Algor-ethics: charting the ethical path for AI in critical care.

The integration of Clinical Decision Support Systems (CDSS) based on artificial intelligence (AI) in healthcare is groundbreaking evolution with enormous potential, but its development and ethical implementation, presents unique challenges, particularly in critical care, where physicians often deal with life-threating conditions requiring rapid actions and patients unable to participate in the decisional process. Moreover, development of AI-based CDSS is complex and should address different sources of bias, including data acquisition, health disparities, domain shifts during clinical use, and cognitive biases in decision-making. In this scenario algor-ethics is mandatory and emphasizes the integration of 'Human-in-the-Loop' and 'Algorithmic Stewardship' principles, and the benefits of advanced data engineering. The establishment of Clinical AI Departments (CAID) is necessary to lead AI innovation in healthcare, ensuring ethical integrity and human-centered development in this rapidly evolving field.

Balance Rehabilitation through Robot-Assisted Gait Training in Post-Stroke Patients: A Systematic Review and Meta-Analysis.

BACKGROUND: Balance impairment is a common disability in post-stroke survivors, leading to reduced mobility and increased fall risk. Robotic gait training (RAGT) is largely used, along with traditional training. There is, however, no strong evidence about RAGT superiority, especially on balance. This study aims to determine RAGT efficacy on balance of post-stroke survivors. METHODS: PubMed, Cochrane Library, and PeDRO databases were investigated. Randomized clinical trials evaluating RAGT efficacy on post-stroke survivor balance with Berg Balance Scale (BBS) or Timed Up and Go test (TUG) were searched. Meta-regression analyses were performed, considering weekly sessions, single-session duration, and robotic device used. RESULTS: A total of 18 trials have been included. BBS pre-post treatment mean difference is higher in RAGT-treated patients, with a pMD of 2.17 (95% CI 0.79; 3.55). TUG pre-post mean difference is in favor of RAGT, but not statistically, with a pMD of -0.62 (95%CI - 3.66; 2.43). Meta-regression analyses showed no relevant association, except for TUG and treatment duration (ß = -1.019, 95% CI - 1.827; -0.210, p-value = 0.0135). CONCLUSIONS: RAGT efficacy is equal to traditional therapy, while the combination of the two seems to lead to better outcomes than each individually performed. Robot-assisted balance training should be the focus of experimentation in the following years, given the great results in the first available trials. Given the massive heterogeneity of included patients, trials with more strict inclusion criteria (especially time from stroke) must be performed to finally define if and when RAGT is superior to traditional therapy.

Effectiveness of robot-assisted arm therapy in stroke rehabilitation: An overview of systematic reviews.

BACKGROUND: Robot-assisted arm therapy (RAT) has been used mainly in stroke rehabilitation in the last 20 years with rising expectations and growing evidence summarized in systematic reviews (SRs). OBJECTIVE: The aim of this study is to provide an overview of SRs about the effectiveness, within the ICF domains, and safety of RAT in the rehabilitation of adult with stroke compared to other treatments. METHODS: The search strategy was conducted using search strings adapted explicitly for each database. A screening base on title and abstract was realized to find all the potentially relevant studies. The methodological quality of the included SRs was assessed using AMSTAR-2. A pre-determined standardized form was used to realize the data extraction. RESULTS: 18 SRs were included in this overview. Generally, positive effects from the RAT were found for motor function and muscle strength, whereas there is no agreement for muscle tone effects. No effect was found for pain, and only a SR reported the positive impact of RAT in daily living activity. CONCLUSION: RAT can be considered a valuable option to increase motor function and muscle strength after stroke. However, the poor quality of most of the included SRs could limit the certainty around the results.

Robot-assisted rehabilitation for children with neurological disabilities: Results of the Italian consensus conference CICERONE.

