 19 (7.9%) 1.27 (0.66 – 2.48) 0.47 Type of stroke, ischemic (reference), n (%) 227 (91.5%) 202 (84.5%) 0.51 (0.28 – 0.88) 0.02 Location of stroke, n (%)     Cortical (reference) 105 (42.3%) 70 (29.3%)   Subcortical 114 (46.0%) 134 (56.1%) 0.58 (0.38 – 0.84) 0.005 Both 29 (11.7%) 35 (14.6%) 0.55 (0.31 – 0.98) 0.04 Initial NIHSS, median (IQR) 4 (2 – 7) 9 (5 – 15.5) 0.90 (0.87 – 0.93) <0.001 Leg impairment, median (IQR) 0 (0 – 1) 1 (0 – 3) 0.60 (0.51 – 0.68) <0.001 Can Walk, Yes, n (%) (reference: No) 201 (81.1%) 62 (25.9%) 12.21 (8.01 – 18.93) <0.001 Walking score, median (IQR) 4 (2 – 5) 0 (0 – 1) 2.19 (1.94 – 2.51) <0.001 Abbreviations: 95% CI, 95% confidence interval; IQR, interquartile range; NIHSS, National Institutes of Health Stroke Scale; OR, odds ratio; ref, reference; SD, standard deviation Bold indicates a significant p-value <0.05  Age, stroke type, stroke location, stroke severity, leg impairment, and walking ability (both binary and ordinal ratings) were identified as relevant predictors of being discharged home. The two multivariate models are presented in Table 2.4: an ordinal walking Model A (containing leg impairment and 6-point walking score) and a binary walking Model B (containing overall stroke severity and binary ambulatory status). Goodness-of-fit was acceptable for both models,   37 by either Nagelkerke R2 or Hosmer-Lemeshow test; no significant interactions were found in either model. As seen from the ROC curves shown in Figure 2.1, Model A had slightly greater accuracy than Model B in predicting home discharge. However, both Model A (AUC: 0.86, 95% CI: 0.83 – 0.90) and Model B (AUC: 0.84, 95% CI: 0.80 – 0.87) were considered to have excellent discriminative ability. Table 2.4 Final multivariate regression models for discharge home (vs. other institutions)  B OR  95% CI p-value Model A  Age -0.008 0.99  0.98 – 1.01 0.33 Hemorrhagic (ref = ischemic) -0.42 0.65  0.31 – 1.36 0.26 Location (ref = cortical)     Subcortical -0.43 0.65  0.38 – 1.10 0.11 Both -0.23 0.79  0.37 – 1.68 0.55 Maximal leg impairment -0.25 0.78 0.65 – 0.92 0.005 Walking score (ordinal) 0.71 2.04 1.79 – 2.35 <0.001 Model B Age -0.01 0.99  0.97 – 1.00 0.08 Hemorrhagic (ref = ischemic) -0.48 0.62  0.30 – 1.26 0.19 Location (ref = cortical)     Subcortical -0.64 0.53 0.32 – 0.86 0.01 Both -0.27 0.76 0.37 – 1.56 0.045 Initial NIHSS -0.07 0.93 0.90 – 0.96 <0.001 Walking score (binary, ref = no) 2.23 9.29 5.98 – 14.66 <0.001 Abbreviations: 95% CI: 95% confidence interval; B, variable coefficient in model; OR, odds ratio Model A: Nagelkerke Pseudo-R2: 0.50; Hosmer-Lemeshow χ2 = 5.68, df = 8, p = 0.68 Model B: Nagelkerke Pseudo-R2: 0.42; Hosmer-Lemeshow: χ2 = 9.88, df = 8, p = 0.27 Bold indicates a significant p-value <0.05  Early walking ability, whether generalized as a binary classification or qualified using an ordinal walking score, was an independent predictor of being discharged home. Those with some ambulatory function had 9.29 times greater odds (95% CI: 5.98 – 14.66, p < 0.001) than non-ambulators for being discharged home. Similarly, for every increment in the 6-point walking score, patients had 2.04 times greater odds (95% CI: 1.79 – 2.35, p < 0.001) than the previous   38 level of walking ability for being discharged home. However, for every point increase of the 5-point leg motor impairment score, the odds of returning home decreased by 22% (OR: 0.78, 95% CI: 0.65 – 0.92, p = 0.005). In Model B (Table 2.4), in addition to walking ability and stroke severity, a stroke affecting subcortical structures of the brain was also associated with lower odds of returning home, relative to a cortical stroke (OR: 0.53, 95% CI: 0.32 – 0.86, p = 0.01). Figure 2.1 Receiver operating characteristic curves for Model A and Model B   2.4 Discussion This study aimed to provide an updated prevalence of lower extremity and walking impairment for patients hospitalized with a first-ever stroke, as well as to characterize the predictive nature of early walking ability for being discharged home. In total, 44.1% of patients presented with some degree of leg impairment when first admitted to hospital with acute stroke. Similarly, 46.0% – 57.9% were unable to walk without varying levels of physical assistance. Just   39 over half (50.9%) of patients were discharged home, a large percentage (81%) of whom had some early walking ability. Whether broadly generalized or quantified at various levels of dependence, walking ability was a significant and independent predictor of being discharged home.   The prevalence findings in this study are slightly lower for leg impairment (44.1%) and walking impairment (57.9%) compared to those presented in the Copenhagen Stroke Study (65% for leg impairment and 63% for walking impairment) (Jørgensen et al., 1995). However, the Copenhagen Stroke study included participants who died during acute hospitalization (Jørgensen et al., 1995), which may have inflated their prevalence estimates relative to ours; assuming those who died presented with walking limitations, exclusion of these individuals would have resulted in an adjusted walking impairment prevalence of 54%. Furthermore, our prevalence estimates aligned with reported baseline walking limitation (56 – 60%) in recent large-scale acute intervention studies (AVERT Trial Collaboration group et al., 2015; Hankey et al., 2020), suggesting that rates of walking dysfunction after stroke have remained stable, despite changes in medical intervention, stroke incidence patterns, and mortality rates.  The finding that leg impairment, early walking ability, and initial stroke severity are predictive of home discharge after acute hospitalization may seem intuitive. Indeed, several studies have shown that the NIHSS, taken at admission, can predict discharge to rehabilitation or home (Dutrieux et al., 2016; Rundek et al., 2000; Schlegel et al., 2004). However, recent research has posited that admission NIHSS is less useful in an era of modern thrombolytic interventions (Reznik et al., 2018). Our model using a binary variable of walking ability and initial NIHSS yielded excellent discriminative ability of home discharge, indicating that inclusion of early walking ability may be sufficient to account for potential post-intervention   40 changes in stroke severity. We also showed that initial leg impairment alone independently predicts discharge home, even after accounting for walking ability in a multivariate model. This is supported by a previous study which found that certain individual items on the NIHSS can predict discharge disposition in minor stroke (Yaghi et al., 2016); the authors of that study also suggested that pairing ambulation with the NIHSS could aid in making treatment decisions. Our study emphasizes the importance of walking at the beginning of the overall trajectory of stroke recovery.  The present study reinforces the utility of completing a walking assessment as early as possible after stroke. While the AlphaFIM taken during acute care has been shown to predict functional outcomes after inpatient rehabilitation (Lo et al., 2012; Stillman et al., 2009), our study shows that the walking domain of the measure alone can be used to predict discharge to home after acute hospitalization. Therapists regularly assess walking ability and can do so quickly; knowing that the odds for being discharged home are 2.04 times greater for every point increase in walking ability may help therapists manage expectations and discharge planning from an earlier standpoint. It is important to note that this analysis does not necessarily convey causality; it should not be interpreted that attempting to improve walking ability during early therapy will lead to better discharge outcomes. Indeed, this was proven otherwise by a large multi-site randomized controlled trial, in which very early mobilization led to poorer outcomes (AVERT Trial Collaboration group et al., 2015). This analysis adds to the existing literature in predicting home discharge after acute hospitalization. Though it may seem obvious that early walking ability is predictive of being discharged home after stroke, it cannot be understated that accurately streamlining patients to the home setting after stroke is of utmost importance. Home-time is defined as the number of days   41 spent at home outside of the acute care setting in the first 90 days post-stroke, and has been shown to be associated with improved functional outcomes in the subacute and chronic phase of stroke (Quinn et al., 2008; Sung et al., 2020). No matter how it is predicted, accurately preparing patients to return home from the acute hospitalization period without further institutionalization may reduce healthcare costs and optimize resource utilization in the long-term (Dewilde et al., 2020). Furthermore, being able to predict discharge home as early and accurately as possible is increasingly important as the push for social distancing and at-home telerehabilitation grows (Bashir, 2020).        A strength of this study is the high discriminative power of a predictive model centered on early leg and walking impairment in a large cohort of patients with stroke. Previous research has identified ethnicity, spousal support, and presence of other comorbidities as key determinants of discharge disposition during the acute hospitalization period (Béjot et al., 2012; Dutrieux et al., 2016; Tanwir et al., 2014). Other impairment outcomes which were not accounted for in this analysis, such as cognition and mood, have also been identified as predictors of discharge disposition (Cho et al., 2017; Geubbels et al., 2015). However, without these additional variables, the predictive ability of our model still falls within the range of reported Nagelkerke R2 in other studies (0.11 – 0.63)  (Dutrieux et al., 2016; Stillman et al., 2009). We suggest that future studies predicting discharge disposition should focus explicitly on those with more severe walking deficits after stroke, as those patients will require the most support whether at home or in another institution.   2.4.1 Limitations This study has several limitations. Generalizability to a more diverse stroke population is reduced by the exclusion of those patients who experienced a recurrent stroke, were already in   42 hospital for other medical reasons, or were admitted more than 48 hours after their stroke. A moderate proportion of admissions had missing data and were excluded; although the patients that were excluded did not differ from the analyzed sample in demographics or stroke characteristics, it is possible that they may have differed in stroke severity or walking ability and a complete dataset would have resulted in different findings. Finally, data were collected from a large hospital with a specialized stroke unit that admits patients from surrounding geographical regions. As such, the analyzed sample may not be representative of the typical stroke admissions to a local acute care hospital.   2.4.2 Conclusion In conclusion, approximately half of patients experiencing a first-ever stroke will have lower extremity weakness and experience walking limitations. Early walking ability is a significant predictor of returning home after acute hospitalization, independent of stroke severity. Knowing this may allow clinicians to begin discharge planning early after stroke with the simple and familiar assessment of walking function.      43 Bridging Statement I Chapter 2 provided updated prevalence statistics of leg and walking impairment after stroke using consecutive admissions from a large tertiary hospital within Canada. Approximately half of patients surviving a first-ever stroke experience leg weakness and are unable to walk at the onset of stroke. Furthermore, 79% of those who are non-ambulatory do not return home after the acute hospitalization period. Chapter 2 also showed that retaining or regaining the ability to walk within the first days of stroke is a critical predictor of discharge disposition after acute hospitalization, adding to the list of other health and functional outcomes predicted by walking ability. Knowing that the prevalence of walking limitations is high and that a large portion of those who are unable to walk will continue on to inpatient rehabilitation, it is important to optimize gait training interventions to target walking recovery during the subacute phase of stroke. Powered robotic exoskeletons are novel electromechanical gait training devices that can be used overground, even with patients requiring significant physical assistance, and so they may be particularly beneficial in stroke rehabilitation. Chapter 3 describes a scoping review that was conducted to map the available literature surrounding the use of powered exoskeletons in stroke rehabilitation for walking recovery. This review sought to summarize the targeted population, training protocols, safety concerns, and preliminary efficacy of exoskeleton-based gait training after stroke in order to make preliminary recommendations for clinical use and future research.     44 Chapter 3: Powered robotic exoskeletons in post-stroke rehabilitation of gait: a scoping review A version of this chapter has been published as Louie DR, Eng JJ. Powered robotic exoskeletons in post-stroke rehabilitation of gait: a scoping review. J Neuroeng Rehabil. 2016;13:53 (doi: 10.1186/s12984-016-0162-5).  3.1 Introduction Stroke is a leading cause of acquired disability globally, with increasing survival rates as medical care and treatment techniques improve (Feigin et al., 2015). This equates to an increasing population with stroke-related disability (Feigin et al., 2015; Mukherjee and Patil, 2011), who experience limitations in communication, activities of daily living, and mobility (Kwakkel and Kollen, 2013). A majority of this population ranks recovering the ability to walk among their top rehabilitation goals (Bohannon et al., 1988; Harris and Eng, 2004); furthermore, the ability to walk is a determining factor as to whether an individual is able to return home after their stroke (Portelli et al., 2005). However, 30 – 40% of stroke survivors have limited or no walking ability even after rehabilitation (Jørgensen et al., 1995; Kollen et al., 2006), and so there is an ongoing need to advance the efficacy of gait rehabilitation for stroke survivors. Powered robotic exoskeletons are a recently developed technology that allows individuals with lower extremity weakness to walk (Chen et al., 2013). These wearable robots strap to the legs and have electrically actuated motors that control joint motion to automate overground walking. Powered exoskeletons were originally designed to be used as an assistive device to allow individuals with complete spinal cord injury to walk (Sale et al., 2012). However, because they allow for walking without overhead body weight support or a treadmill, they have gained   45 attention as an alternate intervention for gait rehabilitation in other populations, such as stroke, where repetitive gait training has been shown to yield improvements in walking function (French et al., 2010; Scrivener et al., 2012). Several powered exoskeletons are already commercially available, such as the EksoGT (Ekso Bionics, Richmond, California, USA), ReWalk (ReWalk Robotics, Yokneam, Israel), and Indego (Parker Hannifin, Macedonia, Ohio, USA) exoskeletons, with more being developed. There have been many forms of gait retraining proposed for stroke survivors. Conventional gait rehabilitation provided in physical therapy leads to improvements in speed and endurance (States et al., 2009), particularly when conducted early post-stroke and at high intensity (Hornby et al., 2016). However, conventional gait retraining using hands-on assistance can be taxing on therapists; the number of steps actually taken in a session reflects this and has been shown to be low in subacute inpatient rehabilitation (Rand and Eng, 2012). Many of the proposed technology-based gait intervention strategies have focused on reducing the physical strain to therapists while increasing the amount of walking repetition that individuals undergo. For example, body weight-supported treadmill training (BWSTT) allows therapists to manually move the hemiparetic leg in a cyclical motion while the patient’s trunk and weight are partially supported by an overhead harness system; this has shown improvements in stroke survivors’ gait speed and endurance compared to conventional gait training (Mehrholz et al., 2014), yet still places a high physical demand on therapists. Advances in technology have led to treadmill-based robotics, such as the Lokomat (Hocoma, Zurich, Switzerland), LOPES (University of Twente, Enschede, Netherlands), and G-EO (Reha Technology, Olten, Switzerland), which have bracing that attaches to the patient’s legs to take them through a walking motion on the treadmill. The appeal of this technology is that it can provide substantially higher repetitions for walking   46 practice than BWSTT without placing strain on therapists; however, there is conflicting evidence regarding the efficacy of treadmill-based robotics for gait training compared to conventional therapy or BWSTT. Some studies have shown that treadmill robotics improve walking independence in stroke (Ada et al., 2010; Mehrholz et al., 2013), but do not improve speed or endurance (Hidler et al., 2009; Mehrholz et al., 2013). There has been some sentiment that such technology has not lived up to the expectations originally predicted based on theory and practice (Dobkin and Duncan, 2012). One argument is that these treadmill robotics with a pre-set belt speed, combined with body weight support, create an environment where the patient has less control over the initiation of each step (Turchetti et al., 2014); another argument against treadmill-based gait training is the lack of variability in visuospatial flow, which is an essential challenge of overground walking (Dobkin and Duncan, 2012). Powered robotic exoskeletons, though similar in structure to treadmill-based robotics, differ in that they require active participation from the user for both swing initiation and foot placement; for example, some exoskeletons have control strategies which will only assist the stepping motion when it detects adequate lateral weight-shifting (Chen et al., 2013). Furthermore, because the powered exoskeletons are used for overground walking, it requires the user to be responsible for maintaining trunk and balance control, as well as navigating their path over varying surfaces. While these powered exoskeletons hold promise, the literature surrounding their use for gait training is only just beginning to gather, with the majority focusing on spinal cord injury (Esquenazi et al., 2012; Kolakowsky-Hayner et al., 2013; Kressler et al., 2014). Several systematic reviews have shown safe usage, positive effects as an assistive device, and exercise benefits for individuals with spinal cord injury (Arazpour et al., 2015; Federici et al., 2015; Louie et al., 2015). Only one systematic review specifically focusing on powered exoskeletons   47 has included studies involving participants with stroke (Wall et al., 2015), though studies in spinal cord injury and other conditions were also included. This review focused exclusively on the Hybrid Assistive Limb (HAL) exoskeleton (Cyberdyne, Tsukuba, Japan), which currently is not approved for clinical use outside of Japan, and found beneficial effects on gait function and walking independence; however, the results were combined generally across all included patient populations and not specifically for stroke. Given that this is a relatively new intervention for stroke, the objective of this scoping review was to map the current literature surrounding the use of powered robotic exoskeletons for gait rehabilitation in post-stroke individuals and to identify gaps in the research. The second objective of this scoping review was to preliminarily explore the efficacy of exoskeleton-based gait rehabilitation in stroke. With research of this technology at its infancy, a scoping review can help guide future research and propose recommendations for advancing the technology. 3.2 Methods This scoping review was conducted in accordance with the framework proposed by Arksey and O’Malley (Arksey and O’Malley, 2005) and guided by the refined process highlighted by Levac et al. (Levac et al., 2010).  3.2.1 Summary of search strategy OVID MEDLINE, Embase, Cochrane Central Register of Controlled Trials, PubMed, and CINAHL databases were accessed and searched from inception on October 14, 2015. We combined the search terms (robot* OR exoskeleton OR “powered gait orthosis” OR PGO OR HAL OR “hybrid assistive limb” OR ReWalk OR Ekso OR Indego) AND (stroke OR CVA OR “cerebrovascular accident” OR “cerebral infarct” OR “cerebral hemorrhage” OR hemiplegia OR   48 hemiparesis OR ABI OR “acquired brain injury”) AND (gait OR walk OR walking OR ambulation), with humans and English language as limits.  3.2.2 Inclusion and exclusion criteria Inclusion criteria were full-text, peer-reviewed articles that used a powered robotic exoskeleton with adults post-stroke as an intervention for gait rehabilitation. Articles were included if they reported functional walking outcomes (e.g., speed, distance, independence). We defined a powered robotic exoskeleton as a wearable robotic device which actuates movement of at least one joint while walking, either unilaterally or bilaterally. We further defined powered robotic exoskeletons as stand-alone devices (i.e., independent from a treadmill) that can be used for overground walking, with programmable control. Articles were excluded if they: reported only technology development; reported only electromyography, physiological cost, or joint kinematic data; combined other interventions (e.g., functional electrical stimulation); included healthy participants or children; utilized a treadmill-based device (i.e., the exoskeleton and treadmill are a single device, where the exoskeleton cannot be used separately overground); included mixed diagnosis participants (<50% stroke); or if only an abstract was available. 3.2.3 Study selection and data extraction Titles and abstracts were screened for relevance by two reviewers (DRL, CC) according to the inclusion and exclusion criteria above. In the event of conflict, a third reviewer (JJE) was consulted for resolution. Full-texts were then screened and reference lists of all selected articles were searched for additional studies. Included articles were then examined to extract data regarding study design, exoskeleton device, participant characteristics, intervention, training period, outcome measures, adverse effects, and results. We examined the changes in functional   49 walking outcomes relative to clinically meaningful change values published in the literature (Table 3.1). Table 3.1 Meaningful change values for functional walking outcomes in stroke Outcome measure Subacute stroke Chronic stroke TUG Not available MDC = 2.9 s (Flansbjer et al., 2005)  6MWT MDC = 61 m (Perera et al., 2006) MCID = 34.4 m (Tang et al., 2012)  10MWT / gait speed MCID = 0.16 m/s (Tilson et al., 2010) MCID = 0.06 m/s (small) (Perera et al., 2006) MCID = 0.14 m/s (substantial) (Perera et al., 2006)  FAC Not available Not available Abbreviations: 6MWT, 6-Minute Walk Test; 10MWT, 10-Metre Walk Test; FAC, Functional Ambulation Category; MCID, minimal clinically important difference; MDC, minimal detectable change; TUG, Timed Up and Go  3.3 Results As seen in Figure 3.1, our electronic database search returned 440 unique titles. Only one additional article was identified through reference list searching. After screening titles, abstracts, and full-texts for eligibility, 11 articles were included (Bortole et al., 2015; Buesing et al., 2015; Byl, 2012; Fukuda et al., 2015; Kawamoto et al., 2013; Maeshima et al., 2011; Nilsson et al., 2014; Stein et al., 2014; Watanabe et al., 2014; Wong et al., 2012; Yoshimoto et al., 2015). All 11 articles were published in the last five years, with seven (Bortole et al., 2015; Buesing et al., 2015; Fukuda et al., 2015; Nilsson et al., 2014; Stein et al., 2014; Watanabe et al., 2014; Yoshimoto et al., 2015) published in the last two years. Five studies were conducted in the United States (Bortole et al., 2015; Buesing et al., 2015; Byl, 2012; Stein et al., 2014; Wong et   50 al., 2012), five in Japan (Fukuda et al., 2015; Kawamoto et al., 2013; Maeshima et al., 2011; Watanabe et al., 2014; Yoshimoto et al., 2015), and one in Sweden (Nilsson et al., 2014). Figure 3.1 Flow diagram of study selection process  3.3.1 Study design Of the included studies, three were randomized controlled trials (RCTs) (Buesing et al., 2015; Stein et al., 2014; Watanabe et al., 2014), and one was a non-randomized controlled study (Yoshimoto et al., 2015). The rest were a variety of single-group pre-post clinical trials as seen in   51 Table 3.2. Of the three RCTs, two were smaller in size (n=24 and n=22) and considered pilot studies (Stein et al., 2014; Watanabe et al., 2014). 3.3.2 Participants Across the 11 studies, there was a total of 216 (male/female:136/80) participants with stroke enrolled (Table 3.2), with variability in the inclusion criteria for participation. Seven studies (Bortole et al., 2015; Buesing et al., 2015; Byl, 2012; Kawamoto et al., 2013; Stein et al., 2014; Wong et al., 2012; Yoshimoto et al., 2015) included participants with chronic stroke (>6 months post-stroke). Four studies (Fukuda et al., 2015; Maeshima et al., 2011; Nilsson et al., 2014; Watanabe et al., 2014) investigated the exoskeleton with subacute participants (<6 months post-stroke) during inpatient rehabilitation. The majority of participants were in the 50 – 70 age range. Six studies (Bortole et al., 2015; Buesing et al., 2015; Byl, 2012; Stein et al., 2014; Wong et al., 2012; Yoshimoto et al., 2015) specifically enrolled participants with the ability to walk without physical assistance from a therapist (gait aids permitted), while three studies (Maeshima et al., 2011; Nilsson et al., 2014; Watanabe et al., 2014) specified a requirement of needing manual physical assistance to walk. The former studies aimed to improve mobility for ambulatory individuals with chronic stroke, whereas the latter studies sought to restore independent ambulation for participants with subacute stroke. The other two studies (Fukuda et al., 2015; Kawamoto et al., 2013) enrolled participants with a mix of functional levels.   52 Table 3.2 Summary of studies included in the review Study & Design Participants  Exoskeleton & Training Period Training Protocol Walking outcomes & Results Subacute Stroke Watanabe et al. (2014)   Unblinded RCT Subacute stroke 1 – 2 person assist ambulation (HAL group n=11, mean 58.9 days post-stroke Conventional group n=11, mean 50.6 days post-stroke) HAL – Unilateral  12 sessions over 4 weeks 20 minute sessions HAL group – gait training while wearing HAL, facilitating improvements in walking ability, partial BWS if needed; progress as able from complete assistance by device to assist-as-needed through bioelectric signal detection  Conventional group – facilitate improvements in walking ability, customized to functional level; speed and duration of walking gradually increased 1) TUG – No significant difference in improvement between groups  2) 6MWT – No significant difference in improvement between groups   3) Gait speed – No significant difference in improvement between groups  4) FAC – HAL group improved significantly (p=0.04) more than Conventional group (change of +1.1 for HAL group; change of +0.6 for Conventional group) Nilsson et al. (2014)   Pre-post study Subacute stroke 1 – 2 person assist ambulation  (n=8, 6 – 46 days post-stroke) HAL – Bilateral   5 sessions / week, median 17 sessions 25 minutes training Progression from weight shift control to bioelectric signalling control, training with BWS on treadmill; progression of speed and BWS as tolerated 1) 10MWT – median change of +0.24 m/s, 4 previously non-ambulatory progressed to ambulatory  2) FAC – median change of +1.5 (from 0 to 1.5) Fukuda et al. (2015)  Subacute stroke  HAL – Uni/bilateral  Walking on treadmill in exoskeleton, progress from 1) 10MWT – change of +0.1 m/s for Brunnstrom stage III (greater severity with lower stage)   53 Study & Design Participants  Exoskeleton & Training Period Training Protocol Walking outcomes & Results Pre-post study (n=53, 12 non-ambulatory, 41 ambulatory) 2 sessions/week, mean 3.9 sessions complete control to bioelectric signalling (n=12); no change for Brunnstrom stage IV (n=7); change of +0.1 m/s for Brunnstrom stage V (n=12); change of +0.4 m/s for Brunnstrom stage VI (N=10)   Maeshima et al. (2011)   Pre-post study Subacute stroke 1 – 2 person assist ambulation  (n=16, 27 – 116 days post-stroke) HAL – Bilateral  Single session  Walking and stair practice after standing practice in exoskeleton 1) 10MWT – positive change for 14 of 16 patients (values not provided) Chronic Stroke Buesing et al. (2015)   Single-blind RCT  Chronic stroke Limited community ambulation (SMA group – n=25, mean 7.1 years post-stroke Functional task specific training group – n=25, mean 5.4 years post-stroke) SMA – Bilateral   18 sessions over 6 – 8 weeks 45 minute sessions SMA group – 30 minutes of high intensity overground walking with SMA (12-16 RPE or 75% HR max) and 15 minutes of dynamic functional gait training with SMA (varied surfaces, multi-directional stepping, stair climbing, obstacles, community mobility)  Functional task specific training group – 15 minutes of high intensity overground walking training and 30 1) Gait speed – No significant difference in improvement between groups   54 Study & Design Participants  Exoskeleton & Training Period Training Protocol Walking outcomes & Results minutes of functional goal-based mobility training Stein et al. (2014)   Single-blind RCT Chronic stroke Independent ambulation (AlterG group n=12, mean 49.1 months post-stroke Exercise group n=12, mean 88.5 months post-stroke) AlterG – Unilateral  18 sessions over 6 weeks 60 minute sessions AlterG group – standardized overground functional tasks including transfers, stepping, turning, reaching, gait training, stairs and curbs while wearing exoskeleton  Exercise group – group exercises including relaxation, meditation, self-stretching, active range of motion of upper and lower limbs, minimal gait training (5 min/session) 1) TUG – No significant difference between groups   2) 6MWT – No significant difference in improvements between groups  3) 10MWT – No significant difference in improvement between groups Yoshimoto et al. (2015)   Non-randomized controlled trial Chronic stroke Independent ambulation (HAL group n=9, mean 92.4 months post-stroke Conventional PT group n=9, mean 80.5 months post-stroke) HAL – Unilateral   8 sessions over 8 weeks 60 minute sessions HAL group – 20 minutes of HAL walking per session, with some BWS, walking at speed 1.5-1.7 times max walking speed without device  Conventional PT group – exercise to improve walking ability including static and dynamic postural tasks, range of motion, and 20 minutes of overground walking training 1) TUG – HAL group improved significantly compared to Conventional PT group (change of -11.5 s for HAL group; change of +0.1 s for Conventional PT group)   2) 10MWT – HAL group improved significantly compared to Conventional PT group (change of +0.21 m/s for HAL group; change of -0.02 m/s for Conventional PT group)    55 Study & Design Participants  Exoskeleton & Training Period Training Protocol Walking outcomes & Results Kawamoto et al. (2013)   Pre-post study Chronic stroke (n=16, 1 – 11 years post-stroke, 8 dependent ambulatory, 8 independent ambulatory) HAL – Bilateral   16 sessions over 8 weeks 20 – 30 minutes training Overground walking with overhead harness for safety and partial BWS; gradual progression from sit-to-stand to walking (gradually increased intensity by changing speed, duration, BWS, and HAL control mechanism) 1) TUG – mean change of -1.1 s  2) 10MWT – mean change of +0.04 m/s Bortole et al. (2015)   Pre-post study  Chronic stroke Independent ambulation  (n=3; 60, 6, 11 months post-stroke) H2 – Bilateral  12 sessions over 4 weeks 30 minute sessions Overground walking over a linear track Participants in charge of speed and encouraged to walk as much as possible, with breaks 1) TUG – change of +1.7 s, -2.5 s, -2.5 s  2) 6MWT – change of -115 m, +16 m, +103 m Byl (2012)  Pre-post study Chronic stroke Independent ambulation  (n=3; 6, 1.3, 10 years post-stroke) AlterG – Unilateral  2 – 4 sessions/week over 4 weeks 90 minute sessions Walking practice, with sit-to-stand transfers, squatting, and stepping activities; obstacle clearance, uneven terrain, community ambulation, stair climbing 1) TUG – change of -6.9 s, +1.9 s, -0.2 s  2) 6MWT – change of +37 m, +47 m, +29 m  3) 10MWT – change of +0.21 m/s, +0.14 m/s, +0.20 m/s Wong et al. (2012)   Pre-post study Chronic stroke Independent ambulation  (n=3; 37, 26, 40 months post-stroke) AlterG – Unilateral   18 sessions over 6 weeks 60 minute sessions 45 minutes while wearing device, standardized weight-bearing functional mobility activities, sit-to-stand transfers, balance exercises, gait practice at various speeds 1) TUG – change of  -11.7 s, -2.3 s, -4.2 s  2) 6MWT – change of +17 m, +14 m, +15 m    56 Study & Design Participants  Exoskeleton & Training Period Training Protocol Walking outcomes & Results on different surfaces, functional task practice 3) 10MWT – change of -0.01 m/s, +0.05 m/s, +0.13 m/s Abbreviations: 6MWT, 6-Minute Walk Test; 10MWT, 10-Metre Walk Test; BWS, body weight support; FAC, Functional Ambulation Category; H2, H2 Exoskeleton; HAL, Hybrid Assistive Limb; HR, heart rate; PT, physical therapy; RCT, randomized controlled trial; RPE, rate of perceived exertion; SMA, Stride Management Assist system; TUG, Timed Up and Go Bold indicates value surpasses established meaningful change score detailed in Table 3.1.   57 3.3.3 Exoskeletons The included studies investigated a variety of exoskeletons, each having different set-ups and control mechanisms. Five studies (Byl, 2012; Stein et al., 2014; Watanabe et al., 2014; Wong et al., 2012; Yoshimoto et al., 2015) used a robotic exoskeleton unilaterally on the affected leg, while another five studies (Bortole et al., 2015; Buesing et al., 2015; Kawamoto et al., 2013; Maeshima et al., 2011; Nilsson et al., 2014) used a bilateral set-up for gait training. One study (Fukuda et al., 2015) progressed participants, as they were able, from a bilateral design to a unilateral configuration. The most studied exoskeleton was the HAL, used in six studies (Fukuda et al., 2015; Kawamoto et al., 2013; Maeshima et al., 2011; Nilsson et al., 2014; Watanabe et al., 2014; Yoshimoto et al., 2015); in these studies, participants’ hip and knee joints were electrically actuated in a walking motion. In one study (Bortole et al., 2015) the H2 exoskeleton (Technaid, Arganda del Rey, Spain), assisted the hip, knee, and ankle joints. Four studies (Buesing et al., 2015; Byl, 2012; Stein et al., 2014; Wong et al., 2012) utilized an exoskeleton powering only one joint of the lower extremity (either hip or knee, uni- or bilaterally); no studies were found in which only the ankle was actuated during gait. Control of the exoskeletons ranged from remote-control button activation (Bortole et al., 2015) to active movement control of stepping; the devices are able to detect movement intention through monitoring joint angles and limb torque (Buesing et al., 2015; Byl, 2012; Stein et al., 2014; Wong et al., 2012), or through bio-electric signalling of muscle activity (Fukuda et al., 2015; Kawamoto et al., 2013; Maeshima et al., 2011; Nilsson et al., 2014; Watanabe et al., 2014; Yoshimoto et al., 2015). All exoskeletons except the HAL provided supplementary gait assistance on an as-needed basis, in which the user generates as much of the walking movements as possible and the device provides extra torque or support to ensure step completion. The HAL   58 has two modes, one that provides complete stepping assistance and one that adapts to user force generation. Table 3.3 further details the exoskeletons, their control strategies, and the level of assistance provided. Table 3.3 Details of powered exoskeletons in this review Exoskeleton Joints actuated Stepping initiation Stepping assistance H2  (Bortole et al., 2015) Hip, knee, ankle Initiated by hand buttons on walker  Pre-set speed Assist-as-needed for swing SMA  (Buesing et al., 2015) Hip Initiated by movement  Internal sensors detect hip joint angle to regulate walking Assist-as-needed for swing HAL  (Fukuda et al., 2015; Kawamoto et al., 2013; Maeshima et al., 2011; Nilsson et al., 2014; Watanabe et al., 2014; Yoshimoto et al., 2015)  Hip, knee Initiated by movement (2 modes)  Internal sensors detect lateral weight shift  Surface electrodes detect muscle activation via bioelectric signals   Full-assistance for swing  Assist-as-needed for swing  AlterG  (Byl, 2012; Stein et al., 2014; Wong et al., 2012) Knee Initiated by movement  Internal sensors detect movement intention via variable force threshold Assist-as-needed for stance, free swing Abbreviations: AlterG, AlterG Bionic Leg, formerly Tibion Bionic Leg; H2, H2 exoskeleton; HAL, Hybrid Assistive Limb; SMA, Stride Management Assist system (Honda R&D Corporation, Saitama, Japan)    59 3.3.4 Training Period There was variability in the training period of the included studies, ranging from a single session (Maeshima et al., 2011) to several weeks (Bortole et al., 2015; Byl, 2012; Fukuda et al., 2015; Nilsson et al., 2014; Watanabe et al., 2014) or months (Buesing et al., 2015; Kawamoto et al., 2013; Stein et al., 2014; Wong et al., 2012; Yoshimoto et al., 2015) of training. Training duration lasted from 20 – 90 minutes per session, and frequency ranged from 2 – 5 sessions per week. Table 3.2 details the different training periods for each study. 3.3.5 Training Protocol The training protocol employed in each study differed, and varied depending on the study design, length of the training period, and exoskeleton used (Table 3.2). Generally, subjects were progressed as tolerated from weight-bearing functional tasks (sit-to-stand, standing balance, weight shifting) to walking practice while wearing the exoskeleton device. Two studies (Fukuda et al., 2015; Nilsson et al., 2014) had participants train on a treadmill, which allowed therapists to adjust the walking speed externally. The most detailed training protocols were described in the controlled trials (Buesing et al., 2015; Stein et al., 2014; Watanabe et al., 2014; Yoshimoto et al., 2015), wherein individuals were progressed according to various intensity guidelines such as rate of perceived exertion (Buesing et al., 2015) and non-exoskeletal walking speed (Yoshimoto et al., 2015). For example, Yoshimoto et al. (Yoshimoto et al., 2015) advanced the training speed to 1.5 – 1.7 times the maximal non-exoskeletal walking speed before each session. Several studies (Kawamoto et al., 2013; Nilsson et al., 2014; Watanabe et al., 2014; Yoshimoto et al., 2015) allowed some body weight support using an overhead harness to improve walking mechanics.   60 3.3.6 Walking measures Ten of the 11 studies included a measure of gait speed in their assessment of walking ability, either measuring it directly or via the 10-Metre Walk Test. Five studies (Bortole et al., 2015; Byl, 2012; Stein et al., 2014; Watanabe et al., 2014; Wong et al., 2012) assessed walking endurance by means of a 6-Minute Walk Test (6MWT), and seven studies (Bortole et al., 2015; Byl, 2012; Kawamoto et al., 2013; Stein et al., 2014; Watanabe et al., 2014; Wong et al., 2012; Yoshimoto et al., 2015) assessed the Timed Up and Go (TUG) test, which is a measure of functional mobility as it includes sit-to-stand and turning. Two studies (Nilsson et al., 2014; Watanabe et al., 2014) also included level of independence or assistance in their assessment of walking ability, using the Functional Ambulation Category (FAC). Participants were not wearing an exoskeleton device when assessed for the above measures in all studies, but gait aids such as canes and walkers were permitted.  3.3.7 Effectiveness of exoskeleton-based gait training Ten studies reported varying degrees of improved walking ability after exoskeleton training (Table 3.2). Of the four subacute stroke studies, only one (Watanabe et al., 2014) was a RCT (n = 22) which showed that participants using the HAL experienced a significant improvement in FAC scores compared to those receiving conventional gait rehabilitation matched for training time (medium effect size). However, they found no significant difference between the HAL intervention and conventional therapy for walking speed or endurance. One small pre-post subacute study (Nilsson et al., 2014) (n = 8) also found an improvement in the median FAC score of their subacute participants from 0 (2-person assist to walk) to 1.5 (1-person assist to walk) after exoskeleton-based gait training. Participants in the two other pre-post studies (Fukuda et al., 2015; Maeshima et al., 2011) in subacute stroke demonstrated improvements in   61 walking speed with only a few sessions, though not all of their participants demonstrated a change greater than the established minimal clinically important difference (MCID) (Table 3.1).  Across the seven studies in chronic stroke, improvements in walking ability were less apparent. In a RCT with 50 participants (Buesing et al., 2015), there was no significant difference in gait speed improvements between the exoskeleton and functional training group, though the improvement in both groups was clinically meaningful. Similarly, participants using the AlterG Bionic Leg (AlterG, Fremont, California, USA) did not demonstrate significant improvements compared to baseline or the control group after 18 training sessions in a small RCT with 24 participants (Stein et al., 2014). In contrast, a non-randomized controlled trial (Yoshimoto et al., 2015) found significant and clinically meaningful improvements in gait speed and TUG time after training using a HAL compared to conventional physical therapy; however, the control group did not receive the same number of exercise sessions. One larger pre-post study (Kawamoto et al., 2013) (n=16) did not find changes in gait speed that were beyond the established MCID (Table 3.1), while three small pre-post studies (Bortole et al., 2015; Byl, 2012; Wong et al., 2012), each with three participants, found varying results. Clinical improvements in endurance were made by four participants in two of the pre-post studies (Bortole et al., 2015; Byl, 2012), using a MCID of 34.4m in the 6MWT (Tang et al., 2012). Three participants across the three smaller pre-post studies (Bortole et al., 2015; Byl, 2012; Wong et al., 2012) made meaningful improvements in TUG scores. Four participants in two of the pre-post studies (Byl, 2012; Wong et al., 2012) demonstrated a clinically meaningful improvement in walking speed, using an MCID of 0.06 – 0.14 m/s (Perera et al., 2006).    62 3.3.8 Adverse effects Eight studies confirmed that no adverse events occurred during the course of the gait training intervention. One study (Nilsson et al., 2014) reported minor and temporary adverse effects such as skin irritation and pain from cuffs and bioelectric detection electrodes. Two studies (Fukuda et al., 2015; Maeshima et al., 2011) did not report on adverse events. No studies reported adverse effects on the therapists. 3.4 Discussion This scoping review was conducted to map the literature surrounding the use of powered robotic exoskeletons for gait retraining for individuals after stroke and to identify preliminary findings and areas where further research is required. This is a relatively new application of powered exoskeletons, as they have only recently become available for clinical use. As expected, there was only a small number of studies published relevant to this topic.  There were four different powered exoskeletons utilized amongst the included studies, ranging from unilateral, single joint devices to bilateral, multi-joint robotics with the capacity to detect volitional bioelectrical signals to initiate powered movement. Other exoskeletons exist on the commercial market for clinical application that have not yet been investigated for stroke, such as the EksoGT, ReWalk, and Indego (Parker Hannifin Corporation, USA). Research with these other exoskeletons is required to determine their clinical usefulness and would also strengthen the literature in general support of exoskeleton use for gait rehabilitation in patients with stroke. Studies comparing unilateral to bilateral designs may also be another avenue for investigating the efficacy of exoskeletal gait retraining. The majority of the included studies investigated exoskeleton-based gait training in participants with chronic stroke. However, the greatest amount of functional and neurological   63 recovery after stroke occurs in the first six weeks after stroke (Jørgensen et al., 1995; Kwakkel and Kollen, 2013). Reflecting this, all four studies in the subacute phase of stroke reported positive effects of exoskeleton training. Two studies (Nilsson et al., 2014; Watanabe et al., 2014) demonstrated improved walking independence with repeated exoskeletal gait training for more limited participants with stroke, which is in line with findings using treadmill-based devices (Ada et al., 2010). In another study (Fukuda et al., 2015), there was significant improvement in walking speed (0.4 m/s) for participants who had some voluntary motor control of the leg, but much less change (0.1 m/s) for those without. The magnitude and parameter (ability, speed) of walking improvement may vary depending on the initial functional presentation of the exoskeleton user; furthermore, the spontaneous recovery following stroke is a confounding factor for the improvements reported that has yet to be rigorously controlled for in the current literature. Study findings were not consistent for participants with chronic stroke. All participants with chronic stroke who were included were ambulatory, and so studies investigated changes in gait parameters rather than functional ability. While there were modest, but not consistent, changes in the pre-post studies, the more rigorous RCTs (Buesing et al., 2015; Stein et al., 2014) did not show a difference from their respective control groups when groups were matched for exercise time and interaction with a physical therapist. Even in studies with longer training protocols (Buesing et al., 2015; Kawamoto et al., 2013; Stein et al., 2014; Wong et al., 2012), there was not a trend for greater improvements. Despite receiving the repetitious practice that is required for motor learning (French et al., 2010; Scrivener et al., 2012), participants with chronic stroke did not respond as positively to exoskeletal gait training as subacute patients. This is consistent with findings in a systematic review of treadmill-based exoskeleton devices for gait training in ambulatory individuals with chronic stroke (Mehrholz et al., 2013). A possible   64 explanation for this is that once an individual is able to walk, they benefit more from unconstrained walking practice with greater variability and unpredictable challenges (Hornby et al., 2016). While powered exoskeletons do not require the participant to use a treadmill, they still constrain the user to a stereotyped movement pattern and may thus under-challenge them.  The majority of included studies had small sample sizes, which may have limited the power of their study findings and analysis. In addition, the majority of these studies were pilot feasibility or pre-post clinical studies; recruitment and lack of a control group may have introduced bias to their findings. For example, one study (Yoshimoto et al., 2015) used a non-randomized controlled design, where the control group was formed of participants who were less able to attend the study training protocol. The findings from this scoping review inform the preliminary evidence in the field and more rigorous, appropriately powered RCTs will continue to advance the clinical application of powered exoskeletons. 3.4.1 Future directions for research and suggestions for clinical practice From our data synthesis, we have identified various considerations when using an exoskeleton for gait retraining and propose several questions for future research: 1. Do non-ambulatory patients with chronic stroke experience the same improvement in walking ability as patients with subacute stroke when using an exoskeleton device for gait retraining? 2. How does initial functional presentation impact the nature of improvement in walking ability when using an exoskeleton device for gait rehabilitation? 3. What is the impact of different exoskeletons (number of joints actuated, level of assistance and control of stepping) on gait rehabilitation in stroke?   65 4. What is the impact of using a bilateral design compared to a unilateral design for gait rehabilitation in hemiparetic stroke? 5. What is the optimal dose of exoskeletal gait training for patients with stroke to regain the most walking ability? 6. How does overground exoskeletal gait training compare to body weight-supported treadmill training? 7. Can exoskeletons be used to safely ambulate 2-person assist patients early after stroke with minimal injury risk to therapists? Additionally, larger sample sizes and rigorous methodology investigating the efficacy of powered exoskeletons in stroke will further strengthen findings for or against their utilization for gait rehabilitation.  At the moment there is insufficient evidence to advocate in favour or against use of powered exoskeletons in clinical practice. The patient’s acuity and functional presentation need to be considered and the extent of benefit has yet to be determined through high quality research. The devices, however, have been shown to be safe and feasible for use with patients with stroke. They can be used to mobilize more impaired individuals without physically straining therapists. It thus remains up to therapists to use their own clinical judgement of whether to utilize powered exoskeletons with their patients for gait rehabilitation, considering its application for weight-bearing, standing, and automated walking. 3.4.2 Limitations There are a few limitations with the present review. This review excluded non-English studies, which may have led to an incomplete synthesis of data, given that some exoskeletons are developed in non-native English-speaking countries such as Japan, Germany, Iran, Israel, and   66 Spain. There was heterogeneity in the studies, especially with variability in the training protocols and exoskeletons utilized (control mechanism, unilateral or bilateral application), which makes interpretation of the results challenging. In addition, type, side, and severity of stroke and comorbid conditions were not considered in this review because of the scarcity of studies in this area. As more research trials in stroke rehabilitation using powered exoskeletons are conducted, a systematic review will be able to address these additional considerations. 3.4.3 Conclusion Currently, clinical trials demonstrate that powered robotic exoskeletons can be used safely as a gait training intervention for subacute and chronic stroke. Preliminary findings suggest that exoskeletal gait training is equivalent to traditional therapy for patients with chronic stroke, while subacute patients may experience added benefit from exoskeletal gait training. Efforts should be invested in designing rigorous, appropriately powered controlled trials before it can be translated into a clinical tool for gait rehabilitation post-stroke.      67 Bridging Statement II  A version of Chapter 3 was published in 2016 and was the first review to exclusively focus on using overground powered exoskeletons for gait retraining after stroke. At the time, the review revealed that only a modest number of clinical studies had been conducted, the majority of which had focused on ambulatory patients with chronic stroke. Only one RCT had been conducted in subacute stroke, which utilized a single-leg exoskeleton with patients requiring minimal assistance to walk. Furthermore, no studies had utilized the bilateral multi-joint exoskeletons that were commercially available in Canada. With the review in Chapter 3 highlighting several directions for future research, including the need for controlled trials comparing exoskeleton-based interventions to conventional gait therapy during subacute rehabilitation and a greater focus on non-ambulatory patients with stroke, the stage was set for the Exoskeleton for post-Stroke Recovery of Ambulation (ExStRA) mixed methods study (Chapter 4 and 5).  Mixed methods research designs include both quantitative and qualitative methods and data in order to obtain a richer and deeper understanding of a phenomenon, experience, process, or intervention (Zhang and Creswell, 2013). This mixing allows researchers to capitalize on the strengths of each method, and is particularly useful in rehabilitation and physical therapy research (Kroll and Morris, 2009; Rauscher and Greenfield, 2009); for intervention research, mixed methods research may help with understanding the success or failure of rehabilitation treatments and the acceptability of new interventions. To this end, we conducted a concurrent nested mixed methods trial to investigate both the efficacy and acceptability of exoskeleton-based gait rehabilitation during subacute stroke rehabilitation.   68  Chapter 4 describes the quantitative component of the ExStRA study — a multi-site RCT comparing an exoskeleton-based physical therapy program to standard physical therapy for non-ambulatory patients during subacute stroke rehabilitation. The RCT addressed several gaps in the existing literature identified in the scoping review in that it: 1) replaced patients’ standard physical therapy program with the exoskeleton intervention; 2) utilized the commercially available EksoGT (first device to receive federal approval for use in acquired brain injury); and 3) only recruited non-ambulatory patients. It also included secondary outcomes surrounding mood, cognition, and quality of life that were not reported in the previous studies identified in Chapter 3.    69 Chapter 4: Efficacy of an exoskeleton-based physical therapy program for non-ambulatory patients during subacute stroke rehabilitation: a randomized controlled multi-site trial 4.1 Introduction Recovering the ability to walk is commonly cited by patients with stroke as a top priority of both rehabilitation and research efforts (Harris and Eng, 2004; Rudberg et al., 2020). Besides predicting long-term mobility and community reintegration after stroke (Bijleveld-Uitman et al., 2013; Mayo et al., 1999), walking outcomes are also associated with cognitive performance, post-stroke depression, and quality of life (Hackett et al., 2014; Liu-Ambrose et al., 2007; Min and Min, 2015). With implications for so many post-stroke outcomes, refining rehabilitation efforts to optimize the timeliness and degree of walking recovery after stroke remains a top priority. It is recommended that early stroke rehabilitation should be goal-oriented, intensive, repetitive, and task-specific to take advantage of neuroplastic recovery and make gains in mobility and walking (Hebert et al., 2016; Krakauer et al., 2012). However, dependent patients requiring substantial assistance from one or two therapists are the least likely to achieve these guidelines and take very few steps during rehabilitation; studies have reported as low as 8 minutes of physical activity and 6 – 16 steps completed during physical therapy (Lacroix et al., 2016; Rand and Eng, 2012). Even with the introduction of body weight-supported treadmill training in stroke rehabilitation and promising findings for walking recovery, the physical demand of having multiple therapists involved to assist moving the lower extremities has dampened its clinical utility (Bogey and Hornby, 2007; Mehrholz et al., 2017). With such   70 minimal levels of walking practice, it is unsurprising that nearly half of patients admitted for stroke are discharged from rehabilitation without the ability to walk independently (Jørgensen et al., 1995), which in turn influences meaningful outcomes such as discharge location and return-to-work (Jarvis et al., 2019; Mayo et al., 1999). Consequently, it is those patients who are more impaired and unable to walk independently who should be the target of novel interventions and research (Dobkin and Duncan, 2012). High intensity and repetition should be included as key components of such efforts. Powered exoskeletons have been commercially developed to assist and automate overground walking for individuals with lower extremity weakness. Such devices strap around the lower limbs and generate joint motion using embedded motors. They may allow patients to achieve the higher duration and repetition of stepping practice recommended for stroke rehabilitation, while offloading therapists’ physical burden. However, previous research of treadmill-based robotic devices (e.g., Lokomat (Hocoma, Zurich, Switzerland)) has found mixed results for gait recovery; several randomized trials did not find superior effects of robotic training on walking outcomes (Hidler et al., 2009; Mayr et al., 2018), yet several reviews have found improved walking independence (Bruni et al., 2018; Mehrholz et al., 2013). Powered exoskeletons may offer more realistic task-specific and goal-oriented overground walking practice than treadmill-based devices, as they address the criticism that suspended robotic devices lack variability in movement and encourage passive participation (Dobkin and Duncan, 2012). Early research of powered exoskeletons in stroke rehabilitation has shown promising findings, though only a few randomized controlled studies have been conducted and none have focused explicitly on non-ambulatory patients during the subacute phase of recovery (Louie and Eng, 2016). From the previous reviews of electromechanically-assisted gait training, it has been   71 recommended that further research and therapy with robotics should only be used with patients in the early phase of stroke recovery and who require more physical assistance to walk. (Dobkin and Duncan, 2012; Hornby et al., 2020; Mehrholz et al., 2013). Although early research of powered exoskeletons has shown they can be safely used in stroke rehabilitation as an adjunct therapy (Louie and Eng, 2016; Molteni et al., 2017), limited research has investigated their effect when integrated within the standard physical therapy component of subacute stroke rehabilitation. The primary aim of this study was to assess the effect of an exoskeleton-based physical therapy program on the recovery of walking ability in the early phase of rehabilitation after stroke. The primary hypothesis was that non-ambulatory patients who regularly utilized an exoskeleton during their physical therapy sessions would have greater walking independence at discharge compared to patients who received standard physical therapy. The secondary aims were to evaluate the effect of exoskeleton-based physical therapy on other walking outcomes (e.g., speed), leg motor impairment, balance, cognition, post-stroke depression, and quality of life, at discharge and after 6 months.  4.2 Methods The full protocol and design of this multi-site, parallel-group, randomized controlled trial have been described elsewhere (Louie et al., 2020). Approval was granted by each respective local research ethics board and operational institute to conduct the study in Vancouver, Edmonton, and London, Canada. The study and intervention are reported using the Consolidated Standards of Reporting Trials (CONSORT) and Template for Intervention Description and Replication guidelines (TIDieR) (Hoffmann et al., 2014; Schulz et al., 2010). Both checklists are provided in Appendix C.   72 4.2.1 Participants Participants were recruited from the following three inpatient rehabilitation hospitals: GF Strong Rehabilitation Centre (May 2017 – March 2020), Glenrose Rehabilitation Hospital (December 2017 – August 2019), and Parkwood Institute (November 2018 – March 2020). Inclusion criteria for the study were: age of 19 years or older; stroke within the last 12 weeks (ischemic or hemorrhagic); one-sided hemiparesis; requiring moderate or maximal assistance to walk according to the Functional Ambulation Category (Holden et al., 1984); ability to understand and follow directions in English; ability to communicate (yes/no verbal or physical indication); and scheduled to receive physical therapy. Individuals were excluded if they had: a significant musculoskeletal or other neurological condition affecting mobility; or co-morbidities that would preclude activity (e.g., cardiovascular contraindications, pain which was intolerably worsened with exercise). Participants were also excluded if they were unable to walk prior to their stroke or had any contraindications to using the exoskeleton (e.g., pregnancy, leg length discrepancy, height/weight restrictions, open ulcerations at device contact points, etc.).  Following baseline testing, participants were randomized and allocated at a one-to-one ratio to either the Exoskeleton group or Usual Care group using a third-party, online randomization service (, Interrand Inc., Ottawa, ON) which generated and concealed the allocation sequence. A permuted block design (block sizes: 2, 4) was stratified by site to control for potential differences in standard of care, and by physical function using the Berg Balance Scale (score of < 12 or ≥ 12, which is the threshold that predicts a non-ambulator to regain unassisted ambulation) (Louie and Eng, 2018). The research coordinator at each site conducted the randomization after the baseline assessment.    73 4.2.2 Exoskeleton device This study utilized the EksoGT powered exoskeleton (Ekso Bionics, Richmond, California, USA). This exoskeleton straps bilaterally to the lower extremities, and has electrically actuated hip and knee joints, and a passive spring-loaded ankle articulation which supports toe-off and foot clearance via a footplate. The EksoGT software can be programmed by the operating therapist to power the user’s lower limbs in a walking pattern, providing partial or complete assistance. Other training considerations, including gait parameters (e.g., step height, length, swing speed) and walking automaticity (i.e., how each step is triggered), can also be programmed to tailor the gait training and challenge the user according to their ability. The assistance provided to each leg can be programmed separately, further allowing clinicians to individualize gait training. The settings are described in more detail in Appendix D (D.1). The device software reports standing and walking time, as well as step counts, per use. The device manufacturer did not play any role (design, conduct, reporting) in this research study. 4.2.3 Interventions  Participants in the Exoskeleton group had 75% of their standard weekly physical therapy sessions replaced with exoskeleton-based gait training. During exoskeleton intervention sessions, participants wore the device and were guided to achieve as much repetitious stepping and walking practice as possible. Hospital therapists who were certified to use the exoskeleton (by the manufacturer) carried out the intervention, progressing the training to reduce the level of assistance provided by the device and to increase the amount of time spent walking. With previous evidence showing that robotics-assisted gait training is no more effective than overground walking for ambulatory patients with stroke (Ada et al., 2010; Mehrholz et al., 2013), therapists had the option to discontinue use of the exoskeleton once the participant   74 achieved a threshold level of independence in walking. However, they were instructed to continue focusing on gait training for 75% of their weekly physical therapy sessions if the exoskeleton was discontinued. Guidelines for adapting and progressing gait training using the exoskeleton device, suggested training duration and step count targets (≥ 25 minutes of walking and ≥ 700 steps per session by the fourth week of exoskeletal gait training), as well as an algorithm to assist decision-making to discontinue use of the exoskeleton, were provided to intervention therapists (Appendix D (D.2 – D.5)). The remaining 25% of weekly physical therapy sessions allowed the therapist to work on other goals of their choice (e.g., discharge planning, upper extremity, pain management). Therapists monitored participants for adverse events before, during, and after each training session.  Usual Care participants received standard physical therapy care throughout their rehabilitation stay. Though standard of care varied between sites, patients typically received physical therapy 4 – 5 days a week, for 45 – 60 minutes per session. Therapists providing Usual Care were not provided specific instructions or limitations, other than avoiding use of the robotic exoskeleton. Physical therapy during stroke rehabilitation is typically provided with patient-specific goals in mind and typically focuses on mobility and gait training.  The respective interventions were delivered to both groups until discharge, to a maximum of 8 weeks; this 8-week maximum duration was selected to reflect recommended and actual rehabilitation stay (Canadian Stroke Network, 2011; Hebert et al., 2016). Time spent physically upright (standing or walking, regardless of assistance) and step count during physical therapy sessions were monitored twice a week using the activPAL3 micro (PAL Technologies, Glasgow, UK) activity tracker to compare the amount of mobility practice provided to each group. Participants in either group who were not yet discharged by 8 weeks received standard physical   75 therapy care (without any exoskeleton use) beyond the intervention period, at the discretion of their care team. 4.2.4 Outcomes  Participants were assessed at recruitment (baseline), at discharge (or after 8 weeks of the intervention), and at 6-month follow-up by a blinded assessor.  The primary outcome was post-intervention walking ability, measured using the Functional Ambulation Category (FAC). The FAC is a 6-item ordinal scale that classifies the level of support needed to walk safely, ranging from 0 (unable to walk without the assistance of two people) to 5 (independent walking on uneven surfaces and on stairs). The FAC has good test-retest reliability and validity in the post-stroke population and is responsive to change in the subacute phase of stroke (Mehrholz et al., 2007); additionally, it can still be scored for individuals who are unable to walk independently, unlike measures of speed or distance. By definition of its values, each gradation of the FAC is an inherently clinically important difference.  Secondary walking outcomes were gait speed measured over 5 metres (Salbach et al., 2001), distance walked in 6 minutes (6MWT) (Fulk et al., 2008), and number of days during the intervention to achieve unassisted ambulation (FAC ≥ 3). Secondary measures also included lower extremity impairment using the Fugl-Meyer Assessment (FMA-LE) (Fugl-Meyer et al., 1975), the Berg Balance Scale (BBS) (Berg et al., 1989), the Montreal Cognitive Assessment (MoCA) (Nasreddine et al., 2005), the Patient Health Questionnaire (PHQ-9) for depressive symptoms (de Man-van Ginkel et al., 2012), and quality of life using the Medical Outcomes Short-Form 36 (SF-36) (Hays and Morales, 2001). More information about these outcome measures is provided in Appendix E.   76 4.2.5 Statistical analysis Data were analyzed using RStudio (Version 1.3.959) (RStudio, Boston, MA, USA) running on R (Version 4.0.1) (R Foundation for Statistical Computing, Vienna, Austria). A sample script of the analysis is provided in Appendix F (F.1). Descriptive statistics were expressed as mean (standard deviation, SD) for continuous variables, median (interquartile range, IQR) for ordinal variables, and counts (percentages) for categorical variables. Following the intention-to-treat principle, participants were analyzed according to their original treatment allocation. Between-group differences in primary and secondary walking outcomes at discharge and 6-month follow-up were analyzed using independent t-test for continuous variables and Mann-Whitney U test for ordinal or non-normally distributed continuous variables (Shapiro-Wilk test p < 0.05). A gait speed of 0 m/s and 6MWT distance of 0 m were appended for those who were unable to complete the respective measures at discharge and at 6-month follow-up. Differences in secondary outcomes of impairment, balance, cognition, mood, and quality of life were examined using analysis of covariance (ANCOVA), using baseline score as covariate. We addressed two missing data points at post-intervention by carrying forward baseline observation, while last observation carried forward was used for missing data points at 6 months (Jørgensen et al., 2014). A sensitivity analysis was performed on the missing data, comparing +/- 25% of the last value carried forward, for any intention-to-treat comparison with significant findings. Significance was set at p < 0.05. We also performed a per-protocol analysis, as some of the participants in the intervention group declined further use of the exoskeleton and instead received standard physical therapy until their discharge assessment. Participants who underwent less than 70% of possible   77 exoskeleton sessions for the time they were in the trial were re-analyzed as part of the Usual Care group. Calculation of sample size was informed by a previous study (Mehrholz et al., 2007) examining the FAC in subacute stroke and assumption of a 2-point difference in improvement during rehabilitation, requiring 20 individuals in each group with power set to 80%. 4.3 Results Thirty-six participants were recruited and randomized between 5 May 2017 and 9 March 2020. Due to a suspension of research activities as a result of COVID-19, the trial was terminated early. The flow of participants through the trial is displayed in Figure 4.1. Five participants dropped out from the Exoskeleton group and received standard physical therapy for the remainder of the intervention period. Of these, three participants reported simply not liking the device and did not wish to continue the training. One participant reported knee pain which persisted only while using the device, which could not be resolved through sizing or kinematic adjustments. Another participant reported severe fatigue as their reason for discontinuing the exoskeleton. All but one participant in each group was assessed for the discharge evaluation. Seven additional participants were lost to follow-up for the 6-month evaluation, either declining or were unable to be reached.   78 Figure 4.1 CONSORT flow diagram of study participants    79 Table 4.1 lists the demographic characteristics of all study participants. Fewer females participated in the study (10 females vs. 26 males), and there was a lower proportion of females randomized to the Exoskeleton group than the Usual Care group (3 (15.8%) vs. 7 (41.2%)). There were more participants with hemorrhagic stroke in the Exoskeleton group than the Usual Care group (7 (36.8%) vs. 4 (23.5%)), but fewer participants hospitalized for recurrent stroke (4 (21.1%) vs. 7 (41.2%)).  Table 4.1 Demographic characteristics  Exoskeleton group  n = 19 Usual Care group n = 17 Age, in years, mean (SD) 59.6 (15.8) 55.3 (10.6) Sex, male, n (%) 16 (84) 10 (59) Days since stroke, mean (SD) 36.7 (19.0) 40.9 (19.8) Side of paresis, left, n (%) 11 (58) 10 (59) Type of stroke   Ischemic, n (%) 12 (63) 13 (76) Hemorrhagic, n (%) 7 (37) 4 (24) Recurrent stroke, yes, n (%) 4 (21) 7 (41) NIHSS, mean (SD) 8.3 (4.2) 7.8 (2.4) Abbreviations: NIHSS, National Institute of Health Stroke Scale; SD, standard deviation  The trial intervention period lasted a mean (SD) of 49 (11) days, with 12 participants reaching the 8-week maximum intervention duration. Participants in the Exoskeleton group underwent a mean (SD) of 11 (5) training sessions in the first 27 (16) days in the trial, at 2.85 (0.41) sessions per week. Exoskeleton participants performed a mean (SD) of 592 (332) steps per physical therapy session, while Usual Care participants performed 330 (355) steps per session. Exoskeleton participants were physically upright for a mean (SD) of 33.4 (7.6) minutes per intervention session, compared to 21.8 (6.0) minutes for the Usual Care group.  