BACKGROUND: The use of robotic technologies in pediatric rehabilitation has seen a large increase, but with a lack of a comprehensive framework about their effectiveness. OBJECTIVE: An Italian Consensus Conference has been promoted to develop recommendations on these technologies: definitions and classification criteria of devices, indications and limits of their use in neurological diseases, theoretical models, ethical and legal implications. In this paper, we present the results for the pediatric age. METHODS: A systematic search on Cochrane Library, PEDro and PubMed was performed. Papers published up to March 1st, 2020, in English, were included and analyzed using the methodology of the Centre for Evidence-Based Medicine in Oxford, AMSTAR2 and PEDro scales for systematic reviews and RCT, respectively. RESULTS: Some positives aspects emerged in the area of gait: an increased number of children reaching the stance, an improvement in walking distance, speed and endurance. Critical aspects include the heterogeneity of the studied cases, measurements and training protocols. CONCLUSION: Many studies demonstrate the benefits of robotic training in developmental age. However, it is necessary to increase the number of trials to achieve greater homogeneity between protocols and to confirm the effectiveness of pediatric robotic rehabilitation.

Reference theories and future perspectives on robot-assisted rehabilitation in people with neurological conditions: A scoping review and recommendations from the Italian Consensus Conference on Robotics in Neurorehabilitation (CICERONE).

BACKGROUND: Robot-based treatments are developing in neurorehabilitation settings. Recently, the Italian National Health Systems recognized robot-based rehabilitation as a refundable service. Thus, the Italian neurorehabilitation community promoted a national consensus on this topic. OBJECTIVE: To conceptualize undisclosed perspectives for research and applications of robotics for neurorehabilitation, based on a qualitative synthesis of reference theoretical models. METHODS: A scoping review was carried out based on a specific question from the consensus Jury. A foreground search strategy was developed on theoretical models (context) of robot-based rehabilitation (exposure), in neurological patients (population). PubMed and EMBASE® databases were searched and studies on theoretical models of motor control, neurobiology of recovery, human-robot interaction and economic sustainability were included, while experimental studies not aimed to investigate theoretical frameworks, or considering prosthetics, were excluded. RESULTS: Overall, 3699 records were screened and finally 9 papers included according to inclusion and exclusion criteria. According to the population investigated, structured information on theoretical models and indications for future research was summarized in a synoptic table. CONCLUSION: The main indication from the Italian consensus on robotics in neurorehabilitation is the priority to design research studies aimed to investigate the role of robotic and electromechanical devices in promoting neuroplasticity.

Effects of robotic upper limb treatment after stroke on cognitive patterns: A systematic review.

BACKGROUND: Robotic therapy (RT) has been internationally recognized for the motor rehabilitation of the upper limb. Although it seems that RT can stimulate and promote neuroplasticity, the effectiveness of robotics in restoring cognitive deficits has been considered only in a few recent studies. OBJECTIVE: To verify whether, in the current state of the literature, cognitive measures are used as inclusion or exclusion criteria and/or outcomes measures in robotic upper limb rehabilitation in stroke patients. METHODS: The systematic review was conducted according to PRISMA guidelines. Studies eligible were identified through PubMed/MEDLINE and Web of Science from inception to March 2021. RESULTS: Eighty-one studies were considered in this systematic review. Seventy-three studies have at least a cognitive inclusion or exclusion criteria, while only seven studies assessed cognitive outcomes. CONCLUSION: Despite the high presence of cognitive instruments used for inclusion/exclusion criteria their heterogeneity did not allow the identification of a guideline for the evaluation of patients in different stroke stages. Therefore, although the heterogeneity and the low percentage of studies that included cognitive outcomes, seemed that the latter were positively influenced by RT in post-stroke rehabilitation. Future larger RCTs are needed to outline which cognitive scales are most suitable and their cut-off, as well as what cognitive outcome measures to use in the various stages of post-stroke rehabilitation.

Evidence-based improvement of gait in post-stroke patients following robot-assisted training: A systematic review.

BACKGROUND: The recovery of walking after stroke is a priority goal for recovering autonomy. In the last years robotic systems employed for Robotic Assisted Gait Training (RAGT) were developed. However, literature and clinical practice did not offer standardized RAGT protocol or pattern of evaluation scales. OBJECTIVE: This systematic review aimed to summarize the available evidence on the use of RAGT in post-stroke, following the CICERONE Consensus indications. METHODS: The literature search was conducted on PubMed, Cochrane Library and PEDro, including studies with the following criteria: 1) adult post-stroke survivors with gait disability in acute/subacute/chronic phase; 2) RAGT as intervention; 3) any comparators; 4) outcome regarding impairment, activity, and participation; 5) both primary studies and reviews. RESULTS: Sixty-one articles were selected. Data about characteristics of patients, level of disability, robotic devices used, RAGT protocols, outcome measures, and level of evidence were extracted. CONCLUSION: It is possible to identify robotic devices that are more suitable for specific phase disease and level of disability, but we identified significant variability in dose and protocols. RAGT as an add-on treatment seemed to be prevalent. Further studies are needed to investigate the outcomes achieved as a function of RAGT doses delivered.