Apart from the above-mentioned reasons for dropouts from the Exoskeleton group, no other notable adverse events relating to the exoskeleton were reported. Three additional   80 participants experienced transient pain or discomfort while using the exoskeleton, which did not affect their intervention adherence, that was easily resolved within the session through device sizing adjustments. One participant in the Usual Care group experienced a second stroke at the end of their rehabilitation stay and was re-admitted to acute care. Table 4.2 shows the median FAC score for each group at baseline, discharge, and 6 months. Improvements in walking independence was observed in both the Exoskeleton and Usual Care group; however, a comparison of change score from baseline, at both discharge and 6 months, did not reveal statistically significant differences between groups. For the secondary walking outcomes, there were no significant between-group differences at discharge or at 6 months (Table 4.3). A total of 26 participants (Exoskeleton: 12, Usual Care: 14) became ambulatory without requiring physical assistance (FAC ≥ 3) during the intervention period; for these participants, there was no difference in the time to achieve unassisted ambulation between groups. Table 4.2 Primary outcome analysis FAC Exoskeleton n = 19 Median (IQR) Usual Care n = 17 Median (IQR) p-value Baseline 0 (0 – 1) 0 (0 – 1)  Discharge 3 (1 – 4) 3 (3 – 4)  Change from baseline 2 (1 – 4) 3 (2 – 3) 0.72a 6-month 4 (2 – 5) 4 (3 – 5)  Change from baseline 3 (2 – 4) 3 (3 – 4) 0.65a Abbreviations: FAC, Functional Ambulation Category; IQR, interquartile range aAnalyzed using Mann-Whitney U test    81 Table 4.3 Secondary walking outcomes Variable Exoskeleton n = 19 Mean (SD) Usual Care n = 17 Mean (SD) p-value Gait speed, m/s Discharge 0.38 (0.3) 0.35 (0.3) 0.99a 6-month follow-up 0.52 (0.5) 0.42 (0.3) 0.74a 6MWT, m Discharge 117.0 (112.7) 93.0 (84.0) 0.72a 6-month follow-up 164.5 (152.8) 123.4 (90.1) 0.60a Days to unassisted walkingb 26.8 (13.3) 35.3 (15.7) 0.16c Abbreviations: 6MWT, 6-Minute Walk Test; SD, standard deviation aAnalyzed using Mann-Whitney U test bExoskeleton n = 12, Usual Care n = 14 cAnalyzed using independent t-test  Secondary outcomes of impairment, balance, mood, cognition, and quality of life at all timepoints are summarized in Table 4.4. After adjusting for baseline score, no significant group effects were found at either discharge or 6-month follow-up. Sensitivity analyses were not required, given the lack of significant findings, but are provided for reference in Appendix F (F.2).  Table 4.4 Secondary outcomes of impairment, balance, mood, cognition, and quality of life Variable Exoskeleton n = 19 Mean (SD) Usual Care n = 17 Mean (SD) Group difference (95% CI) F-statistic p-value FMA-Lower Baseline 17.3 (6.6) 17.5 (7.0)    Discharge 23.0 (5.9) 20.8 (7.1) 2.3  (-0.4 – 5.1) F(1,33) = 2.95 0.09 6-month  23.5 (6.0) 22.0 (5.2) 1.6  (-1.4 – 4.5) F(1,33) = 1.19 0.28 BBS Baseline 15.3 (10.0) 19.2 (15.4)    Discharge 36.6 (15.1) 37.8 (17.3) 1.4  (-8.2 – 10.9) F(1,33) = 0.086 0.77 6-month  40.3 (14.3) 43.0 (15.6) -0.5  (-9.0 – 7.9) F(1,33) = 0.017 0.90   82 Variable Exoskeleton n = 19 Mean (SD) Usual Care n = 17 Mean (SD) Group difference (95% CI) F-statistic p-value PHQ-9 Baseline 7.2 (4.2) 7.7 (6.4)    Discharge 4.1 (3.3) 6.1 (7.4) -1.6  (-4.5 – 1.3) F(1,33) = 1.289 0.26 6-month  5.1 (4.0) 6.8 (6.5) -1.4  (-3.6 – 0.8) F(1,33) = 1.599 0.22 MoCAa Baseline 22.4 (4.3) 23.5 (5.0)    Discharge 24.9 (5.2) 24.6 (5.2) 1.5  (-0.2 – 3.1) F(1,29) = 3.456 0.07 6-month  24.6 (4.8) 25.1 (4.9) 0.4  (-1.6 – 2.4) F(1,29) = 0.185 0.67 SF-36-Physicalb Baseline 30.2 (8.9) 28.2 (6.5)    Discharge 31.8 (9.7) 28.7 (8.8) 2.3  (-3.9 – 8.5) F(1,31) = 0.591 0.45 6-month  33.5 (9.9) 30.8 (10.5) 1.9  (-5.1 – 8.8) F(1,31) = 0.298 0.59 SF-36-Mentalb Baseline 51.0 (10.4) 49.4 (12.4)    Discharge 52.5 (12.6) 52.6 (14.7) -1.1  (-9.3 – 7.0) F(1,31) = 0.079 0.78 6-month  50.1 (12.5) 52.4 (13.2) -3.2  (-11.0 – 4.5) F(1,31) = 0.723 0.40 Abbreviations: BBS, Berg Balance Scale; CI, confidence interval; FMA-LE, Lower extremity component of Fugl-Meyer Assessment; MoCA, Montreal Cognitive Assessment; PHQ-9, Patient Health Questionnaire; SD, standard deviation; SF-36-Mental, Mental Component of 36-Item Short Form Survey; SF-36-Physical, Physical Component of 36-Item Short Form Survey aExoskeleton n = 17; Usual Care n = 15 bExoskeleton n = 18; Usual Care n = 16  Findings from the per-protocol analysis of primary and secondary walking outcomes are presented in Table 4.5. Participants adhering to the Exoskeleton protocol who achieved unassisted ambulation did so significantly earlier in the intervention period than Usual Care participants who became ambulatory (p = 0.03). While not different at discharge, there was a significant difference between groups at 6 months for both gait speed (p = 0.04) and 6MWT (p =   83 0.03). For other secondary outcomes (Appendix F (F.3)), there was a significant effect of exoskeleton training on FMA-LE (group difference: 3.9, 95% CI: 1.3 – 6.6, F(1,33) = 9.33, p = 0.004) and MOCA (group difference: 2.1, 95% CI: 0.6 – 3.7, F(1,29) = 7.96, p = 0.009) at discharge, though this did not carry over at the 6-month evaluation. Table 4.5 Per-protocol analysis of primary and secondary walking outcomes Variable Exoskeleton n = 14 Usual Care n = 22 p-value FAC, median (IQR) Baseline 0 (0 – 1) 0 (0 – 1)  Discharge 4 (3 – 4) 3 (2 – 4)  Change from baseline 3 (2 – 4) 2.5 (2 – 3) 0.12a 6-month follow-up 4.5 (4 – 5) 4 (2.25 – 4)  Change from baseline 4 (3 – 4.75) 3 (2 – 4) 0.09a Gait speed, m/s, mean (SD) Discharge 0.47 (0.3) 0.30 (0.3) 0.15a 6-month follow-up 0.67 (0.5) 0.35 (0.3) 0.04a 6MWT, m, mean (SD) Discharge 145.8 (110.2) 80.1 (5.0) 0.08a 6-month follow-up 211.2 (147.2) 103.0 (93.5) 0.03a Days to unassisted walkingb 24.1 (9.7) 36.7 (16.1) 0.03c Abbreviations: 6MWT, 6-Minute Walk Test; FAC, Functional Ambulation Category; IQR, interquartile range; SD, standard deviation aAnalyzed using Mann-Whitney U test bExoskeleton n = 11, Usual Care n = 15 cAnalyzed independent t-test Bold indicates a significant p-value <0.05  4.4 Discussion  An exoskeleton-based physical therapy program during subacute stroke rehabilitation did not result in greater improvements in walking independence when compared to standard physical therapy care. Secondary measures of walking function, physical impairment, balance, cognition, mood, and quality of life did not differ between groups at discharge or after 6 months using an intent-to-treat analysis.    84 A per-protocol analysis separating participants who adhered to the exoskeleton intervention suggests that exoskeleton-based training may confer a potential benefit over standard therapy. For participants who regained the ability to walk without physical assistance, those who utilized an exoskeleton did so significantly earlier during their hospital stay. Furthermore, those adhering to the exoskeleton intervention had significantly greater gait speed and endurance at 6 months compared to those receiving standard physical therapy, the magnitude of which exceeded the respective minimum clinically important difference (gait speed: 0.16 m/s, 6MWT: 71m) (Fulk and He, 2018; Tilson et al., 2010). A calculation of effect size for these significant differences indicate a large effect of exoskeleton-based physical therapy (Hedge’s g: 0.80 – 0.90) (Cohen, 1988). These findings suggest that the higher repetition of stepping and prolonged standing time are beneficial to those patients with stroke who will go on to become ambulatory. However, the number of participants lost to 6-month follow-up warrants caution in interpreting these results.  Our study adds to the emerging literature surrounding the use of powered exoskeletons in stroke rehabilitation. The majority of early research has focused on chronic stroke, establishing safe usage and modest efficacy for improving gait speed (Calabrò et al., 2018; Louie and Eng, 2016; Sczesny-Kaiser et al., 2019). Few randomized controlled trials have taken place in the subacute setting, and have often supplemented standard physical therapy with adjunct therapy time using an exoskeleton (Tan et al., 2020; Watanabe et al., 2017). In those studies, no differences were found between groups in gait speed, lower extremity impairment, or balance. Only one study showed greater walking independence with adjunct exoskeleton training (Watanabe et al., 2017). A recently published randomized controlled study substituting standard gait training with exoskeleton-assisted walking, similar to our protocol, also did not find a   85 difference between groups in improvements in the FAC, gait speed, endurance, or balance (Wall et al., 2020). Though the specific exoskeleton device differs between emerging research, the findings across studies suggest that exoskeleton-based training is not consistently or comprehensively better than standard physical therapy at the subacute phase of stroke.  We believe a strength of our study was the pragmatic nature of the intervention protocol. Whereas many interventions during subacute stroke rehabilitation are administered and studied as an adjunct therapy, it is not always feasible to apply findings to clinical practice. Many hospitals are operating within financial and staffing constraints (McHugh and Swain, 2014), wherein resources to administer additional therapy beyond conventional rehabilitation are not available. Our intervention protocol provided guidance as to the percentage of weekly therapy to replace with exoskeleton-based training, as well as criteria for discontinuing the exoskeleton. This flexible training protocol was informed by previous findings from robotics-assisted gait training, which have shown that ambulatory patients with stroke make greater improvements without robotics (Ada et al., 2010). We believe a pragmatic protocol that allows flexibility for clinical decision-making is more realistic for the downstream adoption of technology-based interventions, as physical therapists often weigh expected benefits and practicality when adopting a new intervention (Liu et al., 2015). Within this context, our trial findings support the clinical use of an exoskeleton during standard inpatient physical therapy, as it is not detrimental to patient outcomes. This may be particularly relevant in treating patients with severe disability, allowing an opportunity to practice walking that would otherwise be impractical by manual therapy alone. It is possible that the lack of significant differences is attributable to the flexibility, and thus variability, of the delivered exoskeleton-based gait intervention. Participants in our study,   86 who participated in 2.8 weekly exoskeleton sessions in place of their standard physical therapy, likely did not achieve the same daily walking practice as that provided in previous research of electromechanical devices exoskeleton (i.e., 5 days a week or as additional therapy) (Mehrholz et al., 2017; Moucheboeuf et al., 2020). Furthermore, we provided suggested training targets but allowed therapists to make their own clinical decisions; we did not strictly enforce the minimum step count during exoskeleton sessions or even after discontinuing the device. Thus, a sufficient training threshold to generate walking improvements may not have been attained. We contend that the lack of significant findings is partially explained by the low number of steps taken by study participants. Though the mean step count per physical therapy session in the Exoskeleton group was nearly double that of the Usual Care group, 592 steps is relatively low compared to other walking intervention studies. Klassen et al. conducted a RCT comparing one or two daily high-intensity physical therapy sessions to standard physical therapy care during subacute stroke rehabilitation; participants receiving either higher-intensity therapy regimens achieved greater 6MWT distances at follow-up than participants receiving standard care (Klassen et al., 2020). In that study, participants in standard physical therapy took an average of 580 steps per session, whereas the higher intensity groups achieved 2169 and 4747 steps per session (Klassen et al., 2020). Participants in a RCT conducted by Hornby et al. took 2358 steps per intervention session, and demonstrated greater gait speed and 6MWT improvements than participants in the control group (Hornby et al., 2016). Another potential explanation for a lack of training effect in our study is the low aerobic intensity of robotic-assisted gait training, relative to overground walking practice (Lefeber et al., 2020, 2018). However, many of these studies showing association between aerobic intensity or stepping amount and improved walking outcomes investigated patients requiring minimal or no assistance to walk. It would have been unlikely to   87 achieve these training targets by use of the exoskeleton in our sample of non-ambulatory patients without substantial additional resources in time and staffing. It is important to note that the lack of significant difference from standard physical therapy does not suggest that robotic exoskeletons should not be used in clinical practice. Indeed, our per-protocol findings indicate there may be benefits for walking improvement if protocol adherence is maintained. Furthermore, given that functional improvement in post-stroke walking ability is a product of neuromuscular recovery and movement compensations (Ardestani et al., 2019), there may have been differences between groups in the nature of walking improvement that were not captured by our included outcomes. Current research of the impact of robotic-assisted walking on other measures such as brain plasticity, muscle activation symmetry, and kinematic qualities of gait may support alternate reasons for utilizing an exoskeleton in stroke rehabilitation (Calabrò et al., 2018; Hsu et al., 2017; Peters et al., 2020). Furthermore, an exoskeleton may be the only option for therapists wanting to practice walking with more physically dependent patients, whether due to impairment or practical considerations (patient-to-therapist size ratio). It is also important to consider the emotional and psychological benefit of standing and practicing walking for patients after stroke. This study was conducted with a nested qualitative component of patient and therapist acceptance of exoskeleton-based physical therapy; patients viewed exoskeleton-based physical therapy highly favorably and felt a sense of greater opportunity and effectiveness with exoskeleton training (Chapter 5). These findings are supported by other qualitative studies exploring therapists’ perceptions towards the utility of an exoskeleton in general practice (Mortenson et al., 2020; Read et al., 2020). Future research involving overground exoskeletons in stroke rehabilitation is warranted. By nature of design, an exoskeleton can increase the duration and repetition of walking practice   88 while reducing therapist burden (Bruni et al., 2018), and will likely become increasingly prevalent in clinical practice. Thus, it is important for future research to focus on the identification of patients for whom an exoskeleton will truly benefit. For future trials, we recommend assessing for tolerance to the training before randomization, which may even involve a trial session in the device; this may help to exclude participants who stand no chance to benefit from the intervention by way of non-compliance. As highlighted earlier, another important area for further research is the optimal duration and intensity of exoskeleton-based training for patients with subacute stroke.  4.4.1 Limitations The most obvious limitations of this study are the small sample size and loss of participants to follow-up, which increases the risk of type II errors. Due to the small sample size, we did not control for stratification or correlation of data over time in the statistical analysis, which also increases the type II error risk. Last observation carried forward was used to address missing data, which assumes that participants remain static in functional outcomes upon discharge from rehabilitation; this often is not the case, as patients with stroke can decline or improve, depending on follow-up therapies, family support, and community environment. As well, we did not adjust for multiple comparisons in the per-protocol analysis and so the significant findings of this exploratory analysis should be interpreted with caution. Although we viewed the allowance of clinical decision-making as a strength of the protocol, the lack of enforced training targets and high variability in exoskeleton-usage between participants and trainers may have played a role in the lack of significant findings. Using walking dependency for inclusion or exclusion from the trial may have posed another limitation, as it is rated by the assessor and allows for subjectivity depending on therapist-to-patient size   89 differences, walking aids provided, and personal risk assessment. Despite participants being classified as non-ambulatory at enrolment, there were large ranges in physical impairment and balance. Finally, the male-to-female ratio was greater in the intervention group, whereas the stroke population typically has an even ratio.  4.4.2 Conclusion An exoskeleton-based physical therapy program can be safely administered and integrated within inpatient stroke rehabilitation at no detriment to patients with stroke. Exoskeleton-based physical therapy does not result in greater improvements in walking independence, but may improve other walking outcomes, when compared with standard physical therapy. Future research should focus on the identification of patients who will adhere to and benefit from exoskeleton training. As well, exoskeleton-based research should also focus on determining an optimal training regimen, in duration, repetition, and intensity.     90 Bridging Statement III  The multi-site RCT described in Chapter 4 aimed to address several gaps in the early literature surrounding exoskeleton use in stroke rehabilitation by specifically enrolling non-ambulatory patients during the subacute phase of recovery. The training protocol was unique, in that it replaced standard physical therapy care with a flexible exoskeleton-based gait rehabilitation program. Though there were no significant differences in the primary analysis, the per-protocol analysis indicated an added benefit from exoskeleton-training for certain patients with respect to walking speed and capacity. Given that a powered exoskeleton is one of only a few options for therapists to achieve repetitious walking practice with severely dependent patients, it is expected that exoskeletons will continue to be used during inpatient stroke rehabilitation. Thus, it is important to understand how patients and therapists perceive the device and training.  Chapter 5 explores the experience and acceptability of using an exoskeleton as a regimented physical therapy intervention during subacute stroke rehabilitation. Though presented distinctly from Chapter 4, this qualitative descriptive study was embedded within the RCT and conducted concurrently as part of the ExStRA mixed methods study design. This qualitative study warranted its own analysis and manuscript, as studies so far have only investigated therapists’ general thoughts towards the technology and not as a specific intervention. An integrated interpretation of the mixed methods study will be presented in Chapter 6 (general discussion of this dissertation).    91 Chapter 5: Patients’ and therapists’ experience and perception of exoskeleton-based physical therapy during subacute rehabilitation: a qualitative analysis 5.1 Introduction Physical impairments including limb weakness and limited mobility are typical sequalae of stroke (Arene and Hidler, 2009; Hankey, 2017). Up to 65% of survivors initially experience lower extremity weakness and difficulty walking (Jørgensen et al., 1995). Patients admitted to subacute rehabilitation attend intensive therapies to treat their impairments, yet only 53% of those discharged from rehabilitation are able to walk independently (Shum et al., 2014). In the chronic phase of stroke, lower extremity strength and walking ability are routinely identified as predictors of participation (Andrenelli et al., 2015; Faria-Fortini et al., 2017), community mobility (Grau-Pellicer et al., 2019; Pournajaf et al., 2019), and quality of life (Fulk et al., 2010; Grau-Pellicer et al., 2019). Unsurprisingly, research surrounding balance and walking is identified as the top priority from the perspective of people with stroke (Rudberg et al., 2020).  Best practice guidelines recommend patients with stroke to engage in goal-oriented rehabilitation training that is intensive, repetitive, and task-specific to improve mobility (Hebert et al., 2016). However, patients routinely do not reach these physical activity guidelines during stroke rehabilitation, especially those requiring assistance to walk (Lacroix et al., 2016; Skarin et al., 2013). Robotic exoskeletons, standalone electromechanical devices that fasten around the legs to assist walking, have been developed as a novel rehabilitation tool that can increase the duration and repetition of stepping practice (Gassert and Dietz, 2018). Robotic exoskeletons can be used safely for gait training after stroke (Louie and Eng, 2016), with early trials showing   92 modest benefits (Buesing et al., 2015; Watanabe et al., 2014; Yoshimoto et al., 2015). Robotic exoskeletons will likely become more prevalent in stroke rehabilitation settings, as they are further refined and future research clarifies their efficacy (Louie et al., 2020; Tsurushima et al., 2019).  Limited research has investigated the acceptability of exoskeletons in neurological rehabilitation. The sole published review of user (patient or therapist) perspectives towards exoskeleton technology was inconclusive due to the heterogeneity of study populations (Hill et al., 2017). No studies have specifically explored either patient or therapist perspectives towards an exoskeleton-based intervention in stroke rehabilitation. Recent research of physical therapists’ general experience using an exoskeleton in the clinical setting highlighted a tension between potential benefits and the challenges of utilizing an exoskeleton for day-to-day practice (Mortenson et al., 2020; Read et al., 2020; Swank et al., 2020). These perspectives may differ if the exoskeleton is employed as a focused and regimented intervention.  The Exoskeleton for post-Stroke Recovery of Ambulation (ExStRA) randomized controlled trial was conducted at several Canadian rehabilitation hospitals to determine the efficacy of exoskeleton-based physical therapy for improving walking ability during inpatient stroke rehabilitation ( NCT02995265). Non-ambulatory patients with hemiparetic stroke enrolled to the intervention group received exoskeleton-based therapy in place of their 60-minute conventional physical therapy sessions for 3 days of the week, for up to 8 weeks, using the EksoGT (Ekso Bionics, Richmond, California, USA). The results of the trial have not yet been published. Whether this novel intervention is effective, it is equally important to determine if it is well-received by the stakeholders to whom it pertains; acceptance of the exoskeleton and its application is a key component of the downstream decision to adopt the   93 technology into regular practice (Camden et al., 2015; Turchetti et al., 2014). Therefore, the objective of this study was to explore the experience and acceptability of exoskeleton-based physical therapy during inpatient stroke rehabilitation from the perspective of patients with subacute stroke and their therapists. 5.2 Methods 5.2.1 Design This cross-sectional qualitative descriptive study was conducted concurrently during the ExStRA trial using semi-structured interviews and is reported here in accordance with the Standards for Reporting Qualitative Research (O’Brien et al., 2014). A completed checklist is provided in Appendix G. A postpositivist paradigm was used, as it values individual experiences and perceptions yet assumes there is a pattern to be found amongst those individual experiences (Clark, 1998). The qualitative descriptive approach is ideal for health care research, as it allows conclusions about novel interventions to be drawn from participants’ literal description of their experience (Bradshaw et al., 2017; Sandelowski, 2010, 2000). The full methodology for the ExStRA trial and details of the exoskeleton intervention are published elsewhere (Louie et al., 2020). Ethical approval for this study was granted by all relevant research ethics boards; all participants provided informed consent to participate specifically in this qualitative study.  5.2.2 Participants A volunteer sample of patients with stroke and physical therapists were recruited from the larger ExStRA trial. Patients receiving the exoskeleton-based intervention were approached after they had participated in at least five training sessions to inform them of this qualitative study. Patients with substantial aphasia were not eligible to participate. Fourteen out of 16 eligible patients consented to participate (refusal: 1, readmission to hospital: 1). Physical   94 therapists were eligible if they provided the intervention to at least three trial participants. All six therapists who provided the intervention participated in this qualitative study. 5.2.3 Data collection For patient participants, basic demographic information was collected at enrolment to the ExStRA trial, including age, time since stroke, sex, and stroke impairment using the lower extremity subscale of the Fugl-Meyer assessment (Fugl-Meyer et al., 1975; Gladstone et al., 2002). Patients were interviewed within the week of their discharge assessment from the ExStRA trial, to allow maximal experience of the exoskeleton. For physical therapists, age, sex, years of practice, and number of intervention participants were collected. Physical therapists were interviewed after the ExStRA trial had been running for a year at their respective site to ensure familiarity with the device and intervention protocol. All participants were interviewed once. Separate interview guides were developed for patients and therapists in consultation of existing models of technology acceptance, specifically the Technology Acceptance Model (Davis, 1989) and Unified Theory of Acceptance and Use of Technology (Venkatesh et al., 2003). The lead author (DRL) developed the guides, which were reviewed by an interdisciplinary team consisting of a physiatrist, a physical therapist, and an occupational therapist. Both interview guides consisted of open-ended questions framed around the experience and acceptability of using an exoskeleton for stroke rehabilitation. Minor changes to the wording and sequencing of questions were made after review of the first three interviews. The interview guides are provided in Appendix H. Interviews were conducted by the lead author, lasting between 25 and 45 minutes, in-person or by telephone (for out-of-province participants). Except for the case of one individual whose partner was present, all interviews were conducted one-on-one in a private setting in the   95 participant’s hospital. All interviews were audio-recorded, transcribed verbatim, and reviewed for accuracy. Transcripts were anonymized by substituting participant numbers, and all files were encrypted and stored on a password-protected server. 5.2.4 Data analysis Interview transcripts were analyzed thematically, following the phases outlined by Braun and Clarke (Braun and Clarke, 2006). Three coders (DRL, ML, and WBM) became familiar with the data by repeatedly reading the interview transcripts. Initial coding was carried out separately in Microsoft Excel (Microsoft Corporation, Redmond, Washington, USA) by systematically identifying notable features in the data. A preliminary coding scheme was developed jointly by comparing and contrasting initial codes, then verified by returning to the data and applying the coding tree to the transcripts. Codes were refined and finalized through an iterative process of discussion amongst the coding team. In the analysis, patients and physical therapists were considered both as separate user groups and altogether as stakeholders. Initial themes were formed by identifying recurring patterns of thought, whether converging or contradictory, and subsequently grouping those codes together into sub-themes and main themes. These initial groupings were then considered in relation to one another as well as the entire data set, and further refined to ensure the overarching themes appropriately reflected the data and research objectives.  Though the number of eligible patient and therapist participants was limited, we believe that sufficient participants were sampled to reach coding and thematic saturation (Hennink et al., 2017; Saunders et al., 2018; Vasileiou et al., 2018), given the amount of collected data and results of the analysis relative to the research objectives (Braun and Clarke, 2019). No new codes   96 or themes were identified from later interviews (six patient interviews and two therapist interviews did not demonstrate new codes or themes).  5.2.5 Research characteristics The lead author (DRL) is a male physical therapist with prior experience conducting qualitative interviews and using the exoskeleton for clinical treatment. He acted as a coverage therapist to provide the intervention at one of the sites, but otherwise did not have a prior relationship to any of the participants. Two additional researchers were involved in analyzing the data to minimize bias. WBM is a male rehabilitation researcher with extensive qualitative research experience. ML is a female occupational therapist with clinical experience in public and private practice with neurological populations. 5.2.6 Techniques to enhance trustworthiness The primary trustworthiness strategies used in this study were reflexivity, triangulation, negative case analysis, and member checking (Mays and Pope, 2000; Morrow, 2005). The lead author kept a reflexive journal throughout the study period to note the potential impact of personal biases on the interviews. Other authors read early interview transcripts to monitor bias in questioning. Involving multiple coders and holding several meetings to discuss early analysis findings encouraged reflection on different interpretations of the data. This investigator triangulation, as well as participant triangulation (i.e., including both patient and therapist perspectives), ensured comprehensiveness in the inquiry and analysis. Granting attention to those participants with unexpected or minority views of the exoskeleton training ensured other perspectives were explored (Morrow, 2005). Findings from the analysis were shared with all participants in non-scientific writing (Appendix I); participants were invited to elaborate, affirm,   97 or challenge the provided themes in comparison to their own experience in the study. No participants voiced disagreement with the analysis findings.  5.3 Findings A total of 20 interviews (14 patients, 6 physical therapists) were conducted across three sites between September 2017 and February 2020. Table 5.1 shows the demographic characteristics of individual participants, using the identifiers “P” and “T” to denote patients with stroke and physical therapists, respectively. Ranges for age and therapist experience are provided to protect participant identity.     98 Table 5.1 Participant characteristics Patients with stroke Participant No. Sex Age range, y Time since stroke, d FMA-LE No. of exoskeleton sessions P1 M 50 – 54 23 31 17 P2 M 60 – 64 56 9 9 P3 M 60 – 64 16 16 15 P4 M 70 – 74 38 17 15 P5 F < 25 24 25 12 P6 M 45 – 49 19 21 12 P7 M 55 – 59 72 8 19 P8 F 25 – 29 35 20 8 P9 M 65 – 69 63 14 18 P10 M 80 – 84 24 26 7 P11 M 55 – 59 13 6 15 P12 M 75 – 79 11 12 11 P13 M 60 – 64 22 20 10 P14 M 60 – 64 48 17 7 Mean (SD)  57.6 (16.6) 33.1 (19.5) 17.3 (7.2) 12.5 (4.1) Physical therapists Participant No. Sex Age range, y Clinical experience, y Exoskeleton experience, y No. of patients in exoskeleton trial T1 F 40 – 44 20 – 24 3 5 T2 F 35 – 39 10 – 14 2.5 5 T3 M 40 – 44 5 – 9 2.5 3 T4 F 50 – 54 25 – 29 4 3 T5 F 25 – 29 1 – 4 1.5 5 T6 F 30 – 34 5 – 9 1.5 5 Mean (SD)  37.5 (8.4) 13 (9.6) 2.5 (0.9) 4.2 (1.3) Abbreviations: d, day; F, female; FMA-LE, Lower extremity component of Fugl-Meyer Assessment; M, male; No., number; SD, standard deviation; y, year  We identified three main themes: A matter of getting into the swing of things depicted the firsthand experience of using an exoskeleton for stroke rehabilitation; More of a positive experience than anything else described the participants’ mostly favorable attitude towards exoskeleton-based gait training; and The best step forward captured participant-identified recommendations and considerations for the future integration of exoskeleton training into stroke rehabilitation. Sub-themes that formed each theme are provided in Appendix J.   99 5.3.1 Theme 1: A matter of getting into the swing of things Patients and physical therapists consistently identified a steep learning curve associated with using an exoskeleton for stroke rehabilitation, beginning with initial experiences of difficulty, confusion, or intimidation. P12 said of his initial sessions, “it was quite intimidating … giving it up to a machine is a little bit frightening.” Therapists commented on their patients’ early struggles, with T6 stating that “some people, they just can’t get it feeling good until quite a few sessions in.” P4 suggested providing more instruction in the form of a video, to “reduce the amount of time that’s required to learn how to use the machine.” Similarly, three out of six therapists also experienced initial learning hurdles. T1 indicated, “I remember the first session being very intimidating … I had no experience working with robotics.”  Despite their initial struggles with the exoskeleton, 12 patients reported improvements in proficiency with the device after their early sessions. This was attributed to developing an understanding of how the exoskeleton operates only through experiencing its movement, as illustrated by P13:  After one good session and kinda feeling my way into it, and sensing where it was going … by the end of the hour I was starting to [understand] what was required and starting to work with the machine. Therapists similarly indicated the importance of experiencing the exoskeleton firsthand to develop their own expertise in operating the device. T6 commented: To become part of the experience, to know how it feels to be in the device, I think is critical in being able to also educate individuals who are in the device on how things may or may not feel.   100 To maximize learning and rehabilitation gains, several patients extrapolated their experience beyond their exoskeleton sessions. P13 mentioned combining mental visualization and reflection “when you take a break after the Ekso” to translate the exoskeleton experience to overground walking. Likewise, P5 “started consciously mimicking, trying to remember how it felt in the machine, to kick [her] leg out more” when walking around the ward.  The experience of using an exoskeleton for rehabilitation, however, was not without ongoing challenges related to fatigue and effort. Thirteen patients highlighted fatigue as a frequent occurrence after training sessions. P1 stated, “it was never easy in the exoskeleton … all I could do was try to just get back to bed and go to sleep because I had worked so hard in it.” P2, a patient who dropped out from the clinical trial, cited fatigue as being his primary reason for declining further use of the robotics. T1 commented on the challenge of balancing clinician and patient fatigue, saying, “the client’s pooping out too much, they’re gonna start to sit down, right? And you’re getting your own physical fatigue.” In a similar vein, participants emphasized the physical effort of using the exoskeleton. P10 found the exoskeleton “very exerting” and dropped out from the intervention to opt for less strenuous conventional therapy. T2 described the exertional challenge for therapists: It’s supposed to help you take away some of the assistance that you need, and it does, to a point… there are some patients who still require a lot of physical assistance … it can still be quite heavy for the [therapist] that’s doing the weight shifting and assisting the patient. Despite these ongoing challenges, patients frequently expressed trust in their therapist and the training process. P8 commented, “I felt all of the therapists, they all know what they’re doing… I probably could have hurt myself if I had tried to maybe get more out of it.” Therapists   101 also embraced the necessity of having to troubleshoot issues as they arose as part of the experience of using the exoskeleton in rehabilitation. T5 commented, “I think we felt comfortable problem-solving and figuring this out on our own.” 5.3.2 Theme 2: More of a positive experience than anything else Most patients reported enjoyment and satisfaction with using the exoskeleton for their stroke rehabilitation. For example, P14 said, “I looked forward to it, and every time I got it, I enjoyed it.” P9 commented, “[the exoskeleton has] been used the way it should be used.” Therapists also recognized their patients’ delight, with T4 commenting, “people think it’s cool, and neat, and are excited about it.” Therapists also spoke positively towards the provided training protocol for exoskeleton-based rehabilitation. T2 stated, “I like it … it allows me to work with, you know, what I’m seeing in the moment and what I think the Ekso could help that patient with.”  Nine patients described ways in which exoskeleton training provided an opportunity above and beyond conventional physical therapy treatment, frequently centered around the earliness and intensity of walking practice. P3 commented, “to have the opportunity … that much walking, that soon after the stroke was a real plus.” All therapists appreciated the ability to practice walking or pre-gait tasks for longer periods and at higher repetition. T6 stated, “it’s no question at all that people are up in standing and taking more steps in the device than they would outside of the device.” Therapists also identified being able to achieve earlier walking practice and better-quality movement by using the exoskeleton, especially for more dependent patients. T3 listed “dosage, higher step count, more steps, better patterning, more equal steps, more even weightbearing time” as ways in which the exoskeleton can enhance walking rehabilitation. From   102 the firsthand perspective, patients appreciated the biofeedback of experiencing walking again while also feeling safe. P5 shared, It let me feel the motions of walking again, and use a bit of my muscles, takes some weight off my own weight … I didn’t have to worry about my knee buckling or locking, it took care of that … and so, I could actually just focus purely on the motions of walking. Eleven patients strongly attributed their rehabilitation gains to exoskeleton training, citing improvements in body function (e.g., strength, balance) and walking ability (e.g., quality, speed, endurance). P6 described, “it helped me to find my balance, to help me to walk on my left foot, and it helped me with the right pattern how to walk.” Many patients believed the exoskeleton granted an accelerated rate or higher level of recovery than would have been achieved in conventional therapy. P3 stated, “without it I wouldn’t have gotten anywhere near that kind of walking, that quickly.” Therapists were less certain regarding the magnitude of benefit. Only three therapists attributed rehabilitation improvements to the use of an exoskeleton and did not generalize results to all users. T3 commented, I think there’s people who have achieved independence with ambulation who, without the Ekso sessions, would not have. I can think of two … who probably wouldn’t have been walking if it hadn’t been for Ekso.  Despite the multitude of positive sentiments towards using an exoskeleton for stroke rehabilitation, both patient and therapist participants identified limitations of the device. Critiques ranged from the device set-up to functions that could be improved. How participants felt towards the device may have impacted their general perspective towards exoskeleton   103 training. For example, P5 shared, “it wasn’t really my style of walking, so towards the end it was a bit annoying.” A list of participant critiques of the exoskeleton device is provided in Table 5.2. Table 5.2 A list of device critiques Device critiques  • Lengthy set-up process (T2, T3, T4, T5, P14) • Abnormal / unnatural gait pattern (T2, T3, T4, T6, P1, P5) • No stair ability (T1, P4) • Requirement to use a gait aid (T3, T6, P5) • Lack of configurable options to train stance phase (T5) • High cognitive demand on the patient (T4, T5) • Only forward walking (T6) • Uncomfortable device fit (hips, feet) (P13, P8, P9) • Not very responsive to patient movement (P4, P5, P7, P12) • Lacking specific live feedback of performance (P4, P13) • Not breathable / heat retention (P3) Abbreviations: P, Patient; T, Therapist   Finally, the exoskeleton therapist was identified as another factor that impacted how exoskeleton training was received. Patients identified ways in which their therapist enhanced or potentially hindered the experience of exoskeletal gait training. P6, stated, “the expertise of the person handling the device affects, really, the training and how the patient feel[s].” Patients described the ideal exoskeleton therapist as proficient, attentive, and encouraging; they cautioned against being overprotective, either by crowding too close or imparting too much restraint on the patient’s movement. Therapists also recognized their own impact on the training, with T1 commenting: I can’t do it for a long period of time, I end up just ruining the person’s gait. I’m compensating for my own self and I think I’m influencing them negatively.     104 5.3.3 Theme 3: The best step forward Almost all participants recommended the continued use of an exoskeleton in stroke rehabilitation. Twelve patients believed the exoskeleton should be used with other patients, given their own experience. P1 commented, “it was a positive experience and I would recommend it to anybody.” Another patient participant, P8, suggested a widespread adoption of exoskeleton-based rehabilitation, stating that “the program should be available at all rehabilitation centres.” Interestingly, the two patients who dropped out from the trial did not rule out the exoskeleton for all other patients with stroke, instead saying it depends on the individual. Therapists were more hesitant in generally supporting future exoskeleton training, some adding caveats to their recommendations. T5 suggested using the device with specificity, stating, “I don’t think it’s a device that can … be blanketly used for all [patients].” In contrast, T1 was more convinced: “I don’t have any doubts, I think it’s beneficial.” Four therapists indicated that they would prefer to use an exoskeleton on a trial basis, rather than committing their patients to a set training regimen. T2 suggested to “try [patients] in for a couple sessions, and if it was looking like it was helping … then continue on with it.”  Despite seeing a place for exoskeleton training in stroke rehabilitation, therapists highlighted challenges they encountered or foresaw with implementation. T3 said, “the biggest problem … is the amount of set-up time.” This was echoed by all therapists, viewing set-up time as taking away from their patient’s appointment or their own administrative schedule. Therapists commented that having support to prepare the exoskeleton device, whether from therapy assistants or the research team, made providing the exoskeleton training much easier. A list of barriers to implementing exoskeleton training in stroke rehabilitation and suggested solutions is provided in Table 5.3. Many of the solutions proposed came in the form of having more support,   105 whether from managerial leadership or fellow clinicians. As T5 commented, “having two physios that were trained in the exoskeleton working with the patient was really beneficial, just to troubleshoot around some of the more challenging issues.” Table 5.3 Barriers and solutions to implementing an exoskeleton-based training program in stroke rehabilitation Barriers Solutions • Set-up time (T6, T5, T4, T3, T1, T2, P4) • Number of staff needed for exoskeleton training (T6, T5, T4, P11) • Scheduling demands (time for set-up and multiple staff) (T6, T5, T3, T1, P4) • Lack of time for therapy outside of exoskeleton (T6, T5, T3, T1, T2, P3, P1) • Retaining patient interest in exoskeleton (T6, P5, P1) • Space requirements for using exoskeleton (T6, P14, P9, P5) • Lack of resources to use exoskeleton (T5, T4, T1, T2) • Low number of qualified/trained staff (T3, T1) • Leadership and managerial support to use exoskeleton (T1) • Pool of trained colleagues for clinical support (T1, T5) • Staffing support for set-up (T1, T3, T5, T6) • Exoskeleton intervention delivered as adjunct therapy (T5, T6) • Designated exoskeleton therapists (T6) Abbreviations: P, Patient; T, Therapist  In describing challenges experienced during exoskeleton training, several patient participants identified characteristics of the physical environment that impeded exoskeleton training. P14 commented on being distracted by the environment, stating, “it was hard to get a good flow with the exoskeleton without having to get interrupted.” For P9, “if there was more space, it would be better.”  Specifying which patients were ideal for exoskeleton training was another consideration factoring into both therapist and patient recommendations for the future. P13 commented, “it takes a certain individual … to really lever[age] the full benefit of that machine.” Participant-identified characteristics to guide patient selection are listed in Table 5.4. For therapists, the   106 utility of the exoskeleton centered around patients who they would otherwise be unable to treat effectively. T5 stated, “it’s gotta be somebody that we’re not able to mobilize successfully without the robot.” Patients drew on their own experience with the exoskeleton to make suggestions; P10 commented, “I think a younger person might adapt to it a lot quicker or better than I did.”  Table 5.4 Characteristics to guide patient selection for exoskeleton-based rehabilitation Appropriate characteristics Precautionary characteristics • Early post-stroke (P6, P3, T5, T4, T3) • High pre-stroke fitness (P12) • High activity tolerance (P9, T6) • Younger age (<75) (P9, P10, P11) • Slow progressor (T1) • Non-ambulatory / dependent (P2, P5, P6, T1, T2, T3, T4, T5, T6) • Intact trunk control (T1) • Cognitive impairments (memory, concentration, planning) (P12, P14 T2, T4, T5, T6) • Communication impairments (speech, language) (P12, P13, T1, T2, T5, T6) • Inability to follow complex commands (T6) • Visuospatial neglect or perceptual issues (T1, T2, T3, T5, T6)  • Claustrophobic (T5) • Anxious / fearful (T1) • Larger size compared to therapist (T1)  Abbreviations: P, Patient; T, Therapist  There was debate about the exact regimen of exoskeleton training, as many participants had suggestions alternative to the delivered intervention protocol. Some patients desired a more frequent and longer training period, or longer sessions. As P7 simply said, “could have used more time in the unit.” In contrast, several patients commented that they would have preferred stopping exoskeleton training sooner than they did. This may have stemmed from concerns over lacking therapy dedicated to other goals; P1 stated, “I was looking towards having more time with standard physio as we were drawing near the end of the program.” Several patients and therapists highlighted the importance of weaning down the exoskeleton training, depending on the expected benefit. T4 summarized:   107 How is it affecting them out of the unit, and is it doing what you hoped it would do? Sometimes it’s just not right for them, and it’s not good use of therapy time. … You have to use your brain as a normal practicing therapist … to look at it critically and make that judgement.  5.4 Discussion To our knowledge, this is the first qualitative study of stakeholder attitudes towards a specifically designed exoskeleton-based intervention during subacute stroke. This study highlights the ongoing learning process of using the device, the positive attitudes towards and perceived benefits of exoskeleton-based therapy, as well as considerations for the future integration of exoskeletons within stroke rehabilitation.  Similar to studies in other populations, patients with stroke developed proficiency in exoskeletal walking with repeated practice despite initial challenges. In interview-based studies of patients with spinal cord injury, researchers identified a similar learning curve to using an exoskeleton (Manns et al., 2019; Thomassen et al., 2019). This suggests the importance of committing to a minimum number of training sessions when integrating an exoskeleton into stroke rehabilitation, in order to persevere through early difficulties to attain efficient exoskeleton training. Physical therapists similarly developed expertise in operating the exoskeleton through experience, as has been found in other qualitative research focused on the feasibility of integrating an exoskeleton generally into neurological rehabilitation (Mortenson et al., 2020; Read et al., 2020). As such, physical therapists wishing to utilize an exoskeleton in their clinical practice should do so regularly to gain and maintain proficiency in operating and troubleshooting the device.   108 Our study highlights training-related fatigue and high exertion as important considerations during exoskeleton-based walking practice. While fatigue is a common symptom after stroke (Cumming et al., 2016), patients in our study described experiencing fatigue specifically due to exoskeleton training. It has been shown that exoskeletal walking has higher energy costs compared to sedentary postures, such as sitting (Asselin et al., 2015). Incorporating resistance into the exoskeleton training parameters has also been shown to increase metabolic energy expenditure (Chang et al., 2017). Thus, it is important for physical therapists to be vigilant for signs and symptoms of fatigue in their patients, especially when advancing training parameters using the exoskeleton. Therapist fatigue from operating the exoskeleton was also noted as a factor that could influence the experience of exoskeleton training and its efficiency, from the perspective of both patients and therapists. Research of therapist energy costs or injury during exoskeleton use has yet to be conducted but may be worthwhile.  The participants in our study viewed exoskeleton-based physical therapy highly favorably. Most patients believed that they were able to achieve benefits that were otherwise unattainable, while all therapists appreciated a greater sense of opportunity to retrain gait by using the exoskeleton. Though research specific to this new generation of robotic exoskeletons has yet to establish efficacy in subacute stroke, preliminary findings are promising (Louie and Eng, 2016; Molteni et al., 2017), and in line with the benefits shown by research of previous electromechanical devices (Bruni et al., 2018; Mehrholz et al., 2013). Without definitive evidence on efficacy, our findings of strong user acceptance support the future integration of exoskeleton-based gait retraining in stroke rehabilitation.   The considerations and barriers that have been proposed by the patients and therapists in our study may influence the success and uptake of an exoskeleton-based intervention in stroke   109 rehabilitation. For example, while patients were agreeable to participating in multiple sessions over many weeks, therapists had concerns about the expected benefits compared to the effort expenditure to carry out regular exoskeleton training. All therapists did not feel that carrying out an intensive exoskeleton-based therapy program was achievable outside of a research study, wherein extra staffing and scheduling resources were provided. Many of the barriers listed in Table 5.3 were felt to be managerial or logistical in nature, rather than challenges that they could resolve as individual therapists. This suggests the importance of hospital administrators and practice leaders in the sustained use of exoskeletons within the hospital setting. In line with the barriers identified in this study for widespread exoskeleton implementation, other early qualitative studies have identified therapists’ reservations in integrating an exoskeleton into practice relating to set-up time, staffing concerns, and expected effort (Heinemann et al., 2018; Mortenson et al., 2020; Swank et al., 2020). Our novel findings that a specifically designed exoskeleton intervention during stroke rehabilitation is well-received by patients and therapists can serve as a starting point for future qualitative research. Further studies should consider including administrators and managers as participants. Longitudinal qualitative research may also elucidate if and how perceptions towards exoskeleton-based stroke rehabilitation change over time, from before the intervention to months after the intervention is employed. 5.4.1 Limitations There were some limitations to this study, which may reduce the transferability of the findings to other contexts. The sample of patient and therapist participants were skewed by sex, lacking female patients with stroke and male physical therapists, though the latter skew is somewhat representative of the profession. Additionally, this study was nested within a clinical   110 trial and therefore the number of eligible participants was limited. Finally, trial participants receiving conventional physical therapy care were not interviewed, which may have offered a richer and fuller exploration of exoskeleton training by means of comparison. For example, patients in conventional care may have viewed their rehabilitation program with equal optimism as the exoskeleton participants; alternatively, they may have perceived a missed opportunity to make greater gains in physical therapy.  5.4.2 Conclusion Patients with stroke were strongly optimistic towards the experience and future of exoskeleton-based gait training during subacute stroke rehabilitation. In light of initial challenges to become familiar with an exoskeleton, patients believed the intensive and repetitive training allowed for notable improvements in walking recovery. Therapists were more guarded in recommending the exoskeleton for future use, noting benefits but also highlighting challenges to implementation within subacute stroke rehabilitation. Future clinical practice should consider the balance between actual and perceived benefits, as well as the potential barriers to integrating an exoskeleton into stroke rehabilitation.      111 Chapter 6: Overall discussion, future directions, and conclusions  As the leading neurological cause of adult disability, stroke is a global health crisis (World Health Organization, 2020). The prevalence of individuals living with the effects of stroke is rising, both nationally and worldwide (Feigin et al., 2016; Kamal et al., 2015; Krueger et al., 2015). Unsurprisingly, healthcare spending related to stroke is amongst the highest in the domain of noncommunicable diseases (Muka et al., 2015). In addition to the economic burden of stroke, the emotional and psychological toll of stroke is another area of concern, with devastating impact on patients and their caregivers (Caro et al., 2018; Lapadatu and Morris, 2019). With advances in acute medical interventions leading to reduced mortality (Hankey, 2017), the overall burden associated with stroke-related disability is expected to continue growing. It is crucial to maximize the efficacy of stroke rehabilitation for promoting recovery and functional independence, in efforts to address the mounting burden of stroke.   Restoring the ability to walk is a crucial target for rehabilitation and research, given its link to functional independence, emotional well-being, and long-term health outcomes (Muren et al., 2008; Pundik et al., 2012). The incidence of walking limitation after stroke has remained constant, despite the improvements in medical treatment and survival rate. Thus, the sheer number of stroke survivors who are unable to walk is increasing. Targeted walking rehabilitation should be intense and repetitious, but actual practice falls short; step counts in therapy are low, especially for those requiring more assistance from their therapists to practice walking (Lang et al., 2009; Rand and Eng, 2012). Optimizing walking rehabilitation in the subacute phase of stroke, the period when the greatest neurological and functional recovery occurs, may help to combat the long-term effects of stroke (Jørgensen et al., 1999; Krakauer et al., 2012). The use of   112 technology, such as newly developed powered robotic exoskeletons, may offer an opportunity to increase the efficacy of gait rehabilitation. 6.1 Summary of study findings This thesis was comprised of four studies centered on the recovery of walking after stroke:  Chapter 2 was an observational study which provided an update on the incidence of lower extremity impairment and walking limitation early after stroke onset. The secondary objective was to characterize the nature of early walking ability as a predictor of home discharge from acute hospitalization. This study provided evidence that the incidence of lower extremity and gait dysfunction after stroke has remained relatively stable in the 21st century, despite emerging medical interventions and recent trends in stroke mortality and incidence. Those who retained some walking function had less leg impairment and were more likely to have had an ischemic stroke. Early walking ability was confirmed to be a significant predictor of being discharged home after acute care, relative to further hospitalization. This finding reinforces the importance of walking as a predictor of meaningful stroke outcome, from as early as several days after the onset of stroke. Chapter 3 was a scoping review which explored the early research of exoskeleton use for patients with stroke in order to determine the breadth and depth of existing research. This review revealed a dearth of RCTs comparing exoskeleton-based training to conventional stroke rehabilitation. It also found that fewer studies had been conducted in the subacute phase of stroke or with non-ambulatory individuals, with the majority of early investigations concerning independent ambulators with chronic stroke. Preliminary findings showed that exoskeletons are safe for use with patients with stroke and at least equivalent to conventional physical therapy.   113 The early evidence also suggested a larger effect of exoskeleton-based training in subacute stroke, relative to chronic stroke.  Chapter 4 was a multi-site RCT to determine the efficacy of exoskeleton-based gait training during subacute stroke rehabilitation, compared to conventional physical therapy care. The primary outcome was walking ability (independence), and secondary outcomes of gait, lower extremity impairment, balance, mobility, cognition, mood, and quality of life were included. No differences were found between the intervention and control group, at discharge from rehabilitation or after 6 months. However, a per-protocol analysis showed significant differences in impairment, cognition, walking speed, and endurance between groups, indicating that exoskeleton-based gait training may be effective at a certain threshold of intensity or with specific patients.  Chapter 5 was a qualitative descriptive study exploring the experience and perception of exoskeleton-based physical therapy from the perspective of both patients and therapists. This study used semi-structured interviews to examine the acceptability of using an exoskeleton for gait retraining. The findings from a thematic analysis revealed strongly positive attitudes from both patients and therapists towards using an exoskeleton during subacute stroke rehabilitation. Challenges around an initial learning curve and barriers to clinical implementation were also noted in this study.  6.2 Integration and contribution of dissertation to current research This dissertation advances the research surrounding walking limitation and rehabilitation after stroke, specifically relating to the use of exoskeleton technology. The specific contributions of each study have been described in detail within their respective Chapters (2 – 5). The   114 following discussion instead focusses on the combined implication of the studies relative to the current and future landscape of stroke rehabilitation and walking recovery. Knowing that the prevalence of leg impairment and walking limitation after first-ever stroke has remained stable, despite the rise of acute medical interventions for acute stroke and downwards trends in incidence and mortality, highlights the ongoing need to develop and refine novel rehabilitation interventions to target lower extremity and walking recovery. Though the study in Chapter 2 did not collect data on leg and walking presentation at time of discharge, the analysis showed that patients unable to walk soon after stroke were more likely to require further inpatient care beyond acute hospitalization. Even with spontaneous recovery, we can assume that many patients admitted to intensive rehabilitation still experience walking limitations at that time. Indeed, several studies investigating walking improvement during inpatient rehabilitation show that a majority of patients that are admitted to rehabilitation are totally or maximally dependent on therapist assistance to walk (Bland et al., 2012; Brown et al., 2015; Ouellette et al., 2015). With population growth and generally stable incidence of stroke and walking impairments, there is simply an ever-growing number of patients requiring post-stroke care (Kamal et al., 2015).  While the focus of acute hospitalization after stroke is, deservingly, the emergency mitigation of pathophysiological processes in the brain and medical stabilization of the patient, a potential consideration for this period is the timing of rehabilitative efforts. Very early mobilization, within 48 hours of stroke onset, has been shown to reduce the odds of having minimal disability at three months after stroke (AVERT Trial Collaboration group et al., 2015). However, in Chapter 2 we showed that patients who are non-ambulatory early after stroke, as well as those who are not discharged home, remain in the acute hospital for approximately 14   115 days. Even if the stroke is medically stabilized, other issues that may prolong acute hospital stay include bed availability (at the receiving facility), in-hospital infection or pressure wounds, and other comorbidities requiring management (e.g., diabetes mellitus, congestive heart failure) (Spratt et al., 2003; Willems et al., 2012). This time window between medical stabilization and hospital discharge may present an opportunity to begin rehabilitative efforts to restore walking impairment; indeed, several small-scale experimental studies during the acute hospitalization period have shown a positive effect of novel gait training (Rose et al., 2018; Yen et al., 2019). It is important, however, to ensure that the nature of the intervention is appropriate given the patient’s medical status and functional ability. The feasibility of the intervention should also be carefully considered, understanding that staffing and scheduling constraints are higher in the acute hospital than inpatient rehabilitation.  This dissertation focused on the use of powered robotic exoskeletons for targeting walking recovery for patients with stroke. The use of robotics can help therapists to achieve higher intensity and longer durations of task-specific walking practice, even for patients who are non-ambulatory after stroke. This new generation of overground exoskeletons, untethered to a treadmill or other support system, was only developed in the past decade. We found, in Chapter 3, that only 11 studies investigating the use of exoskeletons to improve walking after stroke had been published by the end of 2015. These were of varying study designs, ranging from three-subject pre-post proof-of-concept studies to small-scale (n = 50) RCTs. Few studies had specifically targeted non-ambulatory patients in the subacute phase of stroke recovery. Since Chapter 3 was published in 2016, several RCTs and systematic reviews investigating exoskeleton-based stroke rehabilitation have been published (Jayaraman et al., 2019; Moucheboeuf et al., 2020; Nam et al., 2019; Postol et al., 2019; Rojek et al., 2020;   116 Sczesny-Kaiser et al., 2019; Watanabe et al., 2017). Still, relatively few studies have been conducted in the subacute stroke population, compared to the number of studies in chronic stroke (Postol et al., 2019). At first glance, exoskeleton-based gait interventions appear to be more effective than conventional physical therapy in the subacute stage of stroke (Rojek et al., 2020; Watanabe et al., 2017), but largely equivalent at the chronic stage (Jayaraman et al., 2019; Nam et al., 2019; Sczesny-Kaiser et al., 2019). However, the recent reviews have included a heterogeneous pool of studies (in terms of their training protocols, definition of exoskeletons, outcomes), and their subsequent findings and recommendations are conflicting. Some reviews have recommended robotics-assisted gait training in combination with physical therapy (Bruni et al., 2018; Moucheboeuf et al., 2020), while another has indicated that overground robotic exoskeletons should only be used for research purposes at this time (Postol et al., 2019). Similar to the recommendations made in Chapter 3, one review also recommended that future research using exoskeletons should be conducted with dependent ambulators, and should include additional outcomes such as acceptability and psychological function (Postol et al., 2019). Chapters 4 and 5 answer this call, exclusively recruiting subacute patients who were non-ambulatory to participate in exoskeleton-based gait training and debrief interviews. Despite being presented as separate studies, the RCT and qualitative study were conducted concurrently as a mixed methods trial. Mixed methods research combines quantitative and qualitative methodologies in order to investigate a particular question or phenomenon at a deeper level, integrating findings at one or multiple stages of the research process, such that the research findings are greater than the sum of its individual components (Kroll and Morris, 2009; Zhang and Creswell, 2013). Mixed methods research is particularly useful in rehabilitation and physical therapy; it can contribute to the understanding of failed or successful rehabilitation treatments   117 (Kroll and Morris, 2009; Rauscher and Greenfield, 2009). Furthermore, it can advance research and practice by promoting a holistic understanding of patient injury and recovery, as well as by refining innovative, complex treatment interventions (Rauscher and Greenfield, 2009). We utilized a concurrent nested mixed methods design to do just this (Plano Clark and Creswell, 2007; Zhang and Creswell, 2013), comparing a novel exoskeleton-based gait training program while exploring the experience, both of patients and therapists, of integrating robotics within subacute stroke rehabilitation. We designed the mixed method study with a postpositivist worldview, giving priority to the quantitative RCT in order to investigate causal relationships (Clark, 1998; Creswell, 2013). Since the quantitative and qualitative components were presented in this thesis as separate chapters, the findings from each study will be integrated in detail later in this discussion.   The primary finding of the randomized controlled study in Chapter 4 was that an exoskeleton-based gait training program does not increase walking recovery beyond that which is achieved in standard subacute rehabilitation. The exoskeleton intervention also did not show any significant differences at discharge or 6 months in secondary outcomes of impairment, balance, mood, cognition, or quality of life compared to standard physical therapy. While a lack of significance in a RCT is often considered a disappointing outcome or one of less value, we argue otherwise. A strength of the study lies in the unique and flexible intervention protocol, replacing standard physical therapy with an exoskeleton-based gait retraining program. Many research interventions in subacute stroke rehabilitation are investigated as adjunct therapy, offering extra hours of therapy using additional staffing and resources. This is simply not feasible in many rehabilitation settings, where financial and staffing resources are already strained (McHugh and Swain, 2014). Our study protocol involved hospital therapists, already   118 trained (by the manufacturer) to use an exoskeleton, to administer the intervention; additionally, the intervention protocol allowed therapists to use their clinical reasoning for tailoring (including weaning off) the exoskeleton training for their patient participants. Furthermore, the intervention was employed until discharge (or a maximum of 8 weeks), so as not to interfere with regular hospital standards for rehabilitation duration. We believe this protocol is close to how therapists would realistically employ an exoskeleton within their clinical duties and caseload, rather than the typical research protocols offering 30 – 90 minutes of an intervention 5 days a week, for 6 – 12 consecutive weeks. We interpret the finding of no difference from standard physical therapy to indicate that exoskeleton-based gait training can be substituted for standard physical therapy sessions at no detriment to patient outcomes, to support therapists in choosing to mobilizing their severely dependent patients using robotics. However, the small sample size and loss to follow-up at the 6-month mark warrant caution; the findings from this trial and interpretation should not be considered definitive until further, larger-scale research trials are conducted and show similar results.  A drawback of the flexible intervention protocol was the difficulty in achieving standardized dosing across all study participants. The exoskeleton was not the ‘active ingredient’ of the intervention (Craig et al., 2008); rather, the active ingredient was the higher intensity (i.e., duration, repetition) of walking practice, afforded by means of using an exoskeleton with patients requiring more assistance to walk. On average, participants in the Exoskeleton group took a mean of 592 steps per therapy session, nearly double the mean of 330 steps taken by participants in the Usual Care group. However, compared to studies of walking interventions showing greater improvements during rehabilitation over and above conventional therapy (Hornby et al., 2016; Klassen et al., 2020), our intervention group did not achieve nearly the   119 same quantity of walking practice (2358 – 4747 steps). Indeed, the four participants in our study (3 Exoskeleton participants, 1 Usual Care participant) who did become fully independent in ambulation (FAC = 5) by discharge took a higher average (923 – 1230) number of steps in their physical therapy sessions. We contend that the lack of significant findings is partially explained by the low number of steps taken by participants in the Exoskeleton group. Recent studies which ensured high-intensity and repetitious walking practice for non-ambulatory patients with stroke have demonstrated positive findings for walking improvements (Henderson et al., 2020; Hornby et al., 2015). Though using an exoskeleton does allow therapists to achieve higher repetitions of stepping with dependent patients, it can still be quite taxing to operate the device and stabilize patients with minimal strength or balance function. For severely dependent patients, achieving step counts in the order of several thousand using an exoskeleton may simply not be attainable after taking into account patient or therapist fatigue, the need for rest breaks, and limited therapy time. Additionally, while therapists were given the option to reduce or discontinue the use of the exoskeleton once participants became ambulatory without assistance, we did not rigorously enforce a minimum intensity of walking practice in the remaining therapy time. This may be another way in which the intervention protocol simply did not achieve a high enough intensity to yield a treatment effect.   The additional analysis in Chapter 4 hints that there are rewards from exoskeleton-based gait rehabilitation to be reaped. A number of participants in the Exoskeleton group were not compliant to the training regimen, refusing further exoskeleton sessions that were intended. When these participants were removed from the Exoskeleton group, the mean step count for remaining participants increased to 658 per session. A per-protocol analysis showed greater improvements in lower extremity impairment and cognition at discharge for participants who   120 adhered to the Exoskeleton intervention, specifically by 3.9 and 2.1 points on the FMA-LE and MoCA, respectively. Of these values, only the improvement in MoCA score fell within the MCID range of 1.22 – 2.15 (Wu et al., 2019). However, neither improvement above the Usual Care group was maintained at the 6-month mark. Conversely, the per-protocol analysis revealed significant differences between groups for walking speed and endurance (6MWT) only at the 6-month evaluation. Furthermore, the respective 0.32 m/s and 108 m difference between groups surpass the MCID for gait speed (0.16 m/s (Tilson et al., 2010)) and 6MWT (61m (Perera et al., 2006)). This long-term finding may suggest that a latent benefit of using an exoskeleton may carry over into the home setting, which allows patients to continue improving in their gait function. Unmeasured outcomes, such as gait quality, symmetry, or muscle activation, may provide more insight as to the nature of these delayed improvements (Rozanski et al., 2020). An alternative possibility is simply that the experience of repetitious walking practice during rehabilitation bestowed participants with the confidence to continue practicing walking on their own after discharge, leading to continued improvements. Though participants in both groups achieved similar levels of walking recovery, another per-protocol finding was that participants who became ambulatory without requiring assistance (FAC ≥ 3) during the intervention period did so 8.5 days sooner in the Exoskeleton group than the Usual Care group. The additional analyses in Chapter 4 point to two conjectures that require further investigation: 1) that exoskeleton-based training leads to greater functional gains for non-ambulatory patients, if the amount of walking practice is of sufficient intensity and repetition; and 2) that patients who will become independent ambulators do so faster as a result of exoskeleton training.     