Gait robot-assisted rehabilitation in persons with spinal cord injury: A scoping review.

BACKGROUND: Many robots are available for gait rehabilitation (BWSTRT and ORET) and their application in persons with SCI allowed an improvement of walking function. OBJECTIVE: The aim of the study is to compare the effects of different robotic exoskeletons gait training in persons with different SCI level and severity. METHODS: Sixty-two studies were included in this systematic review; the study quality was assessed according to GRADE and PEDro's scale. RESULTS: Quality assessment of included studies (n = 62) demonstrated a prevalence of evidence level 2; the quality of the studies was higher for BWSTRT (excellent and good) than for ORET (fair and good). Almost all persons recruited for BWSTRT had an incomplete SCI; both complete and incomplete SCI were recruited for ORET. The SCI lesion level in the persons recruited for BWSTRT are from cervical to sacral; mainly from thoracic to sacral for ORET; a high representation of AIS D lesion resulted both for BWSTRT (30%) and for ORET (45%). The walking performance, tested with 10MWT, 6MWT, TUG and WISCI, improved after exoskeleton training in persons with incomplete SCI lesions, when at least 20 sessions were applied. Persons with complete SCI lesions improved the dexterity in walking with exoskeleton, but did not recover independent walking function; symptoms such as spasticity, pain and cardiovascular endurance improved. CONCLUSION: Different exoskeletons are available for walking rehabilitation in persons with SCI. The choice about the kind of robotic gait training should be addressed on the basis of the lesion severity and the possible comorbidities.

Robot-assisted gait training in patients with Parkinson's disease: Implications for clinical practice. A systematic review.

BACKGROUND: Gait impairments are common disabling symptoms of Parkinson's disease (PD). Among the approaches for gait rehabilitation, interest in robotic devices has grown in recent years. However, the effectiveness compared to other interventions, the optimum amount of training, the type of device, and which patients might benefit most remains unclear. OBJECTIVE: To conduct a systematic review about the effects on gait of robot-assisted gait training (RAGT) in PD patients and to provide advice for clinical practice. METHODS: A search was performed on PubMed, Scopus, PEDro, Cochrane library, Web of science, and guideline databases, following PRISMA guidelines. We included English articles if they used a robotic system with details about the intervention, the parameters, and the outcome measures. We evaluated the level and quality of evidence. RESULTS: We included twenty papers out of 230 results: two systematic reviews, 9 randomized controlled trials, 4 uncontrolled studies, and 5 descriptive reports. Nine studies used an exoskeleton device and the remainders end-effector robots, with large variability in terms of subjects' disease-related disability. CONCLUSIONS: RAGT showed benefits on gait and no adverse events were recorded. However, it does not seem superior to other interventions, except in patients with more severe symptoms and advanced disease.

Upper Limb Robotic Rehabilitation for Patients with Cervical Spinal Cord Injury: A Comprehensive Review.

The upper extremities limitation represents one of the essential functional impairments in patients with cervical spinal cord injury. Electromechanics assisted devices and robots are increasingly used in neurorehabilitation to help functional improvement in patients with neurological diseases. This review aimed to systematically report the evidence-based, state-of-art on clinical applications and robotic-assisted arm training (RAT) in motor and functional recovery in subjects affected by cervical spinal cord injury. The present study has been carried out within the framework of the Italian Consensus Conference on "Rehabilitation assisted by robotic and electromechanical devices for persons with disability of neurological origin" (CICERONE). PubMed/MEDLINE, Cochrane Library, and Physiotherapy Evidence Database (PEDro) databases were systematically searched from inception to September 2021. The 10-item PEDro scale assessed the study quality for the RCT and the AMSTAR-2 for the systematic review. Two different authors rated the studies included in this review. If consensus was not achieved after discussion, a third reviewer was interrogated. The five-item Oxford CEBM scale was used to rate the level of evidence. A total of 11 studies were included. The selected studies were: two systematic reviews, two RCTs, one parallel-group controlled trial, one longitudinal intervention study and five case series. One RCT was scored as a high-quality study, while the systematic review was of low quality. RAT was reported as feasible and safe. Initial positive effects of RAT were found for arm function and quality of movement in addition to conventional therapy. The high clinical heterogeneity of treatment programs and the variety of robot devices could severely affect the generalizability of the study results. Therefore, future studies are warranted to standardize the type of intervention and evaluate the role of robotic-assisted training in subjects affected by cervical spinal cord injury.