The qualitative inquiry described in Chapter 5 explored the experience and perception of an exoskeleton-based gait training program during subacute stroke rehabilitation from the   121 perspective of both patients and therapists. Though previous studies have surveyed therapists’ views towards using an exoskeleton, they were not specific to stroke or to a regimented intervention program and instead asked therapists broadly about the general use of exoskeleton technology (Mortenson et al., 2020; Read et al., 2020; Vaughan-Graham et al., 2020). As such, this is the first qualitative study to have examined the acceptability of an exoskeleton device as a targeted intervention within subacute stroke rehabilitation.  We found that patients with stroke were highly accepting of the exoskeleton-based intervention. After overcoming an initial hurdle of learning to use the device and becoming accustomed to the automated movements, patients in large majority felt that using an exoskeleton allowed for higher intensity gait training and thus greater walking improvement than they would have achieved in standard physical therapy. Therapists, who also experienced an initial learning curve with learning to operate an exoskeleton, similarly attached value to exoskeleton-based gait training but were careful to state that the greater benefit is dependent on specifying appropriate patients for the device. It would seem that participants in our study were more enthusiastic than those in other qualitative studies so far. In a recent study conducted by Vaughan-Graham et al. (Vaughan-Graham et al., 2020), patients with stroke were interviewed after a single exoskeleton training session; their regular therapists were interviewed as well, after observing the single training session. It is not surprising that participants in their study had more reservations towards using an exoskeleton, given our finding that it takes several sessions to become familiar with the device (and to experience any benefits first-hand). Our findings that there are several perceived barriers to implementing regular exoskeleton-based training within stroke rehabilitation is consistent amongst all qualitative studies (Mortenson et al., 2020; Read et al., 2020; Vaughan-Graham et al., 2020), with set-up time, staffing, and scheduling issues most commonly   122 identified. Administrative and organizational support are therefore critical to adopting and integrating exoskeletons within stroke rehabilitation, even if the actual technology-users are in favour of its use. Efforts to engage these stakeholders, whether through knowledge translation, involvement in future qualitative research, or discussions of policy change, should be considered for any future attempts at implementation.     In revisiting the overarching mixed methods design of Chapter 4 and 5, integrating the findings from the two studies allows for a more comprehensive evaluation of the role for exoskeleton-based gait training during subacute stroke rehabilitation. On one hand, the main results from the RCT of no greater benefit or any functional outcomes suggest that an exoskeleton is not necessary for stroke rehabilitation; taking into account other factors such as cost, perceived barriers, and expected effort, the logical recommendation would be to forgo acquiring or using such devices. On the other hand, the overwhelmingly positive reaction from patients, endorsement by therapists, as well as anecdotal success stories of achieving higher-intensity walking practice and unparalleled walking recovery, lend strong support to employing an exoskeleton in stroke rehabilitation. By conducting the studies concurrently and integrating the results, whether complementary or contradictory, we can reach more comprehensive conclusions and make more informed recommendations for future clinical practice and research (O’Cathain et al., 2010).   The most straightforward interpretation of the findings of equivalent efficacy and high acceptability is the recommendation that exoskeletons should be used in stroke rehabilitation mainly for the purpose of patient satisfaction and motivation, by providing the subjective experience of receiving what is perceived as better care. However, given that the exoskeleton is no better than standard physical therapy for functional recovery, as well as the reported effort   123 and time cost associated with scheduling adjustments and setting up the device, we recommend clinicians to forgo using an exoskeleton for ambivalent or disinterested patients. Furthermore, we also emphasize that clinicians who choose to utilize the device should commit to a minimum number of sessions, given the positive quantitative findings with compliant participants and qualitative experience of an initial learning curve. There were also some corroborative study results. The per-protocol analysis suggested a potential therapeutic effect of exoskeleton training above standard rehabilitation that was masked by methodological circumstances; qualitatively, therapists similarly voiced that certain patients with stroke excel and achieve tremendous gains from using an exoskeleton more than others. We thus advocate that there is a place for exoskeletons within subacute stroke rehabilitation. Considerations for future research, then, are the identification of the most appropriate patients for exoskeleton-based rehabilitation and determining the most effective training protocol for walking recovery.   Further caveats to these recommendations surround the practical feasibility of implementing an exoskeleton in the clinical setting. Many of the barriers identified in the qualitative study (i.e., scheduling, set-up time, staffing requirements) are outside of individual therapists’ control; policy and funding changes may be required to fully embrace integrating exoskeleton technology in the rehabilitation setting. For example, policy changes to redistribute therapy time according to stroke severity and patient presentation (i.e., longer sessions with lower functioning patients, shorter sessions with higher functioning patients) may allow individual therapists to address some of the time constraints of using an exoskeleton in their regular therapy session. Additional staffing, such as therapy assistants, solely dedicated to exoskeleton-based therapy may help to streamline sessions by taking care of patient preparation and device set-up, or even offering adjunct exoskeleton-based gait training. However, these   124 changes may be difficult to justify without substantive evidence of effectiveness above current standard of care. Furthermore, when taking into account the financial cost of exoskeleton-based therapy, which includes device purchase, device-specific training fees, and ongoing wages for additional staffing, managers may be hard-pressed to commit funding in this direction. Nevertheless, addressing all stakeholders’ perception of the technology is paramount for adoption (Straub, 2009). A counter argument in this cost-benefit analysis in favour of adopting exoskeleton technology into clinical practice may be the potential cost savings yielded by the therapeutic benefits. If we were to accept the per-protocol findings as true, that exoskeleton-based training results in earlier walking independence and greater long-term walking recovery, this might translate to earlier home discharge and reduced stroke-related disability at the population level, which in turn may reduce the economic burden of stroke. The future of powered exoskeletons in stroke rehabilitation will depend heavily on the dynamics between ongoing research efforts, stakeholder perceptions of the technology, and knowledge translation strategies. Such strategies may involve publication of the present research in rehabilitation journals, and educational presentations at site that are considering or have already purchased an exoskeleton device. Overall, the findings from this dissertation reinforce the understanding and ongoing efforts to address severe walking impairment after stroke, specifically with respect to employing a powered exoskeleton in the subacute rehabilitation setting.  6.3 Future research directions and recommendations  Future experimental research seeking to replicate our RCT (Chapter 4) should consider several conjectures derived from our per-protocol findings. Our per-protocol analysis showed improvements for several measures (i.e., walking speed, endurance, leg impairment, cognition,   125 time to reach unassisted walking) for participants who adhered to the intervention. We contend that a fully compliant Exoskeleton group would have shown more improvements than our intention-to-treat findings, given that there would have been an overall higher dose of walking practice achieved during the intervention period. We suggest adapting the recruitment strategy, including participant inclusion and exclusion criteria, to ensure the best chance of full compliance. This may entail including a trial session (or several) with the robotic exoskeleton to allow prospective participants to experience the sensation of exoskeletal-gait training and to opt out of the study before being randomized. Expanding on the concept of dose with exoskeleton-based gait training, researchers should consider high, yet feasible, targets for walking practice. Given our findings and other research interventions, we suggest a minimum of 1000 steps per exoskeleton-based gait training session for participants who are heavily dependent ambulators. In a similar vein is the need for future research to identify the most appropriate patients with stroke who will truly benefit from exoskeleton-based gait training. We believe that higher-intensity walking practice is beneficial for all non-ambulatory patients, though this was not shown in our intention-to-treat analysis. However, qualitative anecdotes and the per-protocol analysis may suggest that some participants simply excel with exoskeleton-based training compared to others. It may be too hopeful, otherwise, to assume that all patients have the same potential to experience high levels of recovery if a sufficient number of steps are taken. It is important to consider ‘responders’ and ‘non-responders’ when it comes to researching therapeutic intervention and recovery, though the classification is much harder to achieve in practice. Studies have considered several biomarkers, such as brain lesion and paretic muscle strength, to predict responsiveness to gait intervention after stroke (Awad et al., 2016; Kaczmarczyk et al., 2012). Our per-protocol findings similarly suggest a relationship between   126 exoskeleton-training and cognition (MoCA) and motor impairment (FMA-LE), potentially indicating these measures as screening tools for recruitment. Another approach to identifying responders for research was described in a recent study protocol, which entailed observing patients in rehabilitation for a week to detect stagnant recovery prior to enrolment (Tsurushima et al., 2019). This sampling method, though diligent, may pose some difficulties for study enrolment and may further reduce the generalizability of the research.  Continuing to employ mixed methods in future research may help in better understanding and exploring various aspects of exoskeleton-based gait training in stroke rehabilitation. The nature of mixing (timing, strategy) is an important consideration in the design and analysis of mixed methods research, as it can dictate which data is collected and what conclusions are drawn (Creswell and Plano Clark, 2007; Plano Clark and Creswell, 2007). This dissertation integrated the quantitative (efficacy) and qualitative (acceptability) findings of the ExStRA trial after analyzing the data separately in order to comment on the adoptability of exoskeleton training during subacute stroke rehabilitation. Though purposely designed this way, this level of mixing only begins to capitalize on the strengths of the mixed methods research design. Future mixed methods research in this area could begin integrating the quantitative and qualitative aspects at an earlier stage, such as in sequential designs (Rauscher and Greenfield, 2009), to investigate a single phenomenon. For example, research seeking to identify ‘responders’ and ‘non-responders’ to exoskeleton training after stroke could first qualitatively survey therapists, neuroscientists, and patients on potential markers for responsiveness. Next, a subsequent quantitative clinical study employing a short exoskeleton training program could be used to verify or discount these markers as predictors of positive training outcomes.    127 Another consideration and area for future research is the use of an exoskeleton during the acute hospitalization period, for those patients who are unable to walk and awaiting transfer to an inpatient rehabilitation facility. If they are medically stable and do not demonstrate any contraindications, using an exoskeleton early after stroke may have numerous positive effects for patients. This can range from blood pressure management and early weight-bearing, to shaping psychological outlook towards recovery and rehabilitation. Only one pilot study has been published so far (Yokota et al., 2019), which enrolled participants within six days of stroke onset and showed no adverse to early exoskeleton-based gait training. They concluded that using an exoskeleton can be a safe method to target patients with severe walking disability during acute stroke hospitalization, and that randomized trials in this area are warranted to clarify efficacy.  An important outcome in this research would be discharge destination, given the predictive nature of walking ability early after stroke. It would be clinically meaningful if early, intensive walking practice by use of an exoskeleton could change the recovery trajectory for patients who would otherwise be sedentary until the subacute rehabilitation period. However, it should be cautioned that such an intervention may not be clinically feasible during acute hospitalization, between the short length of stay, set-up times, and staffing requirements. Definitive and meaningful efficacy findings would be required to justify the resources required to implement exoskeleton use in the acute setting. The acute hospitalization period could instead serve as a time to prepare for future robotic training in subacute rehabilitation, whether by completing required screening measurements or prioritizing early standing practice for blood pressure management.  Exoskeleton technology and electromechanical devices are ever evolving, with new versions of previous models steadily being released alongside completely different and   128 innovative devices for gait retraining. The EksoNR (Ekso Bionics, Richmond, California, USA) exoskeleton is a newer model of the EksoGT (that was used for our intervention), which was cleared by the Food and Drug Administration for use in acquired brain injury earlier this year (Owusu, 2020). The range of electromechanical devices from bilateral, multi-joint exoskeletons to unilateral, single-joint motorized bracing continues to expand, with new ‘exo-suit’ devices that can assist patient movement without using as much exterior machinery. These lower-profile, cable-driven systems offer less balance support and are intended for more independent ambulators; perhaps they may serve the role as a progression of assisted gait-training once patients surpass the need for the more supportive powered exoskeleton. Early research has shown improved walking outcomes in ambulatory patients with chronic stroke after using such devices for gait retraining (Awad et al., 2020). As long as such devices are being developed, both for scientific and commercial purposes, future research is tremendously important. Qualitative research with various stakeholders (patients, therapists, managers) has many useful purposes, for example: 1) to inform the ongoing development of current and future devices; 2) to appreciate the clinical value of using such technology at the frontlines; and 3) to understand how electromechanical devices can be best integrated clinically. Through engaging with downstream users, developers may uncover clinically meaningful design improvements, such as the need to incorporate backwards and side-stepping functions for gait rehabilitation (Gagnon and Aissaoui, 2020; Rose et al., 2018). Experimental research is also required to establish efficacy and effectiveness of using new electromechanical devices, specifically the appropriate target populations and training regimens. Both types of research will ultimately contribute to the adoption of such technology for the treatment of walking impairment in stroke rehabilitation.    129 We primarily considered the use of powered exoskeletons to retrain gait after stroke, but there may be several other reasons to use an exoskeleton for stroke rehabilitation. Owing to the functionality of exoskeleton devices, with adjustable assistance or resistance to movement and various modes of stepping initiation, it is not a reach to suggest exoskeleton training can be used to target impairment-based outcomes such as motor recovery, muscle activation, and muscle strength. This is supported by our per-protocol finding of improved lower extremity motor recovery in the Exoskeleton group. Similarly, given the learning curve described qualitatively and the per-protocol difference in our cognitive outcome (MoCA), we suggest that exoskeleton-based walking can fit within cognitive dual-task walking paradigms, which can improve both walking and balance function after stroke (He et al., 2018). The physical support offered by the exoskeleton is also an opportunity for sensory and proprioceptive training, while in a standing position, which may be of particular benefit to patients with unilateral spatial neglect. To our knowledge, this has not yet been investigated. The auditory cueing associated with repetitive exoskeletal walking may also serve to train rhythm perception, an impairment after stroke that is associated with gait asymmetry (Patterson et al., 2018). Exoskeletons also offer an opportunity to train exercise capacity and aerobic training for patients who are heavily dependent after stroke; though it has been shown to be less effortful compared to overground walking in healthy adults and ambulatory patients (Lefeber et al., 2020, 2017), it may be one of the few ways to achieve any cardiorespiratory endurance training for non-ambulatory patients. Though the amassing evidence surrounding exoskeletons and walking recovery is still mixed, there are numerous other avenues left to explore.   130 6.4 Personal reflection statement In keeping with the qualitative nature of Chapter 5, I have included a personal reflection regarding my overall experience conducting this dissertation research and pursuing a thesis. Not only is reflexivity recommended to improve credibility in qualitative research, as performed for Chapter 5, it is an important appraisal tool for researchers to enhance critical self-awareness (Mortari, 2015; Probst, 2015). This reflection was guided by a rubric proposed by Stynes et al. for self-appraisal after fully completing project work, which involves questioning the topic, research experience, and researcher (Stynes et al., 2018).  I decided to pursue this thesis after working several years in a neurological physical therapy clinic, merging an interest in rehabilitation technology and a desire to improve outcomes for people with stroke on a greater scale. In the clinic, I commonly saw individuals with longstanding chronic stroke whose only exercise in the week consisted of the 60 minutes of therapy provided in the clinic. Despite laborious efforts during body weight-supported treadmill training, both by myself and the patient, there were rarely lasting improvements and we instead framed our therapy goals as maintenance. I was convinced that the stagnant therapy outcomes were a function of chronicity, so I was particularly enthusiastic to focus my thesis in the early and subacute recovery phase of stroke. I began this research with the assumption that change occurs quickly in the early recovery phase after stroke. I admittedly also assumed that the magnitude of recovery after stroke is extremely malleable, simply dependent on the timing and intervention used. It is with some bemusement that I reflect on these assumptions at the tail end of this thesis. While the first assumption was confirmed over the course of conducting this research, that patients with subacute stroke experience quicker rates of improvement than those in the chronic   131 setting, it is the second assumption that has generated the most internal strife. The notion that simply increasing the duration, repetition, and intensity of walking practice can lead to improved outcomes was fantastically alluring to my clinician mindset; furthermore, it is supported by a multitude of evidence and practice recommendations (Hebert et al., 2016; Hornby et al., 2020). However, in reflecting on this, I will admit that I wishfully assumed a limitless extent of recovery, despite existing dialogue on proportional recovery and inherent plateaus in rehabilitation (Prabhakaran et al., 2008; Smith et al., 2017; Stinear et al., 2017). It is this personal assumption that has been most challenging to overcome, for I may have conflated a hypothesis with hopefulness.  I was confident and expectant of a clear, definitive advantage of using an exoskeleton for walking recovery over conventional therapy during subacute stroke rehabilitation. However, the primary analysis in Chapter 4 showed otherwise. The dissonance between my expectations and reality were initially felt as failure, pessimism, and apathy. However, after a period of reflective ruminating, I have come to appreciate that non-significant findings are still very important findings to move a research area forward — a lesson I would not have learned otherwise. Additionally, the experience of designing and conducting a multi-site clinical trial has been a source of learning for which I am extremely grateful. Not only was it an opportunity to develop as a clinical researcher, the non-significant trial outcome has given me a critical lens to reflect on design characteristics and management processes to optimize future research endeavors.  After having conducted both quantitative and qualitative research firsthand, I have a greater appreciation of the critical role the researcher plays in shaping the research process and outcomes. Having acted as a co-investigator, coordinator, and intervention therapist to conduct the clinical trial, I now have more awareness of the various ways in which personal biases can   132 infiltrate research and better strategies to minimize such threats. I also have a greater understanding of the notion that the researcher is inseparable from the data and the necessity of reflection for embracing this aspect of qualitative research. After conducting the interviews for Chapter 5, I can appreciate how my personal assumptions (e.g., of exoskeleton intervention efficacy), language, and interviewing style can shape the data collected; I can also see how such assumptions can dictate the analysis. Through pursuing this thesis, I have learned as much about myself as I have the research method.  6.5 Conclusion This dissertation encompasses a series of studies surrounding walking impairment and rehabilitation after stroke, revolving around novel robotic exoskeleton technology. This dissertation highlighted an ongoing necessity to improve interventions for walking recovery after stroke. Powered robotic exoskeletons offer an opportunity to practice walking with even the most dependent non-ambulatory patients after stroke. Though a definitive assessment of efficacy was not established, secondary analyses point to a positive training effect for certain patients with stroke. Additionally, this research revealed a strongly accepting response to the technology, from both patients and therapists. This dissertation provides current evidence for clinical decision-making and a foundation for future research to further investigate the benefit of this technology in early stroke rehabilitation for walking recovery.    133 References Ada L, Dean CM, Vargas J, Ennis S. Mechanically assisted walking with body weight support results in more independent walking than assisted overground walking in non-ambulatory patients early after stroke: a systematic review. J Physiother. 2010;56:153–61. Andrenelli E, Ippoliti E, Coccia M, Millevolte M, Cicconi B, Latini L, et al. Features and predictors of activity limitations and participation restriction 2 years after intensive rehabilitation following first-ever stroke. Eur J Phys Rehabil Med. 2015;51:575–85. Arazpour M, Samadian M, Bahramizadeh M, Joghtaei M, Maleki M, Ahmadi Bani M, et al. The efficiency of orthotic interventions on energy consumption in paraplegic patients: a literature review. Spinal Cord. 2015;53:168–75. Ardestani MM, Kinnaird CR, Henderson CE, Hornby TG. Compensation or recovery? Altered kinetics and neuromuscular synergies following high-intensity stepping training poststroke. Neurorehabil Neural Repair. 2019;33:47–58. Arene N, Hidler J. Understanding motor impairment in the paretic lower limb after a stroke: a review of the literature. Top Stroke Rehabil. 2009;16:346–56. Arksey H, O’Malley L. Scoping studies: towards a methodological framework. Int J Soc Res Methodol Theory Pract. 2005;8:19–32. Asselin P, Knezevic S, Kornfeld S, Cirnigliaro C, Agranova-Breyter I, Bauman WA, et al. Heart rate and oxygen demand of powered exoskeleton-assisted walking in persons with paraplegia. J Rehabil Res Dev. 2015;52:147–58. AVERT Trial Collaboration group, Bernhardt J, Langhorne P, Lindley RI, Thrift AG, Ellery F, et al. Efficacy and safety of very early mobilisation within 24 h of stroke onset (AVERT): a randomised controlled trial. Lancet. 2015;386:46–55.  134 Awad LN, Esquenazi A, Francisco GE, Nolan KJ, Jayaraman A. The ReWalk ReStoreTM soft robotic exosuit: a multi-site clinical trial of the safety, reliability, and feasibility of exosuit-augmented post-stroke gait rehabilitation. J Neuroeng Rehabil. 2020;17:80. Awad LN, Reisman DS, Pohlig RT, Binder-Macleod SA. Identifying candidates for targeted gait rehabilitation after stroke: better prediction through biomechanics-informed characterization. J Neuroeng Rehabil. 2016;13:84. Balami JS, Chen RL, Buchan AM. Stroke syndromes and clinical management. Q J Med. 2013;106:607–15. Bashir A. Stroke and telerehabilitation- a brief communication (Preprint). JMIR Rehabil Assist Technol. 2020;7:e18919. Béjot Y, Troisgros O, Gremeaux V, Lucas B, Jacquin A, Khoumri C, et al. Poststroke disposition and associated factors in a population-based study: the Dijon stroke registry. Stroke. 2012;43:2071–7. Belda-Lois J-M, Mena-del Horno S, Bermejo-Bosch I, Moreno JC, Pons JL, Farina D, et al. Rehabilitation of gait after stroke: a review towards a top-down approach. J Neuroeng Rehabil. 2011;8:66. Benjamin EJ, Blaha MJ, Chiuve SE, Cushman M, Das SR, Deo R, et al. Heart disease and stroke statistics - 2017 update: a report from the American Heart Association. Circulation. 2017;135:e146–603. Berg K, Wood-Dauphinee S, Williams JI, Gayton D. Measuring balance in the elderly: preliminary development of an instrument. Physiother Canada. 1989;41:304–11. Bernheisel CR, Schlaudecker JD, Leopold K. Subacute management of ischemic stroke. Am Fam Physician. 2011;84:1383–8.  135 Bewick V, Cheek L, Ball J. Statistics review 14: logistic regression. Crit Care. 2005;9:112–8. Bijleveld-Uitman M, Van De Port I, Kwakkel G. Is gait speed or walking distance a better predictor for community walking after stroke? J Rehabil Med. 2013;45:535–40. Billinger SA, Arena R, Bernhardt J, Eng JJ, Franklin BA, Johnson CM, et al. Physical activity and exercise recommendations for stroke survivors: a statement for healthcare professionals from the American Heart Association/American Stroke Association. Stroke. 2014;45:2532–53. Bland MD, Sturmoski A, Whitson M, Connor LT, Fucetola R, Huskey T, et al. Prediction of discharge walking ability from initial assessment in a stroke inpatient rehabilitation facility population. Arch Phys Med Rehabil. 2012;93:1441–7. Blum L, Korner-Bitensky N. Usefulness of the Berg Balance Scale in stroke rehabilitation: a systematic review. Phys Ther. 2008;88:559–66. Boehlen W, Sample M. Rehabilitation culture and its impact on technology: unpacking practical conditions for ultrabilitation. Disabil Rehabil. 2020;42:1490–4. Bogey R, Hornby TG. Gait training strategies utilized in poststroke rehabilitation: are we really making a difference? Top Stroke Rehabil. 2007;14:1–8. Bohannon RW, Andrews AW, Smith MB. Rehabilitation goals of patients with hemiplegia. Int J Rehabil Res. 1988;11:181–4. Bonnechère B, Jansen B, Omelina L, Van Sint Jan S. The use of commercial video games in rehabilitation: a systematic review. Int J Rehabil Res. 2016:277–90. Bortole M, Venkatakrishnan A, Zhu F, Moreno JC, Francisco GE, Pons JL, et al. The H2 robotic exoskeleton for gait rehabilitation after stroke: early findings from a clinical study. J Neuroeng Rehabil. 2015;12:54.  136 Bowden MG, Woodbury ML, Duncan PW. Promoting neuroplasticity and recovery after stroke: future directions for rehabilitation clinical trials. Curr Opin Neurol. 2013;26:37–42. Bradshaw C, Atkinson S, Doody O. Employing a qualitative description approach in health care research. Glob Qual Nurs Res. 2017;4:2333393617742282. Braun V, Clarke V. To saturate or not to saturate? Questioning data saturation as a useful concept for thematic analysis and sample-size rationales. Qual Res Sport Exerc Heal. 2019. Braun V, Clarke V. Using thematic analysis in psychology. Qual Res Psychol. 2006;3:77–101. Brott T, Adams HP, Olinger CP, Marle JR, Barsan WG, Biller J, et al. Measurements of acute cerebral infarction: a clinical examination scale. Stroke. 1989;20:864–70. Brown AW, Therneau TM, Schultz BA, Niewczyk PM, Granger C V. Measure of functional independence dominates discharge outcome prediction after inpatient rehabilitation for stroke. Stroke. 2015;46:1038–44. Bruni MF, Melegari C, De Cola MC, Bramanti A, Bramanti P, Calabrò RS. What does best evidence tell us about robotic gait rehabilitation in stroke patients: a systematic review and meta-analysis. J Clin Neurosci. 2018;48:11–7. Buesing C, Fisch G, O’Donnell M, Shahidi I, Thomas L, Mummidisetty CK, et al. Effects of a wearable exoskeleton stride management assist system (SMA®) on spatiotemporal gait characteristics in individuals after stroke: a randomized controlled trial. J Neuroeng Rehabil. 2015;12:69. Byl NN. Mobility training using a bionic knee orthosis in patients in a post-stroke chronic state: a case series. J Med Case Rep. 2012;6:216. Caceres JA, Goldstein JN. Intracranial hemorrhage. Emerg Med Clin North Am. 2012;30:771– 137 94. Caetano LCG, Pacheco BD, Samora GAR, Teixeira-Salmela LF, Scianni AA. Self-efficacy to engage in physical exercise and walking ability best predicted exercise adherence after stroke. Stroke Res Treat. 2020;2020:2957623. Cahill LS, Carey LM, Mak-Yuen Y, McCluskey A, Neilson C, O’Connor DA, et al. Factors influencing allied health professionals’ implementation of upper limb sensory rehabilitation for stroke survivors: a qualitative study to inform knowledge translation. BMJ Open. 2021;11:1–11. Calabrò RS, Naro A, Russo M, Bramanti P, Carioti L, Balletta T, et al. Shaping neuroplasticity by using powered exoskeletons in patients with stroke: a randomized clinical trial. J Neuroeng Rehabil. 2018;15:35. Camden C, Shikako-Thomas K, Nguyen T, Graham E, Thomas A, Sprung J, et al. Engaging stakeholders in rehabilitation research: a scoping review of strategies used in partnerships and evaluation of impacts. Disabil Rehabil. 2015;37:1390–400. Campbell BCV, Khatri P. Stroke. Lancet. 2020;396:129–42. Campbell BCV, De Silva DA, Macleod MR, Coutts SB, Schwamm LH, Davis SM, et al. Ischaemic stroke. Nat Rev Dis Prim. 2019;5:70. Canadian Institute for Health Information (CIHI). Factors predicting discharge home from inpatient rehabilitation after stroke [Internet]. 2009 [cited 2020 Dec 14]. Available from: Canadian Stroke Network. The quality of stroke care in Canada [Internet]. 2011 [cited 2020 Dec 14]. Available from:  138 Caro CC, Costa JD, Da Cruz DMC. Burden and quality of life of family caregivers of stroke patients. Occup Ther Heal Care. 2018;32:154–71. Cauraugh JH, Kim SB. Stroke motor recovery: active neuromuscular stimulation and repetitive practice schedules. J Neurol Neurosurg Psychiatry. 2003;74:1562–6. Chan L, Sandel ME, Jette AM, Appelman J, Brandt DE, Cheng P, et al. Does postacute care site matter? A longitudinal study assessing functional recovery after a stroke. Arch Phys Med Rehabil. 2013;94:622–9. Chang SR, Kobetic R, Triolo RJ. Effect of exoskeletal joint constraint and passive resistance on metabolic energy expenditure: implications for walking in paraplegia. PLoS One. 2017;12:e0183125. Charalambous C, Bowden MG, Adkins DL. Motor cortex and motor cortical interhemispheric communication in walking after stroke - the roles of TMS and animal models in our current and future understanding. Neurorehabil Neural Repair. 2016;30:94–102. Charalambous CC, Bonilha HS, Kautz SA, Gregory CM, Bowden MG. Rehabilitating walking speed poststroke with treadmill-based interventions: a systematic review of randomized controlled trials. Neurorehabil Neural Repair. 2013;27:709–21. Chen CC, Bode RK. Factors influencing therapists’ decision-making in the acceptance of new technology devices in stroke rehabilitation. Am J Phys Med Rehabil. 2011;90:415–25. Chen G, Chan CK, Guo Z, Yu H. A review of lower extremity assistive robotic exoskeletons in rehabilitation therapy. Crit Rev Biomed Eng. 2013;41:343–63. Cho JS, Hu Z, Fell N, Heath GW, Qayyum R, Sartipi M. Hospital discharge disposition of stroke patients in Tennessee. South Med J. 2017;110:594–600. Clark AM. The qualitative-quantitative debate: moving from positivism and confrontation to  139 post-positivism and reconciliation. J Adv Nurs. 1998;27:1242–9. Clark WE, Sivan M, O’Connor RJ. Evaluating the use of robotic and virtual reality rehabilitation technologies to improve function in stroke survivors: a narrative review. J Rehabil Assist Technol Eng. 2019;6:2055668319863557. Clarke DJ, Burton LJ, Tyson SF, Rodgers H, Drummond A, Palmer R, et al. Why do stroke survivors not receive recommended amounts of active therapy? Findings from the ReAcT study, a mixed-methods case-study evaluation in eight stroke units. Clin Rehabil. 2018;32:1119–32. Cohen J. Statistical power analysis for the behavioural science (2nd ed.). Hillsdale, NJ: Lawrence Erlbaum Associates; 1988. Craig P, Dieppe P, Macintyre S, Michie S, Nazareth I, Petticrew M, et al. Developing and evaluating complex interventions : new guidance. BMJ. 2008;337:a1655. Creswell JW. Qualitative Inquiry and Research Design: Choosing Among Five Approaches. Third Edit. Los Angeles, CA: SAGE Publications; 2013. Creswell JW, Plano Clark VL. Choosing a Mixed Methods Design. Des. Conduct. Mix. Methods Res., Thousand Oaks: Sage Publications; 2007, p. 58–88. Crofts JJ, Higham DJ, Bosnell R, Jbabdi S, Matthews PM, Behrens TEJ, et al. Network analysis detects changes in the contralesional hemisphere following stroke. Neuroimage. 2011;54:161–9. Cromie JE. Work-related musculoskeletal disorders in physical therapists: prevalence, severity, risks, and responses. Phys Ther. 2000;80:336–51. Cumming TB, Packer M, Kramer SF, English C. The prevalence of fatigue after stroke: a systematic review and meta-analysis. Int J Stroke. 2016;11:968–77.  140 Davis FD. Perceived usefulness, perceived ease of use, and user acceptance of information technology. MIS Q. 1989;13:319–40. Dewilde S, Annemans L, Peeters A, Hemelsoet D, Vandermeeren Y, Desfontaines P, et al. The relationship between home-time, quality of life and costs after ischemic stroke: the impact of the need for mobility aids, home and car modifications on home-time. Disabil Rehabil. 2020;42:419–25. Dmytriw AA, Song JSA, Yu E, Poon CS. Cerebral venous thrombosis: state of the art diagnosis and management. Neuroradiology. 2018;60:669–85. Dobkin BH, Duncan PW. Should body weight-supported treadmill training and robotic-assistive steppers for locomotor training trot back to the starting gate? Neurorehabil Neural Repair. 2012;26:308–17. Donnan GA, Fisher M, Macleod M, Davis SM. Stroke. Lancet. 2008;371:1612–23. van Duijnhoven HJR, Heeren A, Peters MAM, Veerbeek JM, Kwakkel G, Geurts ACH, et al. Effects of exercise therapy on balance capacity in chronic stroke: systematic review and meta-analysis. Stroke. 2016;47:2603–10. Duncan PW, Sullivan KJ, Behrman AL, Azen SP, Wu SS, Nadeau SE, et al. Body-weight–supported treadmill rehabilitation after stroke. N Engl J Med. 2011;364:2026–36. Dutrieux RD, van Eijk M, van Mierlo ML, van Heugten CM, Visser-Meily JMA, Achterberg WP. Discharge home after acute stroke: differences between older and younger patients. J Rehabil Med. 2016;48:14–8. Easton JD, Saver JL, Albers GW, Alberts MJ, Chaturvedi S, Feldmann E, et al. Definition and evaluation of transient ischemic attack: a scientific statement for healthcare professionals from the American Heart Association/American Stroke Association stroke council;  141 council on cardiovascular surgery and anesthesia; council on cardio. Stroke. 2009;40:2276–93. Eich H-J, Mach H, Werner C, Hesse S. Aerobic treadmill plus Bobath walking training improves walking in subacute stroke: a randomized controlled trial. Clin Rehabil. 2004;18:640–51. von Elm E, Altman DG, Egger M, Pocock SJ, Gøtzsche PC, Vandenbroucke JP. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. Lancet. 2007;370:1453–7. Engelhardt E. Apoplexy, cerebrovascular disease, and stroke: historical evolution of terms and definitions. Dement Neuropsychol. 2017;11:449–53. English C, Hillier S. Circuit class therapy for improving mobility after stroke: A systematic review. J Rehabil Med. 2011;43:565–71. Esquenazi A, Talaty M, Packel A, Saulino M. The ReWalk powered exoskeleton to restore ambulatory function to individuals with thoracic-level motor-complete spinal cord injury. Am J Phys Med Rehabil. 2012;91:911–21. Fager SK, Burnfield JM. Patients’ experiences with technology during inpatient rehabilitation: opportunities to support independence and therapeutic engagement. Disabil Rehabil Assist Technol. 2014;9:121–7. Faria-Fortini I, Basílio ML, Polese JC, Menezes KKP, Faria CDCM, Scianni AA, et al. Strength deficits of the paretic lower extremity muscles were the impairment variables that best explained restrictions in participation after stroke. Disabil Rehabil. 2017;39:2158–63. Federici S, Meloni F, Bracalenti M, De Filippis ML. The effectiveness of powered, active lower limb exoskeletons in neurorehabilitation: a systematic review. NeuroRehabilitation. 2015;37:321–40.  142 Feigin VL, Krishnamurthi R V., Parmar P, Norrving B, Mensah GA, Bennett DA, et al. Update on the global burden of ischemic and hemorrhagic stroke in 1990-2013: the GBD 2013 study. Neuroepidemiology. 2015;45:161–76. Feigin VL, Roth GA, Naghavi M, Parmar P, Krishnamurthi R, Chugh S, et al. Global burden of stroke and risk factors in 188 countries, during 1990–2013: a systematic analysis for the Global Burden of Disease Study 2013. Lancet Neurol. 2016;15:913–24. Ferrarello F, Baccini M, Rinaldi LA, Cavallini MC, Mossello E, Masotti G, et al. Efficacy of physiotherapy interventions late after stroke: a meta-analysis. J Neurol Neurosurg Psychiatry. 2011;82:136–43. Ferris JK, Edwards JD, Ma JA, Boyd LA. Changes to white matter microstructure in transient ischemic attack: A longitudinal diffusion tensor imaging study. Hum Brain Mapp. 2017;38:5895–5803. Fini NA, Holland AE, Keating J, Simek J, Bernhardt J. How physically active are people following stroke? Systematic review and quantitative synthesis. Phys Ther. 2017;97:707–17. Fisher A, Martin J, Srikusalanukul W, Davis M. Trends in stroke survival incidence rates in older australians in the new millennium and forecasts into the future. J Stroke Cerebrovasc Dis. 2014;23:759–70. Flansbjer UB, Holmbäck AM, Downham D, Patten C, Lexell J. Reliability of gait performance tests in men and women with hemiparesis after stroke. J Rehabil Med. 2005;37:75–82. French B, Thomas L, Leathley M, Sutton C, McAdam J, Forster A, et al. Does repetitive task training improve functional activity after stroke? A Cochrane systematic review and meta-analysis. J Rehabil Med. 2010;42:9–15.  143 Fuentes B, Díez-Tejedor E. Stroke units: many questions, some answers. Int J Stroke. 2009;4:28–37. Fugl-Meyer AR, Jääskö L, Leyman I, Olsson S, Steglind S. The post-stroke hemiplegic patient, 1: a method for evaluation of physical performance. Scand J Rehabil Med. 1975;7:13–31. Fukuda H, Samura K, Hamada O, Saita K, Ogata T, Shiota E, et al. Effectiveness of acute phase Hybrid Assistive Limb rehabilitation in stroke patients classified by paralysis severity. Neurol Med Chir. 2015;55:487–92. Fulk GD, Echternach JL. Test-retest reliability and minimal detectable change of gait speed in individuals undergoing rehabilitation after stroke. J Neurol Phys Ther. 2008;32:8–13. Fulk GD, Echternach JL, Nof L, O’Sullivan S. Clinometric properties of the six-minute walk test in individuals undergoing rehabilitation poststroke. Physiother Theory Pract. 2008;24:195–204. Fulk GD, He Y. Minimal clinically important difference of the 6-minute walk test in people with stroke. J Neurol Phys Ther. 2018;42:235–40. Fulk GD, Reynolds C, Mondal S, Deutsch JE. Predicting home and community walking activity in people with stroke. Arch Phys Med Rehabil. 2010;91:1582–6. Gagnon DH, Aissaoui R. Wearable robotic exoskeletons for overground walking in rehabilitation: specific contexts of use and improvement opportunities. Curr Robot Reports. 2020;1:1–7. Gassert R, Dietz V. Rehabilitation robots for the treatment of sensorimotor deficits: a neurophysiological perspective. J Neuroeng Rehabil. 2018;15:46. Geubbels HJB, Nusselein BAM, Van Heugten CM, Valentijn SAM, Rasquin SMC. Can the Montreal Cognitive Assessment predict discharge destination in a stroke population in the  144 hospital? J Stroke Cerebrovasc Dis. 2015;24:1094–9. Gibson BE, Teachman G. Critical approaches in physical therapy research: investigating the symbolic value of walking. Physiother Theory Pract. 2012;28:474–84. van Gijn J, Kerr RS, Rinkel GJ. Subarachnoid haemorrhage. Lancet. 2007;369:306–18. Gladstone DJ, Danells CJ, Black SE. The Fugl-Meyer assessment of motor recovery after stroke: a critical review of its measurement properties. Neurorehabil Neural Repair. 2002;16:232–40. Grau-Pellicer M, Chamarro-Lusar A, Medina-Casanovas J, Serdà Ferrer BC. Walking speed as a predictor of community mobility and quality of life after stroke. Top Stroke Rehabil. 2019;26:349–58. Hackett ML, Köhler S, O’Brien JT, Mead GE. Neuropsychiatric outcomes of stroke. Lancet Neurol. 2014;13:525–34. Hallett M. Plasticity of the human motor cortex and recovery from stroke. Brain Res Rev. 2001;36:169–74. Hankey GJ. Stroke. Lancet. 2017;389:641–54. Hankey GJ, Hackett ML, Almeida OP, Flicker L, Mead GE, Dennis MS, et al. Safety and efficacy of fluoxetine on functional outcome after acute stroke (AFFINITY): a randomised, double-blind, placebo-controlled trial. Lancet Neurol. 2020;19:651–60. Harris JE, Eng JJ. Goal priorities identified through client-centred measurement in individuals with chronic stroke. Physiother Canada. 2004;56:171–6. Hathidara MY, Saini V, Malik AM. Stroke in the young: a global update. Curr Neurol Neurosci Rep. 2019;19:91. Hays RD, Morales LS. The RAND-36 measure of health-related quality of life. Ann Med.  145 2001;33:350–7. He Y, Yang L, Zhou J, Yao L, Pang MYC. Dual-task training effects on motor and cognitive functional abilities in individuals with stroke: a systematic review. Clin Rehabil. 2018;32:865–77. Heart and Stroke Foundation. 2017 Stroke Report. Different Strokes: Stroke recovery across the ages [Internet]. 2017 [cited 2020 Dec 14]. Available from: Heart and Stroke Foundation. 2016 Stroke Report. Mind the Connection: Preventing stroke and dementia [Internet]. 2016 [cited 2020 Dec 14]. Available from: Hebert D, Lindsay MP, McIntyre A, Kirton A, Rumney PG, Bagg S, et al. Canadian stroke best practice recommendations: stroke rehabilitation practice guidelines, update 2015. Int J Stroke. 2016;11:459–84. Heinemann AW, Jayaraman A, Mummidisetty CK, Spraggins J, Pinto D, Charlifue S, et al. Experience of robotic exoskeleton use at four spinal cord injury model systems centers. J Neurol Phys Ther. 2018;42:256–67. Henderson CE, Fahey M, Brazg G, Moore JL, Hornby TG. Predicting discharge walking function with high-intensity stepping training during inpatient rehabilitation in nonambulatory patients poststroke. Arch Phys Med Rehabil. 2020:6–10. Hennink MM, Kaiser BN, Marconi VC. Code saturation versus meaning saturation. Qual Health  146 Res. 2017;27:591–608. Hesse S, Schattat N, Mehrholz J, Werner C. Evidence of end-effector based gait machines in gait rehabilitation after CNS lesion. NeuroRehabilitation. 2013;33:77–84. Hidler J, Nichols D, Pelliccio M, Brady K, Campbell DD, Kahn JH, et al. Multicenter randomized clinical trial evaluating the effectiveness of the Lokomat in subacute stroke. Neurorehabil Neural Repair. 2009;23:5–13. Hill D, Holloway CS, Morgado Ramirez DZ, Smitham P, Pappas Y. What are user perspectives of exoskeleton technology? A literature review. Int J Technol Assess Health Care. 2017;33:160–7. Hoffmann TC, Glasziou PP, Boutron I, Milne R, Perera R, Moher D, et al. Better reporting of interventions: template for intervention description and replication (TIDieR) checklist and guide. BMJ. 2014;348:g1687. Holden MK, Gill KM, Magliozzi MR, Nathan J, Piehl-Baker L. Clinical gait assessment in the neurologically impaired. Reliability and meaningfulness. Phys Ther. 1984;64:35–40. Holden RJ, Karsh BT. The Technology Acceptance Model: its past and its future in health care. J Biomed Inform. 2010;43:159–72. Hong I, Goodwin JS, Reistetter TA, Kuo YF, Mallinson T, Karmarkar A, et al. Comparison of functional status improvements among patients with stroke receiving postacute care in inpatient rehabilitation vs skilled nursing facilities. JAMA Netw Open. 2019;2:e1916646. Hopman WM, Verner J. Quality of life during and after inpatient stroke rehabilitation. Stroke. 2003;34:801–5. Hornby T, Straube D, Kinnaird C, Holleran C, Echauz A, Rodriguez K, et al. Importance of specificity, amount, and intensity of locomotor training to improve ambulatory function  147 in patients poststroke. Top Stroke Rehabil. 2011;18:293–307. Hornby TG, Holleran CL, Hennessy PW, Leddy AL, Connolly M, Camardo J, et al. Variable intensive early walking poststroke (VIEWS): a randomized controlled trial. Neurorehabil Neural Repair. 2016;30:440–50. Hornby TG, Holleran CL, Leddy AL, Hennessy P, Leech KA, Connolly M, et al. Feasibility of focused stepping practice during inpatient rehabilitation poststroke and potential contributions to mobility outcomes. Neurorehabil Neural Repair. 2015;29:923–32. Hornby TG, Reisman DS, Ward IG, Scheets PL, Miller A, Haddad D, et al. Clinical practice guideline to improve locomotor function following chronic stroke, incomplete spinal cord injury, and brain injury. J Neurol Phys Ther. 2020;44:49–100. Hosmer DW, Lemeshow S. Applied Logistic Regression, Second Edition. New York, NY: Wiley-Interscience; 2000. Hsu CJ, Kim J, Roth EJ, Rymer WZ, Wu M. Forced use of the paretic leginduced by a constraint force applied to the nonparetic leg in individuals poststroke during walking. Neurorehabil Neural Repair. 2017;31:1042–52. Hughes AM, Burridge JH, Demain SH, Ellis-Hill C, Meagher C, Tedesco-Triccas L, et al. Translation of evidence-based assistive technologies into stroke rehabilitation: users’ perceptions of the barriers and opportunities. BMC Health Serv Res. 2014;14:124. Jarvis HL, Brown SJ, Price M, Butterworth C, Groenevelt R, Jackson K, et al. Return to employment after stroke in young adults: how important is the speed and energy cost of walking? Stroke. 2019;50:3198–204. Jayaraman A, Burt S, Rymer WZ. Use of lower-limb robotics to enhance practice and participation in individuals with neurological conditions. Pediatr Phys Ther.  148 2017;29:S48–56. Jayaraman A, O’Brien MK, Madhavan S, Mummidisetty CK, Roth HR, Hohl K, et al. Stride management assist exoskeleton vs functional gait training in stroke: a randomized trial. Neurology. 2019;92:e263–73. Johnson CO, Nguyen M, Roth GA, Nichols E, Alam T, Abate D, et al. Global, regional, and national burden of stroke, 1990–2016: a systematic analysis for the Global Burden of Disease Study 2016. Lancet Neurol. 2019;18:439–58. Jones TA, Adkins DL. Motor system reorganization after stroke: stimulating and training toward perfection. Physiology (Bethesda). 2015;30:358–70. Jørgensen AW, Lundstrøm LH, Wetterslev J, Astrup A, Gøtzsche PC. Comparison of results from different imputation techniques for missing data from an anti-obesity drug trial. PLoS One. 2014;9:e111964. Jørgensen HS. The Copenhagen Stroke Study experience. J Stroke Cerebrovasc Dis. 1996;6:5–16. Jørgensen HS, Nakayama H, Raaschou HO, Olsen TS. Stroke. Neurologic and functional recovery: the Copenhagen Stroke Study. Phys Med Rehabil Clin N Am. 1999;10:887–906. Jørgensen HS, Nakayama H, Raaschou HO, Olsen TS. Recovery of walking function in stroke patients: the Copenhagen Stroke Study. Arch Phys Med Rehabil. 1995;76:27–32. Jørgensen HS, Nakayama H, Reith J, Raaschou HO, Olsen TS. Stroke recurrence: predictors, severity, and prognosis. The Copenhagen Stroke Study. Neurology. 1997;48:891–5. Kaczmarczyk K, Wit A, Krawczyk M, Zaborski J, Gajewski J. Associations between gait patterns, brain lesion factors and functional recovery in stroke patients. Gait Posture.  149 2012;35:214–7. Kamal N, Lindsay MP, Côté R, Fang J, Kapral MK, Hill MD. Ten-year trends in stroke admissions and outcomes in Canada. Can J Neurol Sci. 2015;42:168–75. Kawamoto H, Kamibayashi K, Nakata Y, Yamawaki K, Ariyasu R, Sankai Y, et al. Pilot study of locomotion improvement using hybrid assistive limb in chronic stroke patients. BMC Neurol. 2013;13:141. Kitago T, Krakauer JW. Motor learning principles for neurorehabilitation. In: Barnes MP, Good DC, editors. Handb. Clin. Neurol., vol. 110, Elsevier; 2013, p. 93–103. Klassen TD, Dukelow SP, Bayley MT, Benavente O, Hill MD, Krassioukov A, et al. Higher doses improve walking recovery during stroke inpatient rehabilitation. Stroke. 2020;51:2639–48. Kolakowsky-Hayner SA, Crew J, Moran S, Shah A. Safety and feasibility of using the EksoTM bionic exoskeleton to aid ambulation after spinal cord injury. J Spine. 2013;S4:003. Kollen B, Van De Port I, Lindeman E, Twisk J, Kwakkel G. Predicting improvement in gait after stroke a longitudinal prospective study. Stroke. 2005;36:2676–80. Kollen BJ, Kwakkel G, Lindeman E. Longitudinal robustness of variables predicting independent gait following severe middle cerebral artery stroke: a prospective cohort study. Clin Rehabil. 2006;20:262–8. Krakauer JW. Motor learning: its relevance to stroke recovery and neurorehabilitation. Curr Opin Neurol. 2006;19:84–90. Krakauer JW, Carmichael ST, Corbett D. Getting neurorehabilitation right – what can we learn from animal models? Neurorehabil Neural Repair. 2012;26:923–31. Krakauer JW, Marshall RS. The proportional recovery rule for stroke revisited. Ann Neurol.  150 2015;78:845–7. Kressler J, Thomas CK, Field-Fote EC, Sanchez J, Widerström-Noga E, Cilien DC, et al. Understanding therapeutic benefits of overground bionic ambulation: exploratory case series in persons with chronic, complete spinal cord injury. Arch Phys Med Rehabil. 2014;95:1878–87. Kroenke K, Spitzer RL, Williams JBW. The PHQ-9: validity of a brief depression severity measure. J Gen Intern Med. 2001;16:606–13. Kroll T, Morris J. Challenges and opportunities in using mixed method designs in rehabilitation research. Arch Phys Med Rehabil. 2009;90:S11–6. Krueger H, Koot J, Hall RE, O’Callaghan C, Bayley M, Corbett D. Prevalence of individuals experiencing the effects of stroke in Canada: trends and projections. Stroke. 2015;46:2226–31. Kwakkel G, Kollen B, Twisk J. Impact of time on improvement of outcome after stroke. Stroke. 2006;37:2348–53. Kwakkel G, Kollen BJ. Predicting activities after stroke: what is clinically relevant? Int J Stroke. 2013;8:25–32. Kwakkel G, Wagenaar RC, Koelman TW, Lankhorst GJ, Koetsier JC. Effects of intensity of rehabilitation after stroke: a research synthesis. Stroke. 1997;28:1550–6. Lackland DT, Roccella EJ, Deutsch AF, Fornage M, George MG, Howard G, et al. Factors influencing the decline in stroke mortality a statement from the american heart association/american stroke association. Stroke. 2014;45:315–53. Lacroix J, Daviet J-C, Borel B, Kammoun B, Salle J-Y, Mandigout S. Physical activity level among stroke patients hospitalized in a rehabilitation unit. PM R. 2016;8:97–104.  151 Lam MY. Perceptions of technology and its use for therapeutic application for individuals with hemiparesis: findings from adult and pediatric focus groups. JMIR Rehabil Assist Technol. 2015;2:e1. Lang CE, MacDonald JR, Reisman DS, Boyd L, Jacobson Kimberley T, Schindler-Ivens SM, et al. Observation of amounts of movement practice provided during stroke rehabilitation. Arch Phys Med Rehabil. 2009;90:1692–8. Langan J, Subryan H, Nwogu I, Cavuoto L. Reported use of technology in stroke rehabilitation by physical and occupational therapists. Disabil Rehabil Assist Technol. 2017:1–7. Langhorne P. Organised inpatient (stroke unit) care for stroke. Cochrane Database Syst Rev. 2013;2013:CD000197. Langhorne P, Wagenaar R, Partridge C. Physiotherapy after stroke: more is better? Physiother Res Int. 1996;1:75–88. Lapadatu I, Morris R. The relationship between stroke survivors’ perceived identity and mood, self-esteem and quality of life. Neuropsychol Rehabil. 2019;29:199–213. Laucis NC, Hays RD, Bhattacharyya T. Scoring the SF-36 in orthopaedics: a brief guide. J Bone Jt Surg - Am Vol. 2014;97:1628–34. Lefeber N, De Buyzer S, Dassen N, De Keersmaecker E, Kerckhofs E, Swinnen E. Energy consumption and cost during walking with different modalities of assistance after stroke: a systematic review and meta-analysis. Disabil Rehabil. 2020;42:1650–66. Lefeber N, De Keersmaecker E, Henderix S, Michielsen M, Kerckhofs E, Swinnen E. Physiological responses and perceived exertion during robot-assisted and body weight–supported gait after stroke. Neurorehabil Neural Repair. 2018;32:1043–54. Lefeber N, Swinnen E, Kerckhofs E. The immediate effects of robot-assistance on energy  152 consumption and cardiorespiratory load during walking compared to walking without robot-assistance: a systematic review. Disabil Rehabil Assist Technol. 2017;12:657–71. Levac D, Colquhoun H, O’Brien KK. Scoping studies: advancing the methodology. Implement Sci. 2010;5:69. Levac DE, Miller PA. Integrating virtual reality video games into practice: clinicians’ experiences. Physiother Theory Pract. 2013;29:504–12. Levin MF, Panturin E. Sensorimotor integration for functional recovery and the Bobath approach. Motor Control. 2011;15:285–301. Li S, Francisco GE, Zhou P. Post-stroke hemiplegic gait: new perspective and insights. Front Physiol. 2018;9:1021. Liu-Ambrose T, Pang MYC, Eng JJ. Executive function is independently associated with performances of balance and mobility in community-dwelling older adults after mild stroke: implications for falls prevention. Cerebrovasc Dis. 2007;23:203–10. Liu L, Miguel Cruz A, Rios Rincon A, Buttar V, Ranson Q, Goertzen D. What factors determine therapists’ acceptance of new technologies for rehabilitation – a study using the Unified Theory of Acceptance and Use of Technology (UTAUT). Disabil Rehabil. 2015;37:447–55. Lo A, Tahair N, Sharp S, Bayley MT. Clinical utility of the AlphaFIM® instrument in stroke rehabilitation. Int J Stroke. 2012;7:118–24. Lopez AD, Mathers CD, Ezzati M, Jamison DT, Murray CJL. Global and regional burden of disease and risk factors, 2001: systematic analysis of population health data. Lancet. 2006;367:1747–57. Louie DR, Eng JJ. Berg Balance Scale score at admission can predict walking suitable for  153 community ambulation at discharge from inpatient stroke rehabilitation. J Rehabil Med. 2018;50:37–44. Louie DR, Eng JJ. Powered robotic exoskeletons in post-stroke rehabilitation of gait: a scoping review. J Neuroeng Rehabil. 2016;13:53. Louie DR, Eng JJ, Lam T, SCIRE Research Team. Gait speed using powered robotic exoskeletons after spinal cord injury: a systematic review and correlational study. J Neuroeng Rehabil. 2015;12:82. Louie DR, Mortenson WB, Durocher M, Teasell R, Yao J, Eng JJ. Exoskeleton for post-stroke recovery of ambulation (ExStRA): study protocol for a mixed-methods study investigating the efficacy and acceptance of an exoskeleton-based physical therapy program during stroke inpatient rehabilitation. BMC Neurol. 2020;20:35. Luengo-Fernandez R, Paul NLM, Gray AM, Pendlebury ST, Bull LM, Welch SJ V, et al. Population-based study of disability and institutionalization after transient ischemic attack and stroke: 10-year results of the Oxford Vascular Study. Stroke. 2013;44:2854–61. Luo J, Hendryx M, Safford MM, Wallace R, Rossom R, Eaton C, et al. Newly developed chronic conditions and changes in health-related quality of life in postmenopausal women. J Am Geriatr Soc. 2015;63:2349–57. Lyden P. Using the National Institutes of Health Stroke Scale. Stroke. 2017;48:513–9. Lyden P, Brott T, Tilley B, Welch KMA, Mascha EJ, Levine S, et al. Improved reliability of the NIH stroke scale using video training. Stroke. 1994;25:2220–6. Maaijwee NAMM, Rutten-Jacobs LCA, Schaapsmeerders P, Van Dijk EJ, De Leeuw FE. Ischaemic stroke in young adults: risk factors and long-term consequences. Nat Rev  154 Neurol. 2014;10:315–25. Maeshima S, Osawa A, Nishio D, Hirano Y, Takeda K, Kigawa H, et al. Efficacy of a hybrid assistive limb in post-stroke hemiplegic patients: a preliminary report. BMC Neurol. 2011;11:116. de Man-van Ginkel JM, Gooskens F, Schepers VPM, Schuurmans MJ, Lindeman E, Hafsteinsdóttir TB. Screening for poststroke depression using the Patient Health Questionnaire. Nurs Res. 2012;61:333–41. Mandrekar JN. Receiver operating characteristic curve in diagnostic test assessment. J Thorac Oncol. 2010;5:1315–6. Manns PJ, Hurd C, Yang JF. Perspectives of people with spinal cord injury learning to walk using a powered exoskeleton. J Neuroeng Rehabil. 2019;16:94. Markus HS, Bevan S. Mechanisms and treatment of ischaemic stroke - insights from genetic associations. Nat Rev Neurol. 2014;10:723–30. Mayo NE, Wood-Dauphinee S, Ahmed S, Gordon C, Higgins J, McEwen S, et al. Disablement following stroke. Disabil Rehabil. 1999;21:258–68. Mayr A, Quirbach E, Picelli A, Kofler M, Smania N, Saltuari L. Early robot-assisted gait retraining in non-ambulatory patients with stroke: a single blind randomized controlled trial. Eur J Phys Rehabil Med. 2018;54:819–26. Mays N, Pope C. Assessing quality in qualitative research. BMJ. 2000;320:50–2. McHugh G, Swain ID. A comparison between reported therapy staffing levels and the department of health therapy staffing guidelines for stroke rehabilitation: a national survey. BMC Health Serv Res. 2014;14:1–9. Mehrholz J, Elsner B, Werner C, Kugler J, Pohl M. Electromechanical-assisted training for  155 walking after stroke (Review). Cochrane Database Syst Rev. 2013;2013:CD006185. Mehrholz J, Pohl M, Elsner B. Treadmill training and body weight support for walking after stroke. Cochrane Database Syst Rev. 2014;2014:CD002840. Mehrholz J, Pohl M, Kugler J, Elsner B. The improvement of walking ability following stroke: a systematic review and network meta-analysis of randomized controlled trials. Dtsch Arztebl Int. 2018;115:639–45. Mehrholz J, Thomas S, Werner C, Kugler J, Pohl M, Elsner B. Electromechanical-assisted training for walking after stroke (Review). Cochrane Database Syst Rev. 2017;5:CD006185. Mehrholz J, Wagner K, Rutte K, Meissner D, Pohl M. Predictive validity and responsiveness of the Functional Ambulation Category in hemiparetic patients after stroke. Arch Phys Med Rehabil. 2007;88:1314–9. Mehrholz J, Werner C, Kugler J, Pohl M. Electromechanical-assisted training for walking after stroke. What is the evidence so far? What have we learnt? Physiotherapy. 2015;101:eS990–1. Mercer VS, Freburger JK, Yin Z, Preisser JS. Recovery of paretic lower extremity loading ability and physical function in the first six months after stroke. Arch Phys Med Rehabil. 2014;95:1547–55. Min KB, Min JY. Health-related quality of life is associated with stroke deficits in older adults. Age Ageing. 2015;44:700–4. Minet LR, Peterson E, von Koch L, Ytterberg C. Healthcare utilization after stroke: a 1-year prospective study. J Am Med Dir Assoc. 2020;21:1684–8. Mittmann N, Seung SJ, Hill MD, Phillips SJ, Hachinski V, Coté R, et al. Impact of disability  156 status on ischemic stroke costs in Canada in the first year. Can J Neurol Sci. 2012;39:793–800. Molteni F, Gasperini G, Gaffuri M, Colombo M, Giovanzana C, Lorenzon C, et al. Wearable robotic exoskeleton for overground gait training in sub-acute and chronic hemiparetic stroke patients: preliminary results. Eur J Phys Rehabil Med. 2017;53:676–84. Morone G, Paolucci S, Cherubini A, De Angelis D, Venturiero V, Coiro P, et al. Robot-assisted gait training for stroke patients: current state of the art and perspectives of robotics. Neuropsychiatr Dis Treat. 2017;13:1303–11. Morrow SL. Quality and trustworthiness in qualitative research in counseling psychology. J Couns Psychol. 2005;52:250–60. Mortari L. Reflectivity in research practice: an overview of different perspectives. Int J Qual Methods. 2015;14. Mortenson WB, Pysklywec A, Chau L, Prescott M, Townson A. Therapists’ experience of training and implementing an exoskeleton in a rehabilitation centre. Disabil Rehabil. 2020:1–7. Moseley AM, Stark A, Cameron ID, Pollock A. Treadmill training and body weight support for walking after stroke. Cochrane Database Syst Rev. 2005:CD002840. Moucheboeuf G, Griffier R, Gasq D, Glize B, Bouyer L, Dehail P, et al. Effects of robotic gait training after stroke: a meta-analysis. Ann Phys Rehabil Med. 2020;63:518–34. Muka T, Imo D, Jaspers L, Colpani V, Chaker L, van der Lee SJ, et al. The global impact of non-communicable diseases on healthcare spending and national income: a systematic review. Eur J Epidemiol. 2015;30:251–77. Mukherjee D, Patil CG. Epidemiology and the global burden of stroke. World Neurosurg.  157 2011;76:S85–90. Muren MA, Hütler M, Hooper J. Functional capacity and health-related quality of life in individuals post stroke. Top Stroke Rehabil. 2008;15:51–8. Nam YG, Lee JW, Park JW, Lee HJ, Nam KY, Park JH, et al. Effects of electromechanical exoskeleton-assisted gait training on walking ability of stroke patients: a randomized controlled trial. Arch Phys Med Rehabil. 2019;100:26–31. Nasreddine ZS, Phillips NA, Bédirian V, Charbonneau S, Whitehead V, Collin I, et al. The Montreal Cognitive Assessment, MoCA: a brief screening tool for mild cognitive impairment. J Am Geriatr Soc. 2005;53:695–9. Nilsson A, Vreede KS, Häglund V, Kawamoto H, Sankai Y, Borg J. Gait training early after stroke with a new exoskeleton - the hybrid assistive limb: a study of safety and feasibility. J Neuroeng Rehabil. 2014;11:92. Nudo RJ. Neural bases of recovery after brain injury. J Commun Disord. 2011;44:515–20. Nudo RJ, Plautz EJ, Frost SB. Role of adaptive plasticity in recovery of function after damage to motor cortex. Muscle and Nerve. 2001;24:1000–19. O’Brien BC, Harris IB, Beckman TJ, Reed DA, Cook DA. Standards for reporting qualitative research: a synthesis of recommendations. Acad Med. 2014;89:1245–51. O’Cathain A, Murphy E, Nicholl J. Three techniques for integrating data in mixed methods studies. BMJ. 2010;341:1147–50. van Ommeren AL, Smulders LC, Prange-Lasonder GB, Buurke JH, Veltink PH, Rietman JS. Assistive technology for the upper extremities after stroke: systematic review of users’ needs. JMIR Rehabil Assist Technol. 2018;5:e10510. Ouellette DS, Timple C, Kaplan SE, Rosenberg SS, Rosario ER. Predicting discharge destination  158 with admission outcome scores in stroke patients. NeuroRehabilitation. 2015;37:173–9. Owusu T. Ekso stock doubles - FDA widens marketing OK for exoskeleton [Internet]. TheStreet. 2020 [cited 2020 Dec 14]. Available from: Pace A, Mitchell S, Casselden E, Zolnourian A, Glazier J, Foulkes L, et al. A subarachnoid haemorrhage-specific outcome tool. Brain. 2018;141:1111–21. Page SJ. Intensity versus task-specificity after stroke: how important is intensity? Am J Phys Med Rehabil. 2003;82:730–2. Pandian S, Arya KN, Kumar D. Minimal clinically important difference of the lower-extremity Fugl–Meyer Assessment in chronic-stroke. Top Stroke Rehabil. 2016;23:233–9. Pang M, Eng J, Dawson AS, McKay H, Harris JE. A community-based fitness and mobility exercise program for older adults with chronic stroke: a randomized, controlled Trial. J Am Geriatr Soc. 2005;53:1667–74. Paolucci S, Antonucci G, Grasso MG, Bragoni M, Coiro P, De Angelis D, et al. Functional outcome of ischemic and hemorrhagic stroke patients after inpatient rehabilitation: a matched comparison. Stroke. 2003;34:2861–5. Pare JR, Kahn JH. Basic neuroanatomy and stroke syndromes. Emerg Med Clin North Am. 2012;30:601–15. Park EY, Choi YI. Psychometric properties of the lower extremity subscale of the Fugl-Myer Assessment for community-dwelling hemiplegic stroke patients. J Phys Ther Sci. 2014;26:1775–7. Patterson KK, Mansfield A, Biasin L, Brunton K, Inness EL, McIlroy WE. Longitudinal changes  159 in poststroke spatiotemporal gait asymmetry over inpatient rehabilitation. Neurorehabil Neural Repair. 2015;29:153–62. Patterson KK, Wong JS, Knorr S, Grahn JA. Rhythm perception and production abilities and their relationship to gait after stroke. Arch Phys Med Rehabil. 2018;99:945–51. Van Peppen RPS, Kwakkel G, Wood-Dauphinee S, Hendriks HJM, Van der Wees PJ, Dekker J. The impact of physical therapy on functional outcomes after stroke: what’s the evidence? Clin Rehabil. 2004;18:833–62. Pereira S, Foley N, Salter K, McClure JA, Meyer M, Brown J, et al. Discharge destination of individuals with severe stroke undergoing rehabilitation: a predictive model. Disabil Rehabil. 2014;36:727–31. Perera S, Mody SH, Woodman RC, Studenski SA. Meaningful change and responsiveness in common physical performance measures in older adults. J Am Geriatr Soc. 2006;54:743–9. Peters S, Lim SB, Louie DR, Yang CL, Eng JJ. Passive, yet not inactive: robotic exoskeleton walking increases cortical activation dependent on task. J Neuroeng Rehabil. 2020;17:107. Plano Clark VL, Creswell JW. An Expanded Typology for Classifying Mixed Methods Research Into Designs. Mix. Methods Read., Thousand Oaks: Sage Publications; 2007, p. 159–96. van de Port IGL, Wevers LEG, Lindeman E, Kwakkel G. Effects of circuit training as alternative to usual physiotherapy after stroke: randomised controlled trial. BMJ. 2012;344:e2672–e2672. Portelli R, Lowe D, Irwin P, Pearson M, Rudd AG. Institutionalization after stroke. Clin Rehabil. 2005;19:97–108.  160 Postol N, Marquez J, Spartalis S, Bivard A, Spratt NJ. Do powered over-ground lower limb robotic exoskeletons affect outcomes in the rehabilitation of people with acquired brain injury? Disabil Rehabil Assist Technol. 2019;14:764–75. Pournajaf S, Goffredo M, Agosti M, Massucci M, Ferro S, Franceschini M, et al. Community ambulation of stroke survivors at 6 months follow-up: an observational study on sociodemographic and sub-acute clinical indicators. Eur J Phys Rehabil Med. 2019;55:433–41. Prabhakaran S, Ruff I, Bernstein RA. Acute stroke intervention: A systematic review. JAMA - J Am Med Assoc. 2015;313:1451–62. Prabhakaran S, Zarahn E, Riley C, Speizer A, Chong JY, Lazar RM, et al. Inter-individual variability in the capacity for motor recovery after ischemic stroke. Neurorehabil Neural Repair. 2008;22:64–71. Probst B. The eye regards itself: benefits and challenges of reflexivity in qualitative social work research. Soc Work Res. 2015;39:37–48. Pundik S, Holcomb J, Mccabe J, Daly JJ. Enhanced life-role participation in response to comprehensive gait training in chronic stroke survivors. Disabil Rehabil. 2012;34:1535–9. Quinn TJ, Dawson J, Lees JS, Chang TP, Walters MR, Lees KR. Time spent at home poststroke: “home-time” a meaningful and robust outcome measure for stroke trials. Stroke. 2008;39:231–3. Rabinstein AA. Treatment of acute ischemic stroke. Contin (Minneap Minn). 2017;23:62–81. Rand D, Eng JJ. Predicting daily use of the affected upper extremity 1 year after stroke. J Stroke Cerebrovasc Dis. 2015;24:274–83.  161 Rand D, Eng JJ. Disparity between functional recovery and daily use of the upper and lower extremities during subacute stroke rehabilitation. Neurorehabil Neural Repair. 2012;26:76–84. Rauscher L, Greenfield BH. Advancements in contemporary physical therapy research: use of mixed methods designs. Phys Ther. 2009;89:91–100. Read E, Woolsey C, McGibbon CA, O’Connell C. Physiotherapists’ experiences using the Ekso bionic exoskeleton with patients in a neurological rehabilitation hospital: a qualitative study. Rehabil Res Pract. 2020;2020:2939573. Reznik ME, Yaghi S, Jayaraman M V., McTaggart RA, Hemendinger M, Mac Grory BC, et al. Baseline NIH Stroke Scale is an inferior predictor of functional outcome in the era of acute stroke intervention. Int J Stroke. 2018;13:806–10. Rodríguez-Castro E, López-Dequit I, Santamaría-Cadavid M, Arias-Rivas S, Rodríguez-Yáñez M, Pumar JM, et al. Trends in stroke outcomes in the last ten years in a European tertiary hospital. BMC Neurol. 2018;18:164. Rojek A, Mika A, Oleksy Ł, Stolarczyk A, Kielnar R. Effects of exoskeleton gait training on balance, load distribution, and functional status in stroke: a randomized controlled trial. Front Neurol. 2020;10:1–12. Rose DK, Demark L, Fox EJ, Clark DJ, Wludyka P. A backward walking training program to improve balance and mobility in acute stroke: a pilot randomized controlled trial. J Neurol Phys Ther. 2018;42:12–21. Rozanski GM, Wong JS, Inness EL, Patterson KK, Mansfield A. Longitudinal change in spatiotemporal gait symmetry after discharge from inpatient stroke rehabilitation. Disabil Rehabil. 2020;42:705–11.  162 Rudberg A-S, Berge E, Laska A-C, Jutterström S, Näsman P, Sunnerhagen KS, et al. Stroke survivors’ priorities for research related to life after stroke. Top Stroke Rehabil. 2020:1–6. Rundek T, Mast H, Hartmann A, Boden-Albala B, Lennihan L, Lin IF, et al. Predictors of resource use after acute hospitalization: the Northern Manhattan Stroke Study. Neurology. 2000;55:1180–7. Salbach NM, Mayo NE, Higgins J, Ahmed S, Finch LE, Richards CL. Responsiveness and predictability of gait speed and other disability measures in acute stroke. Arch Phys Med Rehabil. 2001;82:1204–12. Sale P, Franceschini M, Waldner A, Hesse S. Use of the robot assisted gait therapy in rehabilitation of patients with stroke and spinal cord injury. Eur J Phys Rehabil Med. 2012;48:111–21. Sandelowski M. What’s in a name? Qualitative description revisited. Res Nurs Heal. 2010;33:77–84. Sandelowski M. Whatever happened to qualitative description? Res Nurs Health. 2000;23:334–40. Saposnik G, Barinagarrementeria F, Brown RD, Bushnell CD, Cucchiara B, Cushman M, et al. Diagnosis and management of cerebral venous thrombosis: a statement for healthcare professionals from the American Heart Association/American Stroke Association. Stroke. 2011;42:1158–92. Saunders B, Sim J, Kingstone T, Baker S, Waterfield J, Bartlam B, et al. Saturation in qualitative research: exploring its conceptualization and operationalization. Qual Quant. 2018;52:1893–907. Schlegel D, Kolb SJ, Luciano JM, Tovar JM, Cucchiara BL, Liebeskind DS, et al. Utility of the  163 NIH Stroke Scale as a predictor of hospital disposition. Stroke. 2003;34:134–7. Schlegel DJ, Tanne D, Demchuk AM, Levine SR, Kasner SE. Prediction of hospital disposition after thrombolysis for acute ischemic stroke using the National Institutes of Health Stroke Scale. Arch Neurol. 2004;61:1061–4. Schulz KF, Altman DG, Moher D. CONSORT 2010 statement: updated guidelines for reporting parallel group randomized trials. Ann Intern Med. 2010;152:726–32. Schwartz I, Meiner Z. Robotic-assisted gait training in neurological patients: who may benefit? Ann Biomed Eng. 2015;43:1260–9. Scrivener K, Sherrington C, Schurr K. Exercise dose and mobility outcome in a comprehensive stroke unit: description and prediction from a prospective cohort study. J Rehabil Med. 2012;44:824–9. Sczesny-Kaiser M, Trost R, Aach M, Schildhauer TA, Schwenkreis P, Tegenthoff M. A randomized and controlled crossover study investigating the improvement of walking and posture functions in chronic stroke patients using HAL exoskeleton – the HALESTRO study (HAL-Exoskeleton STROke study). Front Neurosci. 