What does evidence tell us about the use of gait robotic devices in patients with multiple sclerosis? A comprehensive systematic review on functional outcomes and clinical recommendations.

INTRODUCTION: There is growing evidence on the efficacy of gait robotic rehabilitation in patients with multiple sclerosis (MS), but most of the studies have focused on gait parameters. Moreover, clear indications on the clinical use of robotics still lack. As part of the CICERONE Italian Consensus on Robotic Rehabilitation, the aim of this systematic review was to investigate the existing evidence concerning the role of lower limb robotic rehabilitation in improving functional recovery in patients with MS. EVIDENCE ACQUISITION: We searched for and systematically reviewed evidence-based studies on gait robotic rehabilitation in MS, between January 1st, 2010 and December 31st, 2020, in the following databases: Cochrane Library, PEDro, PubMed and Google Scholar. The study quality was assessed by the 16-item assessment of multiple systematic reviews 2 (AMSTAR 2) and the 10-item PEDro scale for the other research studies. EVIDENCE SYNTHESIS: After an accurate screening, only 17 papers were included in the review, and most of them (13 RCT) had a level II evidence. Most of the studies used the Lokomat as a grounded robotic device, two investigated the efficacy of end-effectors and two powered exoskeletons. Generally speaking, robotic treatment has beneficial effects on gait speed, endurance and balance with comparable outcomes to those of conventional treatments. However, in more severe patients (EDSS >6), robotics leads to better functional outcomes. Notably, after gait training with robotics (especially when coupled to virtual reality) MS patients also reach better non-motor outcomes, including spasticity, fatigue, pain, psychological well-being and quality of life. Unfortunately, no clinical indications emerge on the treatment protocols. CONCLUSIONS: The present comprehensive systematic review highlights the potential beneficial role on functional outcomes of the lower limb robotic devices in people with MS. Future studies are warranted to evaluate the role of robotics not only for walking and balance outcomes, but also for other gait-training-related benefits, to identify appropriate outcome measures related to a specific subgroup of MS subjects' disease severity.

Robot-assisted arm therapy in neurological health conditions: rationale and methodology for the evidence synthesis in the CICERONE Italian Consensus Conference.

BACKGROUND: Robot-assisted Arm Therapy (RAT) has been increasingly applied in the last years for promoting functional recovery in patients with disabilities related to neurological health conditions. Evidence of a knowledge-to-action gap for applying robot-assisted technologies in the rehabilitation of patients with neurological health conditions and the difficulty to apply and tailor the knowledge to the local contexts solicited the need for a national consensus conference on these interventions. AIM: The aim of this paper was to explain the methodology used by the working group dedicated to synthesizing evidence on the effectiveness of RAT in neurological health conditions in the context of the CICERONE Italian Consensus Conference. DESIGN: The methodological approach of the working group. SETTING: All rehabilitation settings. POPULATION: Patients with disability following a neurological health condition. METHODS: Following the indications proposed by the Methodological Manual published by the Italian National Institute of Health, a Promoting Committee and a Technical Scientific Committee have been set up. Six working groups (WGs) have been composed to collect evidence on different questions, among which WG2.2 was focused on the effectiveness of RAT in neurological health conditions. RESULTS: WG2.2 started its work defining the specific research questions. It was decided to adopt the ICF as the reference framework for the reporting of all outcomes. Literature search, data extraction and qualitative assessment, evidence analysis and synthesis have been performed. CONCLUSIONS: This paper summarized the methodological approaches used by the WG2.2 of the CICERONE Italian Consensus Conference to define the effectiveness of RAT in the management of patients with neurological health conditions. CLINICAL REHABILITATION IMPACT: WG2.2 synthesis might help clinicians, researchers, and all rehabilitation stakeholders to address the use of RAT in the Individualized Rehabilitation Plan, to guide the allocation of resources and define clinical protocols and indications for the management of patients with different neurological health conditions.