2019;13:1–13. Shum ST, Chiu JKW, Tsang CPL, Wong CHP, Tsang RCC, Ma SL, et al. Predicting walking function of patients one month poststroke using Modified Rivermead Mobility Index on admission. J Stroke Cerebrovasc Dis. 2014;23:2117–21. Sidler MR. Impact of technology on rehabilitation. Occup Ther Heal Care. 1986;3:55–84. Silvis SM, De Sousa DA, Ferro JM, Coutinho JM. Cerebral venous thrombosis. Nat Rev Neurol. 2017;13:555–65. Simmons FM. Hospital overcrowding: an opportunity for case managers. Case Manager. 2005;16:52–4.  164 Skarin M, Sjöholm A, Nilsson ÅL, Nilsson M, Bernhardt J, Lindén T. A mapping study on physical activity in stroke rehabilitation: establishing the baseline. J Rehabil Med. 2013;45:997–1003. Smith MC, Byblow WD, Barber PA, Stinear CM. Proportional recovery from lower limb motor impairment after stroke. Stroke. 2017;48:1400–3. Smith TJ, McKenna CM. A comparison of logistic regression pseudo R2 indices. Mult Linear Regres Viewpoints. 2013;39:17–26. Spratt N, Wang Y, Levi C, Ng K, Evans M, Fisher J. A prospective study of predictors of prolonged hospital stay and disability after stroke. J Clin Neurosci. 2003;10:665–9. States RA, Salem Y, Pappas E. Overground gait training for individuals with chronic stroke: a Cochrane systematic review. J Neurol Phys Ther. 2009;33:179–86. Statistics Canada. Lower Mainland--Southwest [Economic region], British Columbia and British Columbia [Province] (table) [Internet]. Census Profile 2016 Census. 2017 [cited 2020 June 30]. Availablel from: Mainland--Southwest&SearchType=Begins&SearchPR=01&B1=All&GeoLevel=PR&GeoCode=5920&TABID=1&type=0 Stein J, Bishop L, Stein DJ, Wong CK. Gait training with a robotic leg brace after stroke. Am J Phys Med Rehabil. 2014;93:987–94. Stephenson A, Stephens J. An exploration of physiotherapists’ experiences of robotic therapy in upper limb rehabilitation within a stroke rehabilitation centre. Disabil Rehabil Assist Technol. 2018;13:245–52.  165 Stevenson TJ. Detecting change in patients with stroke using the Berg Balance Scale. Aust J Physiother. 2001;47:29–38. Stillman G, Granger C, Niewczyk P. Projecting function of stroke patients in rehabilitation using the AlphaFIM instrument in acute care. PM R. 2009;1:234–9. Stinear CM, Byblow WD, Ackerley SJ, Smith MC, Borges VM, Barber PA. Proportional motor recovery after stroke: implications for trial design. Stroke. 2017;48:795–8. Stinear CM, Petoe MA, Byblow WD. Primary motor cortex excitability during recovery after stroke: implications for neuromodulation. Brain Stimul. 2015;8:1183–90. Straub ET. Understanding technology adoption: theory and future directions for informal learning. Rev Educ Res. 2009;79:625–49. Stynes M, Murphy T, McNamara G, O’Hara J. Reflection-on-action in qualitative research: a critical self-appraisal rubric for deconstructing research. Issues Educ Res. 2018;28:153–67. Sung SF, Su CC, Hsieh CY, Cheng CL, Chen CH, Lin HJ, et al. Home-time as a surrogate measure for functional outcome after stroke: a validation study. Clin Epidemiol. 2020;12:617–24. Sutherland JM, Trafford Crump R. Alternative level of care: Canada’s hospital beds, the evidence and options. Healthc Policy. 2013;9:26–34. Swank C, Sikka S, Driver S, Bennett M, Callender L. Feasibility of integrating robotic exoskeleton gait training in inpatient rehabilitation. Disabil Rehabil Assist Technol. 2020;15:409–17. Tan CK, Kadone H, Watanabe H, Marushima A, Hada Y, Yamazaki M, et al. Differences in muscle synergy symmetry between subacute post-stroke patients with bioelectrically- 166 controlled exoskeleton gait training and conventional gait training. Front Bioeng Biotechnol. 2020;8:1–19. Tang A, Eng JJ, Krassioukov A V, Madden KM, Mohammadi A, Tsang MYC, et al. Exercise-induced changes in cardiovascular function after stroke: a randomized controlled trial. Int J Stroke. 2014;9:883–9. Tang A, Eng JJ, Rand D. Relationship between perceived and measured changes in walking after stroke. J Neurol Phys Ther. 2012;36:115–21. Tanwir S, Montgomery K, Chari V, Nesathurai S. Stroke rehabilitation: availability of a family member as caregiver and discharge destination. Eur J Phys Rehabil Med. 2014;50:355–62. Taraldsen K, Askim T, Sletvold O, Einarsen EK, Grüner Bjåstad K, Indredavik B, et al. Evaluation of a body-worn sensor system to measure physical activity in older people with impaired function. Phys Ther. 2011;91:277–85. Taub E, Uswatte G, Mark VW, Morris DM. The learned nonuse phenomenon: implications for rehabilitation. Eura Medicophys. 2006;42:241–56. Taylor JK, Buchan IE, van der Veer SN. Assessing life-space mobility for a more holistic view on wellbeing in geriatric research and clinical practice. Aging Clin Exp Res. 2019;31:439–45. Teasell R, Hussein N. 4. Motor Rehabilitation 4A. Lower Extremity and Mobility 4.1 Motor Recovery of the Lower Extremity Post Stroke Factors that Predict Motor Recovery. Stroke Rehabil Clin Handb. 2016:1–27. Thomassen GKK, Jørgensen V, Normann B. “Back at the same level as everyone else”-user perspectives on walking with an exoskeleton, a qualitative study. Spinal Cord Ser Cases.  167 2019;5:103. Tieges Z, Mead G, Allerhand M, Duncan F, Van Wijck F, Fitzsimons C, et al. Sedentary behavior in the first year after stroke: a longitudinal cohort study with objective measures. Arch Phys Med Rehabil. 2015;96:15–23. Tilson JK, Sullivan KJ, Cen SY, Rose DK, Koradia CH, Azen SP, et al. Meaningful gait speed improvement during the first 60 days poststroke: minimal clinically important difference. Phys Ther. 2010;90:196–208. Toglia J, Fitzgerald KA, O&apos;Dell MW, Mastrogiovanni AR, Lin CD. The mini-mental state examination and montreal cognitive assessment in persons with mild subacute stroke: Relationship to functional outcome. Arch Phys Med Rehabil. 2011;92:792–8. Tourangeau AE, Squires ME, Wodchis W, McGilton K, Teare G, Widger KA. Post-acute stroke patient outcomes in Ontario, Canada complex continuing care settings. Disabil Rehabil. 2011;33:98–104. Treger I, Ring H, Schwartz R, Tsabari R, Bornstein NM, Tanne D. Hospital disposition after stroke in a national survey of acute cerebrovascular diseases in Israel. Arch Phys Med Rehabil. 2008;89:435–40. Tsurushima H, Mizukami M, Yoshikawa K, Ueno T, Hada Y, Gosho M, et al. Effectiveness of a walking program involving the Hybrid Assistive Limb robotic exoskeleton suit for improving walking ability in stroke patients: protocol for a randomized controlled trial. JMIR Res Protoc. 2019;8:e14001. Turchetti G, Vitiello N, Trieste L, Romiti S, Geisler E, Micera S. Why effectiveness of robot-mediated neurorehabilitation does not necessarily influence its adoption. IEEE Rev Biomed Eng. 2014;7:143–53.  168 Uniform Data System for Medical Rehabilitation, editor. The AlphaFIM Instrument Guide, Version 4.01. Buffalo, NY: UDSMR; 2009. Uniform Data System for Medical Rehabilitation, editor. Guide for the Uniform Data Set for Medical Rehabilitation (Including the FIM(TM) Instrument), Version 5.1. Buffalo, NY: UDSMR; 1997. Vasileiou K, Barnett J, Thorpe S, Young T. Characterising and justifying sample size sufficiency in interview-based studies: systematic analysis of qualitative health research over a 15-year period. BMC Med Res Methodol. 2018;18:148. Vaughan-Graham J, Brooks D, Rose L, Nejat G, Pons J, Patterson K. Exoskeleton use in post-stroke gait rehabilitation: a qualitative study of the perspectives of persons post-stroke and physiotherapists. J Neuroeng Rehabil. 2020;17:1–15. Venkatesh V, Morris MG, Davis GB, Davis FD. User accceptance of information technology: toward a unified view. MIS Q. 2003;27:425–78. Vieira ER, Schneider P, Guidera C, Gadotti IC, Brunt D. Work-related musculoskeletal disorders among physical therapists: a systematic review. J Back Musculoskelet Rehabil. 2016;29:417–28. Virani SS, Alonso A, Benjamin EJ, Bittencourt MS, Callaway CW, Carson AP, et al. Heart disease and stroke statistics—2020 update: a report from the American Heart Association. 2020. Vluggen TPMM, Van Haastregt JCM, Tan FES, Kempen GIJM, Schols JMGA, Verbunt JA. Factors associated with successful home discharge after inpatient rehabilitation in frail older stroke patients. BMC Geriatr. 2020;20:25. Wafa HA, Wolfe CDA, Bhalla A, Wang Y. Long-term trends in death and dependence after  169 ischaemic strokes: a retrospective cohort study using the South London Stroke Register (SLSR). PLoS Med. 2020;17:1–17. Wall A, Borg J, Palmcrantz S. Clinical application of the Hybrid Assistive Limb (HAL) for gait training - a systematic review. Front Syst Neurosci. 2015;9:48. Wall A, Borg J, Vreede K, Palmcrantz S. A randomized controlled study incorporating an electromechanical gait machine, the Hybrid Assistive Limb, in gait training of patients with severe limitations in walking in the subacute phase after stroke. PLoS One. 2020;15:e0229707. Watanabe H, Goto R, Tanaka N, Matsumura A, Yanagi H. Effects of gait training using the Hybrid Assistive Limb® in recovery-phase stroke patients: a 2-month follow-up, randomized, controlled study. NeuroRehabilitation. 2017;40:363–7. Watanabe H, Tanaka N, Inuta T, Saitou H, Yanagi H. Locomotion improvement using a hybrid assistive limb in recovery phase stroke patients: a randomized controlled pilot study. Arch Phys Med Rehabil. 2014;95:2006–12. WHO. International Classification of Functioning, Disability and Health. Int. Classif. Funct. Disabil. Heal., 2001. Wilkinson DA, Pandey AS, Thompson G, Keep RF, Hua Y, Xi G. Injury mechanisms in acute intracerebral hemorrhage. Neuropharmacology. 2018;134:240–8. Willems D, Salter K, Meyer M, Mcclure A, Teasel R, Foley N. Determining the need for in-patient rehabilitation services post-stroke: results from eight Ontario hospitals. Healthc Policy. 2012;7:105–18. Williams LR, Leggett RW. Reference values for resting blood flow to organs of man. Clin Phys Physiol Meas. 1989;10:187–217.  170 De Wit L, Theuns P, Dejaeger E, Devos S, Gantenbein AR, Kerckhofs E, et al. Long-term impact of stroke on patients’ health-related quality of life. Disabil Rehabil. 2017;39:1435–40. Wolpert DM, Goodbody SJ, Husain M. Maintaining internal representations: The role of the human superior parietal lobe. Nat Neurosci. 1998;1:529–33. Wondergem R, Veenhof C, Wouters EMJ, De Bie RA, Visser-Meily JMA, Pisters MF. Movement behavior patterns in people with first-ever stroke. Stroke. 2019;50:3553–60. Wong CK, Bishop L, Stein J. A wearable robotic knee orthosis for gait training: a case-series of hemiparetic stroke survivors. Prosthet Orthot Int. 2012;36:113–20. World Health Organization. Global Health Estimates [Internet]. 2020 [cited 2020 Aug 11]. Available from: Wu CY, Hung SJ, Lin KC, Chen KH, Chen P, Tsay PK. Responsiveness, minimal clinically important difference, and validity of the MoCA in stroke rehabilitation. Occup Ther Int. 2019;2019:2517658. Xing CY, Tarumi T, Liu J, Zhang Y, Turner M, Riley J, et al. Distribution of cardiac output to the brain across the adult lifespan. J Cereb Blood Flow Metab. 2017;37:2848–56. Yaghi S, Willey JZ, Andrews H, Boehme AK, Marshall RS, Boden-Albala B. The itemized NIHSS scores are associated with discharge disposition in patients with minor stroke. The Neurohospitalist. 2016;6:102–6. Yen HC, Chen WS, Jeng JS, Luh JJ, Lee YY, Pan GS. Standard early rehabilitation and lower limb transcutaneous nerve or neuromuscular electrical stimulation in acute stroke patients: a randomized controlled pilot study. Clin Rehabil. 2019;33:1344–54. Yew KS, Cheng EM. Diagnosis of acute stroke. Am Fam Physician. 2015;91:528–36.  171 Yokota C, Yamamoto Y, Kamada M, Nakai M, Nishimura K, Ando D, et al. Acute stroke rehabilitation for gait training with cyborg type robot Hybrid Assistive Limb: a pilot study. J Neurol Sci. 2019;404:11–5. Yoshimoto T, Shimizu I, Hiroi Y, Kawaki M, Sato D, Nagasawa M. Feasibility and efficacy of high-speed gait training with a voluntary driven exoskeleton robot for gait and balance dysfunction in patients with chronic stroke: nonrandomized pilot study with concurrent control. Int J Rehabil Res. 2015;38:338–43. Yu AYX, Fang J, Porter J, Austin PC, Smith EE, Kapral MK. Hospital-based cohort study to determine the association between home-time and disability after stroke by age, sex, stroke type and study year in Canada. BMJ Open. 2019;9:e031379. Zhang W, Creswell J. The use of “mixing” procedure of mixed methods in health services research. Med Care. 2013;51:e51–7. Zhang Z. Model building strategy for logistic regression: purposeful selection. Ann Transl Med. 2016;4:4–10. Zuurbier SM, Coutinho JM. Cerebral venous thrombosis. Adv Exp Med Biol. 2017;906:183–93.   172 Appendices Appendix A  Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) Statement  Item No. Recommendation Page No. Title and abstract 1 (a) Indicate the study’s design with a commonly used term in the title or the abstract No abstract (b) Provide in the abstract an informative and balanced summary of what was done and what was found No abstract Introduction Background/rationale 2 Explain the scientific background and rationale for the investigation being reported 27 – 28 Objectives 3 State specific objectives, including any prespecified hypotheses 28 – 29 Methods Study design 4 Present key elements of study design early in the paper 28 Setting 5 Describe the setting, locations, and relevant dates, including periods of recruitment, exposure, follow-up, and data collection 28 Participants 6 (a) Cohort study—Give the eligibility criteria, and the sources and methods of selection of participants. Describe methods of follow-up Case-control study—Give the eligibility criteria, and the sources and methods of case ascertainment and control selection. Give the rationale for the choice of cases and controls Cross-sectional study—Give the eligibility criteria, and the sources and methods of selection of participants 28 (b) Cohort study—For matched studies, give matching criteria and number of exposed and unexposed Case-control study—For matched studies, give matching criteria and the number of controls per case N/A Variables 7 Clearly define all outcomes, exposures, predictors, potential confounders, and effect modifiers. Give diagnostic criteria, if applicable 30 – 31 Data sources/ measurement 8  For each variable of interest, give sources of data and details of methods of assessment (measurement). Describe comparability of assessment methods if there is more than one group 30 Bias 9 Describe any efforts to address potential sources of bias 30 – 31 Study size 10 Explain how the study size was arrived at 32 – 33 Quantitative variables 11 Explain how quantitative variables were handled in the analyses. If applicable, describe which groupings were chosen and why 30 – 31 Statistical methods 12 (a) Describe all statistical methods, including those used to control for confounding 31 – 32 (b) Describe any methods used to examine subgroups and interactions 31 – 32 (c) Explain how missing data were addressed 32 – 33 (d) Cohort study—If applicable, explain how loss to follow-up was addressed Case-control study—If applicable, explain how matching of cases and controls was addressed Cross-sectional study—If applicable, describe analytical methods taking account of sampling strategy 32 – 33 (e) Describe any sensitivity analyses N/A  173 Results Participants 13 (a) Report numbers of individuals at each stage of study—eg numbers potentially eligible, examined for eligibility, confirmed eligible, included in the study, completing follow-up, and analysed 32 – 33 (b) Give reasons for non-participation at each stage 32 – 33 (c) Consider use of a flow diagram N/A Descriptive data 14 (a) Give characteristics of study participants (eg demographic, clinical, social) and information on exposures and potential confounders 33 – 34 (b) Indicate number of participants with missing data for each variable of interest 32 – 33 (c) Cohort study—Summarise follow-up time (eg, average and total amount) 34 Outcome data 15 Cohort study—Report numbers of outcome events or summary measures over time 36 Case-control study—Report numbers in each exposure category, or summary measures of exposure N/A Cross-sectional study—Report numbers of outcome events or summary measures 34 Main results 16 (a) Give unadjusted estimates and, if applicable, confounder-adjusted estimates and their precision (eg, 95% confidence interval). Make clear which confounders were adjusted for and why they were included 36 – 37 (b) Report category boundaries when continuous variables were categorized N/A (c) If relevant, consider translating estimates of relative risk into absolute risk for a meaningful time period N/A Other analyses 17 Report other analyses done—eg analyses of subgroups and interactions, and sensitivity analyses 37 – 38 Discussion Key results 18 Summarise key results with reference to study objectives 38 – 41 Limitations 19 Discuss limitations of the study, taking into account sources of potential bias or imprecision. Discuss both direction and magnitude of any potential bias 41 – 42 Interpretation 20 Give a cautious overall interpretation of results considering objectives, limitations, multiplicity of analyses, results from similar studies, and other relevant evidence 42 Generalisability 21 Discuss the generalisability (external validity) of the study results 41 – 42 Other information Funding 22 Give the source of funding and the role of the funders for the present study and, if applicable, for the original study on which the present article is based N/A A full-text article describes the development and completion of this reporting guideline (von Elm et al., 2007).   174  Appendix B  Sample R script of analysis from Chapter 2  #Load required packages library(readxl) library(psych) library(janitor) library(fmsb) library(ResourceSelection) library(pROC)  #read table mydata <- read_excel("Desktop/Research/REACH AlphaFIM Data/Thesis Chapter/RExperiment.xlsx") View(mydata)  ###Preparing dataset #remove those who were in hospital at the time of their stroke mydata<-mydata[!(mydata$DaysAdmit<(-1)),] #remove those who were admitted >48 hours after their stroke mydata<-mydata[!(mydata$DaysAdmit>(2)),] #remove those who were admitted for recurrent stroke mydata<-mydata[!(mydata$Recurrent==1),] #removing those who died in hospital mydata<-mydata[!(mydata$Discharge=="deceased"),]  #convert walking scores table(mydata$CanWalk) table(mydata$WalkScore) #if WalkScore was 1 or 2, change CanWalk to 0 mydata$CanWalk[mydata$WalkScore=="1"]<-"0" mydata$CanWalk[mydata$WalkScore=="2"]<-"0" #Adjust WalkScore to 6-point (0-5) scale mydata$WalkScore[mydata$CanWalk=="0"]<-"0" mydata$WalkScore[mydata$WalkScore=="1"]<-"0" mydata$WalkScore[mydata$WalkScore=="2"]<-"0" mydata$WalkScore[mydata$WalkScore=="3"]<-"1" mydata$WalkScore[mydata$WalkScore=="4"]<-"2" mydata$WalkScore[mydata$WalkScore=="5"]<-"3" mydata$WalkScore[mydata$WalkScore=="6"]<-"4" mydata$WalkScore[mydata$WalkScore=="7"]<-"5"  #check variable types str(mydata) #convert appropriate variables to categorical mydata$Discharge<-factor(mydata$Discharge) mydata$CanWalk<-factor(mydata$CanWalk) mydata$HomeDC<-factor(mydata$HomeDC) mydata$Sex<-factor(mydata$Sex)  175 mydata$Recurrent<-factor(mydata$Recurrent) mydata$Side<-factor(mydata$Side) mydata$Cortical<-factor(mydata$Cortical) mydata$StrokeType<-factor(mydata$StrokeType) mydata$tPA<-factor(mydata$tPA) mydata$OtherMedical<-factor(mydata$OtherMedical) mydata$EVT<-factor(mydata$EVT) #convert appropriate variables to numeric mydata$WalkScore<-as.double(mydata$WalkScore) mydata$Ques1b<-as.double(mydata$Ques1b) mydata$Comm1c<-as.double(mydata$Comm1c) mydata$Gaze2<-as.double(mydata$Gaze2) mydata$Visual3<-as.double(mydata$Visual3) mydata$Face4<-as.double(mydata$Face4) mydata$LArm5<-as.double(mydata$LArm5) mydata$RArm5<-as.double(mydata$RArm5) mydata$LLeg6<-as.double(mydata$LLeg6) mydata$RLeg6<-as.double(mydata$RLeg6) mydata$Ataxia7<-as.double(mydata$Ataxia7) mydata$Sens8<-as.double(mydata$Sens8) #convert screening ID to string mydata$ScreeningID<-as.character(mydata$ScreeningID)  #create total NIHSS score - InitialNIHSS mydata$InitialNIHSS<-mydata$LOC1a+mydata$Ques1b+mydata$Comm1c+mydata$Gaze2+mydata$Visual3+   mydata$Face4+mydata$LArm5+mydata$RArm5+mydata$LLeg6+mydata$RLeg6+mydata$Ataxia7+mydata$Sens8+   mydata$Lang9+mydata$Dysarth10+mydata$Neglect11  #create new variable - greatest leg impairment MaxLeg<-pmax(mydata$LLeg6,mydata$RLeg6) mydata$MaxLeg<-MaxLeg  ###Handling Missing Data #quantify amount of missing data #create function to determine percent of missing data in a given column percentmiss=function(x){sum( / length(x) * 100} apply(mydata,2,percentmiss) #apply function to columns(2)  #Create new variable to indicate missing either NIHSS or walking data mydata$missEither<-0 mydata$missEither[$LOC1a)|$Ques1b)|$Comm1c)|$Gaze2)|           $Visual3)|$Face4)|$LArm5)|$RArm5)|           $LLeg6)|$RLeg6)|$Ataxia7)|$Sens8)|           $Lang9)|$Dysarth10)|$Neglect11)|  176           $CanWalk)|$WalkScore)]<-1 mydata$missEither<-factor(mydata$missEither)  #Comparing those missing either and those who had complete data describeBy(mydata,group=mydata$missEither) by(mydata,mydata$missEither,summary) t.test(Age~missEither,data=mydata,paired=FALSE,equal.var=TRUE) prop.test(table(mydata$missEither,mydata$Sex),correct=FALSE) print(chisq.test(table(mydata$missEither,mydata$Side),correct=FALSE)) print(chisq.test(table(mydata$missEither,mydata$Cortical),correct=FALSE)) print(chisq.test(table(mydata$missEither,mydata$StrokeType),correct=FALSE)) print(chisq.test(table(mydata$missEither,mydata$Discharge),correct=FALSE))  #remove observations that are missing any component value of InitialNIHSS or walking data mydata<-mydata[!($LOC1a)|$Ques1b)|$Comm1c)|$Gaze2)|          $Visual3)|$Face4)|$LArm5)|$RArm5)|          $LLeg6)|$RLeg6)|$Ataxia7)|$Sens8)|          $Lang9)|$Dysarth10)|$Neglect11)),] mydata<-mydata[!$CanWalk),] mydata<-mydata[!$WalkScore),]  ###Complete Case Analysis ###Primary objective: Contemporary prevalence estimates of leg and walking impairment #descriptive statistics of whole group describe(mydata) summary(mydata) tabyl(mydata$Sex) tabyl(mydata$Side) tabyl(mydata$StrokeType) tabyl(mydata$Cortical) tabyl(mydata$EVT) tabyl(mydata$tPA) tabyl(mydata$WalkScore) tabyl(mydata$MaxLeg) tabyl(mydata$Discharge) tabyl(mydata$CanWalk)  #descriptive statistics by CanWalk Yes/No describeBy(mydata,group=mydata$CanWalk) by(mydata,mydata$CanWalk,summary) mydata %>% tabyl(CanWalk,Sex) %>% adorn_percentages("row") %>% adorn_pct_formatting(digits=2) %>% adorn_ns() mydata %>% tabyl(CanWalk,Side) %>% adorn_percentages("row") %>% adorn_pct_formatting(digits=2) %>% adorn_ns()  177 mydata %>% tabyl(CanWalk,StrokeType) %>% adorn_percentages("row") %>% adorn_pct_formatting(digits=2) %>% adorn_ns() mydata %>% tabyl(CanWalk,Cortical) %>% adorn_percentages("row") %>% adorn_pct_formatting(digits=2) %>% adorn_ns() mydata %>% tabyl(CanWalk,tPA) %>% adorn_percentages("row") %>% adorn_pct_formatting(digits=2) %>% adorn_ns() mydata %>% tabyl(CanWalk,EVT) %>% adorn_percentages("row") %>% adorn_pct_formatting(digits=2) %>% adorn_ns() mydata %>% tabyl(CanWalk,WalkScore) %>% adorn_percentages("row") %>% adorn_pct_formatting(digits=2) %>% adorn_ns() mydata %>% tabyl(CanWalk,MaxLeg) %>% adorn_percentages("row") %>% adorn_pct_formatting(digits=2) %>% adorn_ns() mydata %>% tabyl(CanWalk,Discharge) %>% adorn_percentages("row") %>% adorn_pct_formatting(digits=2) %>% adorn_ns()  #Comparison of Walkers and Non-Walkers t.test(Age~CanWalk,data=mydata,paired=FALSE,var.equal=TRUE) prop.test(table(mydata$CanWalk,mydata$Sex),correct=FALSE) print(chisq.test(table(mydata$CanWalk,mydata$Side),correct=FALSE)) prop.test(table(mydata$CanWalk,mydata$StrokeType),correct=FALSE) print(chisq.test(table(mydata$CanWalk,mydata$Cortical),correct=FALSE)) wilcox.test(InitialNIHSS~CanWalk,data=mydata) wilcox.test(MaxLeg~CanWalk,data=mydata) print(chisq.test(table(mydata$CanWalk,mydata$MaxLeg),correct=FALSE)) print(chisq.test(table(mydata$CanWalk,mydata$Discharge),correct=FALSE)) wilcox.test(LOS~CanWalk,data=mydata) mydataISCHEMIC<-mydata[!(mydata$StrokeType==1),] prop.test(table(mydataISCHEMIC$CanWalk,mydataISCHEMIC$tPA),correct=FALSE) prop.test(table(mydataISCHEMIC$CanWalk,mydataISCHEMIC$EVT),correct=FALSE)  ###Secondary objective: Predicting Home Discharge #descriptive statistics by HomeDC (discharged home or not) describeBy(mydata,group=mydata$HomeDC) mydata %>% tabyl(HomeDC,Sex) %>% adorn_percentages("row") %>% adorn_pct_formatting(digits=2) %>% adorn_ns() mydata %>% tabyl(HomeDC,Side) %>% adorn_percentages("row") %>% adorn_pct_formatting(digits=2) %>% adorn_ns() mydata %>% tabyl(HomeDC,StrokeType) %>% adorn_percentages("row") %>% adorn_pct_formatting(digits=2) %>% adorn_ns() mydata %>% tabyl(HomeDC,Side) %>% adorn_percentages("row") %>% adorn_pct_formatting(digits=2) %>% adorn_ns() mydata %>% tabyl(HomeDC,Cortical) %>% adorn_percentages("row") %>% adorn_pct_formatting(digits=2) %>% adorn_ns() mydata %>% tabyl(HomeDC,CanWalk) %>% adorn_percentages("row") %>% adorn_pct_formatting(digits=2) %>% adorn_ns() by(mydata,mydata$HomeDC,summary)  #Sample Univariate Logistic Regression code - Age glm(HomeDC~Age,data=mydata,family=binomial)%>%summary() glm(HomeDC~Age,data=mydata,family=binomial)%>%coef()%>%exp()  178 glm(HomeDC~Age,data=mydata,family=binomial)%>%confint()%>%exp() #Assumption 1: All observations are independent (yes, by design) #Assumption 2: Response variable (and residuals) should be matched by an appropriate distribution (binomial in this case) #Assumption 3: All observations are equally influential in determining the trends (no observations overly influential) #Assumption 4: Linear relationship between any continuous independent variables and the logit transformation of the dependent variable hist(mydata$Age) #not overly skewed plot(HomeDC~Age,mydata) with(data=mydata,lines(lowess(HomeDC~Age))) #standard sigmoidal curve glm(HomeDC~Age,data=mydata,family=binomial)%>%plot() #no patterns in residual plots  #Multivariate Logistic Regression analyses #Ordinal Walking multivariate model modelA<-glm(HomeDC~Age+StrokeType+Cortical+MaxLeg+WalkScore,data=mydata,family="binomial") summary(modelA) exp(coef(modelA)) exp(confint(modelA)) NagelkerkeR2(modelA) #0.498 roc(mydata$HomeDC,modelA$fitted.values,plot=TRUE,legacy.axes=TRUE, print.auc=TRUE) #AUC=0.863 (0.83 - 0.90) ci.auc(roc(mydata$HomeDC,modelA$fitted.values,plot=TRUE,legacy.axes=TRUE, print.auc=TRUE)) glm(HomeDC~Age+StrokeType+Cortical+MaxLeg+WalkScore+MaxLeg:WalkScore,data=mydata,family="binomial") %>% summary() #interaction not significant hoslem.test(modelA$y,fitted(modelA)) #x=5.68, df=8 p=0.68, no significant difference between observed and predicted values  #Binary Walking multivariate model modelB<-glm(HomeDC~Age+StrokeType+Cortical+InitialNIHSS+CanWalk,data=mydata,family="binomial") summary(modelB) exp(coef(modelB)) exp(confint(modelB)) NagelkerkeR2(modelB) #0.422 roc(mydata$HomeDC,modelB$fitted.values,plot=TRUE,legacy.axes=TRUE, print.auc=TRUE) #AUC=0.836 (0.80 - 0.87) ci.auc(roc(mydata$HomeDC,modelB$fitted.values,plot=TRUE,legacy.axes=TRUE, print.auc=TRUE)) glm(HomeDC~Age+StrokeType+Cortical+InitialNIHSS+CanWalk+InitialNIHSS:CanWalk,data=mydata,family="binomial") %>% summary() #interaction term not significant hoslem.test(modelB$y,fitted(modelB)) #x=9.88, df=8 p=0.27, no significant difference between observed and predicted values  #ROC Figure rocmodelA<-roc(mydata$HomeDC,modelA$fitted.values)  179 rocmodelB<-roc(mydata$HomeDC,modelB$fitted.values) plot(rocmodelA,print.auc=FALSE,col="black",lty=1,lwd=2,legacy.axes=TRUE,print.auc.y=.35,grid=TRUE) plot(rocmodelB,print.auc=FALSE,col="black",lty=3,lwd=2,legacy.axes=TRUE,print.auc.y=.3,grid=TRUE,add=TRUE) text(0.4,0.4,paste("AUC for Model A:", round(rocmodelA$auc,2))) text(0.4,0.35,paste("AUC for Model B:", round(rocmodelB$auc,2))) legend("bottomright",        legend=c("Model A","Model B"),        col=c("black","black"),        lty=c(1,3),        lwd=c(2,2))  180 Appendix C  Reporting guidelines for Chapter 4 C.1 CONSORT 2010 checklist of information to include when reporting a randomized trial Section/Topic Item No Checklist item Page No(s). Title and abstract  1a Identification as a randomised trial in the title 69 1b Structured summary of trial design, methods, results, and conclusions  No abstract Introduction Background and objectives 2a Scientific background and explanation of rationale 69 – 70 2b Specific objectives or hypotheses 70 Methods Trial design 3a Description of trial design (such as parallel, factorial) including allocation ratio 71 – 72 3b Important changes to methods after trial commencement (such as eligibility criteria), with reasons N/A Participants 4a Eligibility criteria for participants 72 4b Settings and locations where the data were collected 71 Interventions 5 The interventions for each group with sufficient details to allow replication, including how and when they were actually administered 73 – 74 Outcomes 6a Completely defined pre-specified primary and secondary outcome measures, including how and when they were assessed 75  (Appendix E) 6b Any changes to trial outcomes after the trial commenced, with reasons N/A Sample size 7a How sample size was determined 77 7b When applicable, explanation of any interim analyses and stopping guidelines 77 Randomisation:     Sequence generation 8a Method used to generate the random allocation sequence 72 8b Type of randomisation; details of any restriction (such as blocking and block size) 72  Allocation concealment mechanism 9 Mechanism used to implement the random allocation sequence (such as sequentially numbered containers), describing any steps taken to conceal the sequence until interventions were assigned 72  Implementation 10 Who generated the random allocation sequence, who enrolled participants, and who assigned participants to interventions 72 Blinding 11a If done, who was blinded after assignment to interventions (for example, participants, care providers, those assessing outcomes) and how 75 11b If relevant, description of the similarity of interventions 74  181 Statistical methods 12a Statistical methods used to compare groups for primary and secondary outcomes 76 12b Methods for additional analyses, such as subgroup analyses and adjusted analyses 76 – 77 Results Participant flow (a diagram is strongly recommended) 13a For each group, the numbers of participants who were randomly assigned, received intended treatment, and were analysed for the primary outcome 77 – 78 13b For each group, losses and exclusions after randomisation, together with reasons 77 – 78 Recruitment 14a Dates defining the periods of recruitment and follow-up 77 14b Why the trial ended or was stopped 77 Baseline data 15 A table showing baseline demographic and clinical characteristics for each group 79 Numbers analysed 16 For each group, number of participants (denominator) included in each analysis and whether the analysis was by original assigned groups 80 – 83 Outcomes and estimation 17a For each primary and secondary outcome, results for each group, and the estimated effect size and its precision (such as 95% confidence interval) 80 – 82 17b For binary outcomes, presentation of both absolute and relative effect sizes is recommended N/A Ancillary analyses 18 Results of any other analyses performed, including subgroup analyses and adjusted analyses, distinguishing pre-specified from exploratory 82 – 83 Harms 19 All important harms or unintended effects in each group 79 – 80 Discussion Limitations 20 Trial limitations, addressing sources of potential bias, imprecision, and, if relevant, multiplicity of analyses 88 Generalisability 21 Generalisability (external validity, applicability) of the trial findings 83 – 88 Interpretation 22 Interpretation consistent with results, balancing benefits and harms, and considering other relevant evidence 83 – 88 Other information  Registration 23 Registration number and name of trial registry vii Protocol 24 Where the full trial protocol can be accessed, if available 71 Funding 25 Sources of funding and other support (such as supply of drugs), role of funders xx A full-text article describes the development and completion of this reporting guideline (Schulz et al., 2010).    182 C.2 The TIDieR (Template for Intervention Description and Replication) Checklist Item number Item  Where located   Primary paper Other  (details)  BRIEF NAME   1. Provide the name or a phrase that describes the intervention. ______42______ _____________  WHY   2. Describe any rationale, theory, or goal of the elements essential to the intervention. ____69 – 71____ _____________  WHAT   3. Materials: Describe any physical or informational materials used in the intervention, including those provided to participants or used in intervention delivery or in training of intervention providers. Provide information on where the materials can be accessed (e.g. online appendix, URL). _____42______  _Appendix D__ 4. Procedures: Describe each of the procedures, activities, and/or processes used in the intervention, including any enabling or support activities. ___73 – 74____ _Appendix D__  WHO PROVIDED   5. For each category of intervention provider (e.g. psychologist, nursing assistant), describe their expertise, background and any specific training given. ___73 – 74____ _____________  HOW   6. Describe the modes of delivery (e.g. face-to-face or by some other mechanism, such as internet or telephone) of the intervention and whether it was provided individually or in a group. ___73 – 74____ _____________  WHERE    183 7. Describe the type(s) of location(s) where the intervention occurred, including any necessary infrastructure or relevant features. ____72_______ _____________  WHEN and HOW MUCH   8. Describe the number of times the intervention was delivered and over what period of time including the number of sessions, their schedule, and their duration, intensity or dose. ___73 – 74____ _Appendix D__  TAILORING   9. If the intervention was planned to be personalised, titrated or adapted, then describe what, why, when, and how. ___73 – 74___ _Appendix D__  MODIFICATIONS   10. If the intervention was modified during the course of the study, describe the changes (what, why, when, and how). ____N/A_______ _____________  HOW WELL   11. Planned: If intervention adherence or fidelity was assessed, describe how and by whom, and if any strategies were used to maintain or improve fidelity, describe them. ____N/A______ _____________ 12.  Actual: If intervention adherence or fidelity was assessed, describe the extent to which the intervention was delivered as planned. _____77_____ _____________ A full-text article describes the development and completion of this reporting checklist (Hoffmann et al., 2014) .   184 Appendix D  Guidelines for progressing participant in Exoskeleton Group  To standardize the rehabilitation received by participants in the Exoskeleton Group to a degree, the following guidelines to progress a participant are provided:  Initial settings: Walk Mode First Step Step Length Individualized Step Height 0.3-0.5” Lateral Target 2 Side Assisted Bilateral Level of Assistance Adapt  Ideal target settings for hemiparetic stroke: Walk Mode Pro Step+ Step Length Individualized Step Height 0.1” Lateral Target -1, -2 Side Assisted Free leg on strong side Level of Assistance Fixed 0-50 There may be participants who improve rapidly during the course of the intervention (whether due to spontaneous recovery or from functional practice). If a participant is improving faster than expected, i.e. is able to perform longer bouts of walking with the exoskeleton, achieving higher steps, walking with minimal assistance with the device, it is permissible to advance the participant faster than the provided guidelines. Certain progressions can be made simultaneously (e.g. fixed assistance with longer duration of walking).  D.1 Definition of EksoGT settings: Walk mode (how each step is triggered) First Step: Each step manually triggered by therapist by pressing a button Pro Step+: Steps are triggered by exoskeleton user (by weight shift onto stance side, hip flexion moment on swing side) Step Length (length of each assisted step) Step Height (height of foot clearance during swing, in inches) Lateral Target (amount of lateral weight shift, units set by manufacturer) Side Assisted (side of assistance from robotics) Bilateral: Both legs receiving assistance from robotics Free leg: Unilateral assistance from robotics Level of Assistance (amount of assistance received from robotics) Adapt: fluctuating assistance provided by robotics for each step as needed Fixed (0-100): set maximum of assistance provided by robotics for any given step (%)     185 D.2 Progression of settings to create best quality, fully-assisted gait in exoskeleton: This process may be completed as quickly as the first or second session with the participant in the device. Some of these target values may not need to be reached in order to create best quality gait and will also be guided by the visual appearance of the gait cycle.      186 D.3 Algorithm to increase active participation and challenge while in exoskeleton This flow chart should be used once the participant is comfortable walking in the device, requiring minimal verbal and physical guidance to maintain their balance in the device, to shift weight onto each stance limb, and to complete a swing. It may be very challenging for a participant to reach the final settings.       187 D.4 Progression of time spent walking while standing in exoskeleton As the participant becomes more familiar with walking in the exoskeleton, it is expected that they will be able to tolerate longer bouts of walking during each exoskeleton session. While the guidelines for progressing the exoskeleton device settings are more standardized by order, progression of time spent walking versus standing will be dependent on the participant’s endurance, presence of pain or discomfort, as well as therapist fatigue (if participant requires much assistance). The following are rough guidelines for progressing the proportion of time spent walking during each week of exoskeleton training.   Training time guidelines Week 1 (i.e., first 2 – 3 sessions) Require 30 minutes of upright time in exoskeleton, no set requirement for time in walking (expect approximately 10 minutes) Aim for at least 250 steps per session. Week 2 (i.e., 4th session and on) Require 15 minutes of walking time, of 30 minutes of upright time Aim for 400 steps per session. Week 3 (i.e., 7th session and on) Require 20 minutes of walking time, of 30 minutes of upright time Aim for 550 steps per session. Week 4 (i.e., beyond 9 sessions) Require 25 minutes of walking time, of 30 minutes of upright time Aim for 700+ steps per session  These guidelines may not always be reached if a participant experiences significant fatigue or if the challenge of the device settings (e.g. reduced assistance) requires more breaks for the participant. However, if the participant were to remain at the same settings (unable to progress the challenge of the device walking) then it is expected that they instead are making improvements in their walking tolerance. If a participant progresses to independent walking, these guidelines should still be followed even without wearing the device.  At times, the exoskeleton therapist may opt to increase the challenge and demand on the participant to improve the quality of their walking in the device. For example, standing and stepping exercises. In this case, the walking time may not be reached but the 30 minutes of upright time is still required each session. The clinical decision to deviate from the guideline should be purposeful, in that any other exercises done in the device are expected to help the participant progress in their exoskeleton training for their next session (either time, quality, or challenge).     188 D.5 Algorithm to continue or discontinue daily exoskeleton training To ensure a level of standardization for exoskeleton-based rehabilitation in the Exoskeleton Group, the following flow-chart will be used to decide if an individual has advanced to a point where the exoskeleton should be discontinued. In essence, the Exoskeleton should be used for therapy as long as it provides an advantage to the therapist to get the participant up in walking, whether it is for time, speed, assistance, or quality.  Discontinuation of daily exoskeleton training indicates that the device no longer needs to be worn 75% of the time for gait training. Instead, gait training can be a combination of unassisted overground walking (for functional task practice, speed training, etc.) and exoskeleton training (for quality of movement, speed training, etc.), based on clinical judgement.     189 Appendix E  Data collection schedule and outcome measures used in Chapter 4 E.1 Schedule of data collection Study Procedures Screening  Baseline evaluation  Post-intervention evaluation 6-month evaluation Informed consent +    Inclusion/exclusion criteria +    Demographics  +   Randomization  +   Primary outcome measure     Functional Ambulation Category   + + + Secondary outcome measures     Impairment     Fugl-Meyer Assessment (Lower extremity)  + + + Functional     Gait speed over 5 metres   (+) (+) 6-Minute Walk test   (+) (+) Berg Balance Scale  + + + Step counts (during therapy)   +*  Days to unassisted ambulation   +*  Mood     Patient Health Questionnaire  + + + Cognition     Montreal Cognitive Assessment  + + + Quality of life     36-Item Short Form Survey  + + + Adverse events screen   +* + ( ) Parentheses indicate that the outcome will be assessed if the participant is able to walk without physical assistance *indicates that the measure will be taken or monitored throughout the intervention period  E.2 Outcome measures Primary Outcome Measure The primary outcome was walking ability, measured using the Functional Ambulation Category (FAC) (Holden et al., 1984). This is a 6-item scale designed to classify the level of physical support required by subjects in order to walk safely over 10 feet, extending from 0 (unable to walk without the assistance of two people) to 5 (independent walking overground on various surfaces, including stairs). It has been shown to have good test-retest reliability and validity in the hemiparetic stroke population (Mehrholz et al., 2007). The FAC is also responsive to change within the first four weeks post-stroke as well as between four weeks and six months post-stroke (Mehrholz et al. 2007). Unlike other walking measures that consider speed or  190 distance, a value is assigned for the FAC even if the patient is not yet independent in walking. Every increment in the FAC is considered a clinically meaningful difference.   Secondary Outcome Measures • Physical impairment: The Fugl-Meyer assessment is a measure of body function impairment after stroke (Fugl-Meyer et al., 1975) and includes a subscale of the upper and lower extremity. The assessment of the lower extremity (FMA-LE) tests reflexes, synergistic movement, and coordination and is scored from 0 to 34. The FMA-LE alone has been demonstrated to be valid and reliable for measuring lower extremity motor impairment in hemiparetic stroke patients (Park and Choi, 2014). The only available MCID for the FMA-LE is in chronic stroke, which is 6 points (Pandian et al., 2016).   • Self-paced walking speed: The timed 5-Metre Walk Test (5MWT) to measure walking speed is a reliable, valid, and responsive measure for subacute stroke (Fulk and Echternach, 2008; Salbach et al., 2001). Gait speed is a viable outcome for documenting rehabilitation progress and longitudinal change in walking disability post-stroke (Salbach et al., 2001). The MCID for walking speed is 0.16 m/s for patients with subacute stroke (Tilson et al., 2010).   • Walking capacity: The 6-Minute Walk Test (6MWT) is an outcome measure that records the distance walked in six minutes, with or without breaks. It reflects the typical demands of everyday functional mobility. While the FAC captures information regarding the amount of assistance required for walking, the 5MWT and 6MWT quantify the nature of independent walking. The 6MWT is reliable and valid for use with the subacute stroke population undergoing rehabilitation (Fulk and Echternach, 2008). The MCID is 61 m for patients with subacute stroke (Perera et al., 2006).   • Balance function: The Berg Balance Scale (BBS) (Berg et al., 1989) consists of 14 balance tasks, each rated from 0 to 4 for a total score ranging from 0 to 56, and is a valid measure of balance in stroke. It has high intra- and inter-rater reliability, as well as excellent sensitivity to change (Blum and Korner-Bitensky, 2008). The tasks range in difficulty from static sitting to standing on one foot and turning on the spot. The minimal detectable change (MDC) for the BBS in acute and subacute stroke is 6 points (Stevenson, 2001).  • Step counts and standing time: The number of steps during physical therapy sessions and time spent in standing was monitored twice a week to provide an indication of the walking activities of participants. Participants in either group were monitored using the activPAL activity monitor, which captures time spent in standing and sitting/supine, as well as step counts. It has been shown to be a valid measure of standing time in older adults with impaired walking ability (Taraldsen et al., 2011). Participants were monitored for the duration of the intervention (until discharge, or a maximum of eight weeks).   • Days to unassisted walking: The number of days from enrolment to first achievement of unassisted walking (FAC ≥ 3, i.e., no contact from a therapist) was measured during the intervention period, by therapist report or by the discharge assessor.   191 • Depression: The Patient Health Questionnaire (PHQ-9) is a self-reported outcome measure that screens participants for depressive symptoms over the last two weeks and is reliable and valid for use in stroke (Kroenke et al., 2001; de Man-van Ginkel et al., 2012). A previous study in subacute stroke identified a cut-off score of 10 to indicate a potential diagnosis of depression (de Man-van Ginkel et al., 2012). No research has identified a MCID in the stroke population.   • Cognition: The Montreal Cognitive Assessment (MoCA) (Nasreddine et al., 2005) is a screening tool to evaluate cognitive impairment. It assesses attention and concentration, executive function, memory, language, visuoconstructional skills, conceptual thinking, calculations, and orientation. The assessment is scored from 0 to 30, with scores less than 26 points indicating presence of cognitive impairment. The MoCA is valid and reliable for use in the subacute stroke population (Toglia et al., 2011). The MCID is 1.22 – 2.15, investigated in a chronic stroke population (Wu et al., 2019).  • Health-related quality of life: The Medical Outcomes Study 36-Item Short Form Health Survey (SF-36) is a multi-purpose health survey with 36 questions on functional health and well-being (Hays and Morales, 2001). The SF-36 has been shown to be sensitive to changes in the inpatient stroke setting (Hopman and Verner, 2003), whereas some other stroke-specific scales (e.g. Stroke Impact Scale) require the subject to be in a home setting. The items on the questionnaire are pooled and standardized to provide a physical (SF-36-Physical) and mental (SF-36-Mental) component summary score, centred around a mean of 50 (standard deviation: 10) for the general population (Laucis et al., 2014). No studies have specifically determined a MCID value in stroke for either the physical or mental component summary scores; a 5-point MCID has been used in generalized studies (Luo et al., 2015).     192  Appendix F  Data analysis script, and results from per-protocol and sensitivity analyses F.1 Sample R script of analysis #load required packages library(readxl) library(psych) library(rstatix) library(janitor) library(car) library(ggplot2) library(effsize)  ###Data preparation #read data file mydata <- read_excel("Desktop/Research/_ExStRA Study/DATA ENTRY/ExStRA_Rdata.xlsx", na="NA") View(mydata)  #check variable types str(mydata) #convert appropriate variables into categorical mydata$Group<-factor(mydata$Group) #0 is Usual Care, 1 is Intervention mydata$GroupPP2<-factor(mydata$GroupPP2) mydata$BBSAdjust<-factor(mydata$BBSAdjust) #0 is <12, 1 is 12 or higher mydata$Sex<-factor(mydata$Sex) #0 is female, 1 is male mydata$StrokeSide<-factor(mydata$StrokeSide) #0 is right, 1 is left mydata$StrokeType<-factor(mydata$StrokeType) #0 is ischemic, 1 is hemorrhagic mydata$Recurrent<-factor(mydata$Recurrent) #0 is first stroke, 1 is recurrent stroke  ###Descriptive statistics (Table 1) describeBy(mydata,group=mydata$Group) #to get mean, SD by(mydata,mydata$Group,summary) #median, IQR #get counts and percentages for specified variable by intervention group mydata %>% tabyl(Group,Sex) %>% adorn_percentages("row") %>% adorn_pct_formatting(digits=2) %>% adorn_ns() mydata %>% tabyl(Group,StrokeSide) %>% adorn_percentages("row") %>% adorn_pct_formatting(digits=2) %>% adorn_ns() mydata %>% tabyl(Group,StrokeType) %>% adorn_percentages("row") %>% adorn_pct_formatting(digits=2) %>% adorn_ns() mydata %>% tabyl(Group,Recurrent) %>% adorn_percentages("row") %>% adorn_pct_formatting(digits=2) %>% adorn_ns()  #Checking normality of data (examples) mydata %>% group_by(Group) %>% shapiro_test(Velocity2allLast) #Usual Care: p=0.11, Exo: p=0.01 mydata %>% group_by(Group) %>% shapiro_test(Velocity3allLast) #Usual Care: p= 193 0.38, Exo: p=0.03 mydata %>% group_by(Group) %>% shapiro_test(SixMWT2allLast) #Usual Care: p=0.14, Exo: p=0.01 mydata %>% group_by(Group) %>% shapiro_test(SixMWT3allLast) #Usual Care: p=0.34, Exo: p=0.02 mydata %>% group_by(Group) %>% shapiro_test(DaysIndep) #Usual Care: p=0.14, Exo: p=0.13  ###Primary Outcome analysis: Walking ability (FAC) #Comparison of change score from baseline, at discharge mydata$FACDiff<-(mydata$FAC2Last-mydata$FAC1) wilcox.test(mydata$FACDiff~mydata$Group) #Comparison of change score from baseline, at 6-months mydata$FACDiff2<-(mydata$FAC3Last-mydata$FAC1) wilcox.test(mydata$FACDiff2~mydata$Group)  ###Secondary Outcomes analysis: Pairwise comparisons of walking speed, 6MWT, days to independence #Walking speed - Mann-Whitney U test wilcox.test(Velocity2allLast~Group,data=mydata) wilcox.test(Velocity3allLast~Group,data=mydata) #6MWT - Mann-Whitney U test wilcox.test(SixMWT2allLast~Group,data=mydata) wilcox.test(SixMWT3allLast~Group,data=mydata) #Days to independence - independent samples t-test leveneTest(DaysIndep~Group,data=mydata) #assumption of homogeneity of variances, p=0.16 t.test(DaysIndep~Group,data=mydata,paired=FALSE,var.equal=TRUE) #effect sizes cohen.d(mydata$Velocity2allLast,mydata$Group,na.rm=TRUE,hedges.correction=TRUE) cohen.d(mydata$Velocity3allLast,mydata$Group,na.rm=TRUE,hedges.correction=TRUE) cohen.d(mydata$SixMWT2allLast,mydata$Group,na.rm=TRUE,hedges.correction=TRUE) cohen.d(mydata$SixMWT3allLast,mydata$Group,na.rm=TRUE,hedges.correction=TRUE) cohen.d(mydata$DaysIndep,mydata$Group,na.rm=TRUE,hedges.correction=TRUE)  ###Secondary Outcomes analysis: ANCOVA of impairment, balance, etc. (using baseline as covariate) #Sample analysis script: Fugl-Meyer Assessment - Lower Extremity (at discharge is dependent variable) #assumption checking #visual inspection of outcomes ggplot(mydata,aes(y=FM2Last,x=FM1,group=Group))+geom_point()+geom_smooth(method="lm") #relationship appears linear, slopes broadly similar ggplot(mydata,aes(y=FM2Last,x=Group))+geom_boxplot() #no obvious non-normality, variability approximately equal #assumption: homogeneity of slopes anova(lm(FM2Last~FM1*Group,data=mydata)) #non-significant interaction term #assumption: normality of residuals  194 modelFM<-lm(FM2Last~FM1+Group,data=mydata) plot(modelFM)  #plot 1: visual inspection of residual vs fitted plot, no wedge-shaped pattern (non-homogeneity of variance) #plot 2: visual inspection of Q-Q normal plot, no evidence of non-normality #plot 3 and 4: no obviously large residual or leverage values, no outliers near Cook's D contour, i.e., no overly influential observations #ANCOVA results summary(modelFM) #difference between groups and p-value confint(modelFM) #confidence intervals  #Sample analysis script: Fugl-Meyer Assessment - Lower Extremity (at 6 months is dependent variable) #assumption checking #visual inspection of outcomes ggplot(mydata,aes(y=FM3Last,x=FM1,group=Group))+geom_point()+geom_smooth(method="lm") #relationship appears linear, slopes broadly similar ggplot(mydata,aes(y=FM3Last,x=Group))+geom_boxplot() #no obvious non-normality, variability approximately equal #assumption: homogeneity of slopes anova(lm(FM3Last~FM1*Group,data=mydata)) #non-significant interaction term #assumption: normality of residuals modelFM2<-lm(FM3Last~FM1+Group,data=mydata) plot(modelFM2)  #plot 1: visual inspection of residual vs fitted plot, no wedge-shaped pattern (non-homogeneity of variance) #plot 2: visual inspection of Q-Q normal plot, negligible degree of non-normality #plot 3 and 4: no obviously large residual or leverage values, no outliers near Cook's D contour, i.e., no overly influential observations #ANCOVA results summary(modelFM2) #difference between groups and p-value confint(modelFM2) #confidence intervals #Sample Sensitivity analysis lm(FM3Minus25~FM1+Group,data=mydata)%>%summary() #only interested in a significant group effect lm(FM3Plus25~FM1+Group,data=mydata)%>%summary() #only interested in a significant group effect  ###Per-protocol analysis (As above, except using GroupPP2 for grouping factor) #Descriptive statistics, by per-protocol grouping describeBy(mydata,group=mydata$GroupPP2) #to get mean, SD by(mydata,mydata$GroupPP2,summary) #median, IQR  #Primary outcome: Comparison of change score from baseline, at discharge wilcox.test(mydata$FACDiff~mydata$GroupPP2) #at 6-months wilcox.test(mydata$FACDiff2~mydata$GroupPP2)   195 #Secondary Outcomes analysis: Pairwise comparisons of walking speed (example) #Walking speed - Mann-Whitney U test wilcox.test(Velocity2allLast~GroupPP2,data=mydata) wilcox.test(Velocity3allLast~GroupPP2,data=mydata) cohen.d(mydata$Velocity2allLast,mydata$GroupPP2,na.rm=TRUE,hedges.correction=TRUE) cohen.d(mydata$Velocity3allLast,mydata$GroupPP2,na.rm=TRUE,hedges.correction=TRUE)  #Secondary outcomes analysis: ANCOVA of impairment (using baseline as covariate) #Example analysis script: Fugl-Meyer Assessment - Lower Extremity (at discharge is dependent variable) #assumption checking #visual inspection of outcomes ggplot(mydata,aes(y=FM2Last,x=FM1,group=GroupPP2))+geom_point()+geom_smooth(method="lm") #relationship appears linear, slopes broadly similar ggplot(mydata,aes(y=FM2Last,x=GroupPP2))+geom_boxplot() #no obvious non-normality, variability approximately equal #assumption: homogeneity of slopes anova(lm(FM2Last~FM1*GroupPP2,data=mydata)) #non-significant interaction term #assumption: normality of residuals modelFMPP<-lm(FM2~FM1+GroupPP2,data=mydata) plot(modelFMPP)  #plot 1: visual inspection of residual vs fitted plot, slight wedge-shaped (non-homogeneity of variance) #plot 2: visual inspection of Q-Q normal plot, no evidence of non-normality #plot 3 and 4: no obviously large residual or leverage values, no outliers near Cook's D contour, i.e., no overly influential observations #ANCOVA results summary(modelFMPP) #difference between groups and p-value confint(modelFMPP) #confidence intervals    196 F.2 Sensitivity analysis Participants that did not attend the 6-month assessment were given their last observation carried forward. The sensitivity analysis, which was the last observation carried forward +/- 25%, served to determine if any significant findings were robust to systematic loss to follow-up (i.e., verifying that participants’ absenteeism was not related to their outcomes). Since no significant differences were found in the primary analysis, this sensitivity analysis was therefore only conducted and provided here as a reference.  Primary outcome: (Rather than 25%, we used +/- 1 on the FAC scale) FAC Exoskeleton n = 19 Median (IQR) Usual Care n = 17 Median (IQR) p-value Baseline 0 (0 – 1) 0 (0 – 1)  6-month 4 (2 – 5) 4 (3 – 5)  Change from baseline 3 (2 – 4) 3 (3 – 4) 0.65a LOCF -1 4 (1 – 5) 4 (3 – 5)  Change from baseline 3 (1 – 4) 3 (2 – 4) 0.83a LOCF +1 4 (2.5 – 5) 4 (4 – 5)  Change from baseline 3 (2 – 4) 4 (3 – 4) 0.54a Abbreviations: FAC, Functional Ambulation Category; IQR, interquartile range aAnalyzed using Mann-Whitney U test  Secondary walking outcomes: Variable Exoskeleton n = 19 Mean (SD) Usual Care n = 17 Mean (SD) p-value Gait speed, m/s 6-month follow-up 0.52 (0.5) 0.42 (0.3) 0.74a -25% 0.51 (0.5) 0.40 (0.3) 0.74a +25% 0.52 (0.5) 0.44 (0.3) 0.84a 6MWT, m 6-month follow-up 164.5 (152.8) 123.4 (90.1) 0.60a -25% 163.2 (153.6) 117.8 (89.0) 0.58a +25% 165.9 (152.5) 128.9 (93.3) 0.60a Abbreviations: 6MWT, 6-Minute Walk Test; SD, standard deviation aAnalyzed using Mann-Whitney U test     197 Secondary outcomes of impairment, balance, mood, cognition, and quality of life: Variable Exoskeleton n = 19 Mean (SD) Usual Care n = 17 Mean (SD) Group difference (95% CI) F-statistic p-value FMA-Lower Baseline 17.3 (6.6) 17.5 (7.0)    6-month  20.2 (8.9) 19.2 (8.2) 1.2  (-3.7 – 6.1) F(1,33) = 0.24 0.63 6-month  -25% 18.5 (9.0) 17.7 (7.4) 0.9 (-4.0 – 5.9) F(1,33) = 0.15 0.71 6-month +25% 21.9 (9.5) 20.3 (9.2) 1.7 (-3.7 – 7.1) F(1,33) = 0.42 0.52 BBS Baseline 15.3 (10.0) 19.2 (15.4)    6-month  40.3 (14.3) 43.0 (15.6) -0.5  (-9.0 – 7.9) F(1,33) = 0.02 0.90 6-month -25%  38.0 (15.8) 40.9 (13.9) -0.9 (-10.2 – 8.4) F(1,33) = 0.04 0.84 6-month  +25% 42.5 (13.8) 44.4 (13.9) 0.2 (-8.1 – 8.6) F(1,33) = 0.003 0.95 PHQ-9 Baseline 7.2 (4.2) 7.7 (6.4)    6-month  5.1 (4.0) 6.8 (6.5) -1.4  (-3.6 – 0.8) F(1,33) = 1.599 0.22 6-month  -25% 4.7 (4.0) 6.1 (5.3) -1.0  (-3.2 – 1.1) F(1,33) = 0.95 0.34 6-month  +25% 5.4 (4.1) 7.4 (7.3) -1.6  (-4.0 – 0.8) F(1,33) = 1.84 0.19 MoCAa Baseline 22.4 (4.3) 23.5 (5.0)    6-month  24.6 (4.8) 25.1 (4.9) 0.4  (-1.6 – 2.4) F(1,29) = 0.19 0.67 6-month  -25% 22.8 (6.3) 23.7 (5.9) 0.2  (-3.0 – 3.3) F(1,29) = 0.01 0.91 6-month  +25% 26.1 (4.1) 26.1 (4.5) 0.7  (-1.4 – 2.8) F(1,29) = 0.49 0.49 SF-36-Physicalb Baseline 30.2 (8.9) 28.2 (6.5)    6-month  33.5 (9.9) 30.8 (10.5) 1.9  (-5.1 – 8.8) F(1,31) = 0.30 0.59 6-month  -25% 31.1 (9.8) 29.2 (11.6) 1.2 (-6.3 – 8.7) F(1,31) = 0.11 0.74 6-month  +25% 35.8 (11.6) 32.4 (10.1) 2.5 (-4.8 – 9.8) F(1,31) = 0.49 0.49  198 Variable Exoskeleton n = 19 Mean (SD) Usual Care n = 17 Mean (SD) Group difference (95% CI) F-statistic p-value SF-36-Mentalb Baseline 51.0 (10.4) 49.4 (12.4)    6-month  50.1 (12.5) 52.4 (13.2) -3.2  (-11.0 – 4.5) F(1,31) = 0.72 0.40 6-month  -25% 46.7 (13.0) 49.4 (14.0) -3.7  (-11.7 – 4.2) F(1,31) = 0.92 0.34 6-month  +25% 53.5 (14.6) 55.5 (14.9) -2.7  (-12.4 – 6.9) F(1,31) = 0.34 0.57 Abbreviations: BBS, Berg Balance Scale; FMA-LE, Lower extremity component of Fugl-Meyer Assessment; MoCA, Montreal Cognitive Assessment; PHQ-9, Patient Health Questionnaire; SD, standard deviation; SF-36-Mental, Mental Component of 36-Item Short Form Survey; SF-36-Physical, Physical Component of 36-Item Short Form Survey aExoskeleton n = 17; Usual Care n = 15 bExoskeleton n = 18; Usual Care n = 16  F.3 Per-protocol analysis For the per-protocol analysis, any participants in the Exoskeleton group refused further training using the device, who also received <70% of their planned exoskeleton sessions (based on number of days in the trial until the discharge assessment, assuming three exoskeleton sessions per week in that time). Five participants met this criterion and were analyzed instead as part of the Usual Care group.  Secondary outcomes of impairment, balance, mood, cognition, and quality of life: Variable Exoskeleton n = 14 Mean (SD) Usual Care n = 22 Mean (SD) Group difference (95% CI) F-statistic p-value FMA-Lower Baseline 16.6 (6.7) 17.9 (6.8)    Discharge 23.7 (6.5) 20.8 (6.4) 3.9  (1.3 – 6.6) F(1,33) = 9.33 0.004 6-month  24.0 (6.9) 22.0 (4.7) 2.7 (-0.2 – 5.7) F(1,33) = 3.5 0.07 BBS Baseline 16.7 (10.3) 17.4 (14.4)    Discharge 40.9 (14.6) 34.8 (16.7) 6.5 (-6.9 – 9.7) F(1,33) = 1.95 0.17 6-month  45.6 (11.7) 39.1 (14.7) 6.9 (-1.3 – 15.1) F(1,33) = 2.94 0.10  199 Variable Exoskeleton n = 14 Mean (SD) Usual Care n = 22 Mean (SD) Group difference (95% CI) F-statistic p-value PHQ-9 Baseline 7.0 (4.8) 7.6 (5.7)    Discharge 3.5 (3.3) 6.0 (6.6) -2.1 (-5.0 – 0.9) F(1,33) = 2.05 0.16 6-month  4.7 (4.3) 6.6 (5.8) -1.4 (-3.7 – 0.9) F(1,33) = 1.61 0.21 MoCAa Baseline 23.5 (3.5) 22.4 (5.2)    Discharge 26.7 (3.4) 23.5 (5.7) 2.1 (0.6 – 3.7) F(1,29) = 7.96 0.009 6-month  26.2 (3.0) 23.9 (5.6) 1.4 (-0.6 – 3.4) F(1,29) = 2.07 0.16 SF-36-Physicalb Baseline 29.9 (10.0) 28.9 (6.4)    Discharge 30.3 (9.3) 30.4 (9.4) -0.6 (-7.0 – 5.8) F(1,31) = 0.04 0.85 6-month  31.7 (10.4) 33.0 (10.0) 0.9 (-6.2 – 8.1) F(1,31) = 0.07 0.79 SF-36-Mentalb Baseline 51.3 (11.9) 49.6 (11.1)    Discharge 54.1 (11.7) 51.6 (14.6) 1.3 (-7.9 – 9.7) F(1,31) = 0.10 0.75 6-month  50.8 (11.9) 51.5 (13.4) -1.7 (-9.7 – 6.4) F(1,31) = 0.18 0.67 Abbreviations: BBS, Berg Balance Scale; CI, confidence interval; FMA-LE, Lower extremity component of Fugl-Meyer Assessment; MoCA, Montreal Cognitive Assessment; PHQ-9, Patient Health Questionnaire; SD, standard deviation; SF-36-Mental, Mental Component of 36-Item Short Form Survey; SF-36-Physical, Physical Component of 36-Item Short Form Survey aExoskeleton n = 17; Usual Care n = 15 bExoskeleton n = 18; Usual Care n = 16    200 Appendix G  Standards for Reporting Qualitative Research (SRQR) Title and abstract Page No(s).  Title - Concise description of the nature and topic of the study Identifying the study as qualitative or indicating the approach (e.g., ethnography, grounded theory) or data collection methods (e.g., interview, focus group) is recommended 91  Abstract - Summary of key elements of the study using the abstract format of the intended publication; typically includes background, purpose, methods, results, and conclusions No abstract    Introduction   Problem formulation - Description and significance of the problem/phenomenon studied; review of relevant theory and empirical work; problem statement 91 – 92  Purpose or research question - Purpose of the study and specific objectives or questions 92 – 93    Methods   Qualitative approach and research paradigm - Qualitative approach (e.g., ethnography, grounded theory, case study, phenomenology, narrative research) and guiding theory if appropriate; identifying the research paradigm (e.g., postpositivist, constructivist/ interpretivist) is also recommended; rationale 93  Researcher characteristics and reflexivity - Researchers’ characteristics that may influence the research, including personal attributes, qualifications/experience, relationship with participants, assumptions, and/or presuppositions; potential or actual interaction between researchers’ characteristics and the research questions, approach, methods, results, and/or transferability 96  Context - Setting/site and salient contextual factors; rationale 94 – 95  Sampling strategy - How and why research participants, documents, or events were selected; criteria for deciding when no further sampling was necessary (e.g., sampling saturation); rationale 93 – 94  Ethical issues pertaining to human subjects - Documentation of approval by an appropriate ethics review board and participant consent, or explanation for lack thereof; other confidentiality and data security issues 93  Data collection methods - Types of data collected; details of data collection procedures including (as appropriate) start and stop dates of data collection and analysis, iterative process, triangulation of sources/methods, and modification of procedures in response to evolving study findings; rationale 94 – 95  201  Data collection instruments and technologies - Description of instruments (e.g., interview guides, questionnaires) and devices (e.g., audio recorders) used for data collection; if/how the instrument(s) changed over the course of the study 94  Units of study - Number and relevant characteristics of participants, documents, or events included in the study; level of participation (could be reported in results) 97 – 98  Data processing - Methods for processing data prior to and during analysis, including transcription, data entry, data management and security, verification of data integrity, data coding, and anonymization/de-identification of excerpts 94 – 95  Data analysis - Process by which inferences, themes, etc., were identified and developed, including the researchers involved in data analysis; usually references a specific paradigm or approach; rationale 95 – 96  Techniques to enhance trustworthiness - Techniques to enhance trustworthiness and credibility of data analysis (e.g., member checking, audit trail, triangulation); rationale 96 – 97    Results/findings   Synthesis and interpretation - Main findings (e.g., interpretations, inferences, and themes); might include development of a theory or model, or integration with prior research or theory 98 – 107  Links to empirical data - Evidence (e.g., quotes, field notes, text excerpts, photographs) to substantiate analytic findings 99 – 107    Discussion   Integration with prior work, implications, transferability, and contribution(s) to the field - Short summary of main findings; explanation of how findings and conclusions connect to, support, elaborate on, or challenge conclusions of earlier scholarship; discussion of scope of application/generalizability; identification of unique contribution(s) to scholarship in a discipline or field 107 – 109  Limitations - Trustworthiness and limitations of findings 109 – 110    Other   Conflicts of interest - Potential sources of influence or perceived influence on study conduct and conclusions; how these were managed N/A  Funding - Sources of funding and other support; role of funders in data collection, interpretation, and reporting xx A full-text article describes the development and completion of this reporting guideline (O’Brien et al., 2014).   202 Appendix H  Interview guides H.1 Patient Interview Guide Use of a robotic exoskeleton to promote walking recovery after stroke  [ExStRA Study]   How did you feel about your walking ability when you first came to GF Strong?  How do you feel about your walking ability now? (e.g., confidence, fear of falling, endurance)  What training did you receive on the Exoskeleton? What did you like about using the exoskeleton for your stroke rehabilitation? What didn’t you like about using the exoskeleton for your stroke rehabilitation?  How were you progressed on the Exoskeleton? Prompt: What did you like about this progression? Prompt: How could it be improved?  What do you see as barriers to using the exoskeleton?  Prompt: with the technology (e.g., ease of use, noise, comfort) Prompt: with the therapist (e.g., difficulties setting up, confidence in using device, attitude) Prompt: with the current health care system (e.g., space, scheduling within the hour, access to equipment) Prompt: yourself (client) (e.g., fatigue, pain, perceived assumptions and stereotypes, expectations)  What was easy about using the exoskeleton? What are the benefits to using the exoskeleton for stroke rehabilitation? (may query similar to 2)  What changes do you think should be made in the way the Exoskeleton was used in this study?  What are your thoughts on using the Exoskeleton beyond this trial, for other stroke patients?    203 H.2 Physical Therapist Interview Guide Use of a robotic exoskeleton to promote walking recovery after stroke  [ExStRA Study]   How long have you been working in stroke rehabilitation?  What training have you received to use the Exoskeleton?  What did you think about the training you received? Prompt: How could it be improved? Prompt: What did you like about it?  Since you’ve been trained, what opportunities have you had to use the Exoskeleton with patients? Prompt: What type of patients (diagnosis) did you use it with?  What has been your experience using the device in your regular practice?  What has been your experience using the device in the ExStRA study?  What are the barriers to using the exoskeleton for stroke rehabilitation? Prompt: with the technology (e.g., ease of use, noise, quality of movement) Prompt: with the patient (e.g., fatigue, pain, perceived assumptions and stereotypes) Prompt: with the current health care system logistics (e.g., space, scheduling within the hour, access to equipment) Prompt: yourself (therapist) (e.g., your training, your confidence in using the device, your own perceptions of the potential benefits)  What are the benefits to using the exoskeleton for stroke rehabilitation? (may query similar to 4a)  How do you think using technology like the exoskeleton affects patient motivation, if at all?  How would you describe the ideal stroke patient for whom you think the Exoskeleton might be most suitable for? Why? Prompt: What level of functioning would they be? (Consider cognitive and physical) Prompt: At what point in their rehabilitation should they use it?  Describe how you have progressed the patient on the Exoskeleton? Prompt:  What did you like about it? Prompt:  How could it be improved? Prompt:  How many sessions does it take to see progression / improvement?  What changes do you think should be made in the way the Exoskeleton was used in this study?  204  How has exoskeleton-use affected how your work is organized? Prompt:  Is there sufficient support from your work setting to implementing exoskeleton training?  What are your thoughts on using the Exoskeleton beyond this trial, and in your everyday practice? Prompt: Would you recommend exoskeleton device training to your clinical colleagues? All, or some? Prompt: Would you recommend exoskeleton training to your patients?  Is there anything else you would like to add?    205 Appendix I  Study findings from the ExStRA interviews (1-page summary for participants) After running for several years at GF Strong Rehab Hospital (Vancouver), Glenrose Rehab Hospital (Edmonton), and Parkwood Institute (London), recruitment for the ExStRA study has finished. In this clinical trial, participants were randomly assigned to receive an exoskeleton-based intervention or standard physiotherapy during their rehab stay. Participants who received the exoskeleton protocol were also invited to participate in one-on-one interviews to explore their experience and perception of using the robotic exoskeleton for stroke rehabilitation. Therapists who provided the exoskeleton training were interviewed as well. The dialogue from these interviews was analyzed for meaningful patterns and insight, to help guide the future use of this technology within stroke rehabilitation. The findings are organized into three overarching topics, summarized below.  THEME 1: Getting into the swing of things – experiential learning and growth This theme is centered around the learning experience that participants with stroke and therapists had with using an exoskeleton. Many participants found that there was an initial ‘hump’ to get over when beginning to use the exoskeleton, for both patients and therapists alike. This often took the form of confusion, difficulty, or fatigue when trying to move alongside the demands of exoskeleton device. Some participants with stroke suggested more verbal and visual instruction would have helped to reduce the initial difficulty of ‘figuring out’ the machine.  However, this theme also captures the shared experience of progression and proficiency with exoskeleton-based walking. Many participants found that they became more confident and skilled at using the exoskeleton, that only came about with repeated practice. There was also an element of active self-improvement, where participants described applying elements of exoskeleton training to their self-practice of walking outside of therapy.   THEME 2: Balancing perspectives towards exoskeleton training – strengths and limitations  This second theme revolves around participants’ attitudes towards the exoskeleton device, the training itself, as well as perceived benefits offered by using the exoskeleton for stroke rehabilitation. Almost all participants felt positively about having used the exoskeleton, expressing appreciation for the opportunity to stand and practice walking at higher duration and intensity. Additionally, participants described the exoskeleton training as a chance for their body to feel what walking is again, after having a stroke. Many participants with stroke attributed a large part of their walking recovery specifically to having used the exoskeleton, which was reinforced by therapist opinion.  This praise for the exoskeleton is not without criticism, and many participants had suggestions for improving the device or its use. From device fit to function, there were areas identified that could be revamped for better usability; for example, improving the feedback mechanisms of the exoskeleton could streamline the training experience. Furthermore, it was identified that the therapist could impact how the exoskeleton training progressed. Depending on therapist confidence and how they directed the machine, participants felt that it could hinder the walking practice.   THEME 3: Future directions and recommendations  The final theme surrounds the future of using an exoskeleton in stroke rehabilitation. All patients and therapists recommended using an exoskeleton in some capacity in stroke care. A large debate surrounded the exact regimen of exoskeleton training, from only a few trial sessions to a longer and more intense program. Participants also identified considerations for optimizing exoskeleton rehabilitation, including criteria for exoskeleton candidates, staffing and scheduling barriers, and how to integrate exoskeleton training with other therapies and technology.    206 Appendix J  Sub-themes within each theme  


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items