Robotic-assisted gait rehabilitation following stroke: a systematic review of current guidelines and practical clinical recommendations.

INTRODUCTION: Stroke is the third leading cause of adult disability worldwide, and lower extremity motor impairment is one of the major determinants of long-term disability. Although robotic therapy is becoming more and more utilized in research protocols for lower limb stroke rehabilitation, the gap between research evidence and its use in clinical practice is still significant. The aim of this study was to determine the scope, quality, and consistency of guidelines for robotic lower limb rehabilitation after stroke, in order to provide clinical recommendations. EVIDENCE ACQUISITION: We systematically reviewed stroke rehabilitation guideline recommendations between January 1, 2010 and October 31, 2020. We explored electronic databases (N.=4), guideline repositories and professional rehabilitation networks (N.=12). Two independent reviewers used the Appraisal of Guidelines for Research and Evaluation (AGREE) II instrument, and brief syntheses were used to evaluate and compare the different recommendations, considering only the most recent version. EVIDENCE SYNTHESIS: From the 1219 papers screened, ten eligible guidelines were identified from seven different regions/countries. Four of the included guidelines focused on stroke management, the other six on stroke rehabilitation. Robotic rehabilitation is generally recommended to improve lower limb motor function, including gait and strength. Unfortunately, there is still no consensus about the timing, frequency, training session duration and the exact characteristics of subjects who could benefit from robotics. CONCLUSIONS: Our systematic review shows that the introduction of robotic rehabilitation in standard treatment protocols seems to be the future of stroke rehabilitation. However, robot assisted gait training (RAGT) for stroke needs to be improved with new solutions and in clinical practice guidelines, especially in terms of applicability.

State of the art and challenges for the classification of studies on electromechanical and robotic devices in neurorehabilitation: a scoping review.

INTRODUCTION: The rapid development of electromechanical and robotic devices has profoundly influenced neurorehabilitation. Growth in the scientific and technological aspects thereof is crucial for increasing the number of newly developed devices, and clinicians have welcomed such growth with enthusiasm. Nevertheless, improving the standard for the reporting clinical, technical, and normative aspects of such electromechanical and robotic devices remains an unmet need in neurorehabilitation. Accordingly, this study aimed to analyze the existing literature on electromechanical and robotic devices used in neurorehabilitation, considering the current clinical, technical, and regulatory classification systems. EVIDENCE ACQUISITION: Within the CICERONE Consensus Conference framework, studies on electromechanical and robotic devices used for upper- and lower-limb rehabilitation in persons with neurological disabilities in adulthood and childhood were reviewed. We have conducted a literature search using the following databases: MEDLINE, Cochrane Library, PeDro, Institute of Electrical and Electronics Engineers, Science Direct, and Google Scholar. Clinical, technical, and regulatory classification systems were applied to collect information on the electromechanical and robotic devices. The study designs and populations were investigated. EVIDENCE SYNTHESIS: Overall, 316 studies were included in the analysis. More than half (52%) of the studies were randomised controlled trials (RCTs). The population investigated the most suffered from strokes, followed by spinal cord injuries, multiple sclerosis, cerebral palsy, and traumatic brain injuries. In total, 100 devices were described; of these, 19% were certified with the CE mark. Overall, the main type of device was an exoskeleton. However, end-effector devices were primarily used for the upper limbs, whereas exoskeletons were used for the lower limbs (for both children and adults). CONCLUSIONS: The current literature on robotic neurorehabilitation lacks detailed information regarding the technical characteristics of the devices used. This affects the understanding of the possible mechanisms underlying recovery. Unfortunately, many electromechanical and robotic devices are not provided with CE marks, strongly hindering the research on the clinical outcomes of rehabilitation treatments based on these devices. A more significant effort is needed to improve the description of the robotic devices used in neurorehabilitation in terms of the technical and functional details, along with high-quality RCT studies.

Effects of robot-assisted gait training on postural instability in Parkinson's disease: a systematic review.

INTRODUCTION: Postural instability is a cardinal feature of Parkinson's disease, together with rest tremor, rigidity and bradykinesia. It is a highly disabling symptom that becomes increasingly common with disease progression and represents a major source of reduced quality of life in patients with Parkinson's disease. Rehabilitation aims to enable patients with Parkinson's disease to maintain their maximum level of mobility, activity and independence. To date, a wide range of rehabilitation approaches has been employed to treat postural instability in Parkinson's disease, including robotic training. Our main aim was to conduct a systematic review of current literature about the effects of robot-assisted gait training on postural instability in patients with Parkinson's disease. EVIDENCE ACQUISITION: A systematic search using the following MeSH terms "Parkinson disease," "postural balance," "robotics," "rehabilitation" AND string "robotics [mh]" OR "robot-assisted" OR "electromechanical" AND "rehabilitation [mh]" OR "training" AND "postural balance [mh]" was conducted on PubMed, Cochrane Library and Pedro electronic databases. Full text articles in English published up to December 2020 were included. Data about patient characteristics, robotic devices, treatment procedures and outcome measures were considered. Every included article got checked for quality. Level of evidence was defined for all studies. EVIDENCE SYNTHESIS: Three authors independently extracted and verified data. In total, 18 articles (2 systematic reviews, 9 randomized controlled trials, 4 uncontrolled studies and 3 case series/case reports) were included. Both end-effector and exoskeleton devices were investigated as to robot-assisted gait training modalities. No clear relationship between treatment parameters and clinical conditions was observed. We found a high level of evidence about the effects of robot-assisted gait training on balance and freezing of gait in patients with Parkinson's disease. CONCLUSIONS: This systematic review provides to the reader a complete overview of current literature and levels of evidence about the effects of robot-assisted gait training on postural instability issues (static and dynamic balance, freezing of gait, falls, confidence in activities of daily living and gait parameters related to balance skills) in patients with Parkinson's disease.

What is the impact of robotic rehabilitation on balance and gait outcomes in people with multiple sclerosis? A systematic review of randomized control trials.

INTRODUCTION: In recent years, robot-assisted gait training (RAGT) has been proposed as therapy for balance and gait dysfunctions in people with multiple sclerosis (PwMS). Through this systematic review, we aimed to discuss the impact of RAGT on balance and gait outcomes. Furthermore, characteristics of the training in terms of robots used, participants characteristics, protocols and combined therapeutic approaches have been described. EVIDENCE ACQUISITION: As part of the Italian Consensus on robotic rehabilitation "CICERONE" a systematic search was provided in PubMed, the Cochrane Library and PEDro to identify relevant studies published before December 2019. Only randomized control trials (RCT) involving RAGT for PwMS were included. PEDro scale was used to assess the risk of bias and the Oxford Center for Evidence-Based Medicine (OCEBM) was used to assess level of evidence of included studies. EVIDENCE SYNTHESIS: The search on databases resulted in 336 records and, finally, 12 studies were included. RAGT was provided with Exoskeleton in ten studies (6-40 session, 2-5 per week) and with end-effector in two studies (12 sessions, 2-3 per week) with large variability in terms of participants' disability. All the exoskeletons were combined with bodyweight support treadmill and movement assistance varied from 0% to 100% depending on participants' disability, two studies combined exoskeleton with virtual reality. The end-effector speed ranged between 1.3 and 1.8 km/h, with bodyweight support starting from 50% and progressively reduced. In seven out of twelve studies RAGT was provided in a multimodal rehabilitation program or in combination with standard physical therapy. There is level 2 evidence that RAGT has positive impact in PwMS, reaching the minimally clinically importance difference in Berg Balance Scale, six-minute walking test and gait speed. CONCLUSIONS: In available RCT, RAGT is mostly provided with exoskeleton devices and improves balance and gait outcomes in a clinically meaningful way. Considering several advantages in terms of safety, motor assistance and intensity of training provided, RAGT should be promoted for PwMS with severe disability in a multimodal rehabilitation context as an opportunity to maximize recovery.

Systematic review of guidelines to identify recommendations for upper limb robotic rehabilitation after stroke.

INTRODUCTION: Upper limb motor impairment is one of the most frequent stroke consequences. Robot therapy may represent a valid option for upper limb stroke rehabilitation, but there are still gaps between research evidence and their use in clinical practice. The aim of this study was to determine the quality, scope, and consistency of guidelines clinical practice recommendations for upper limb robotic rehabilitation in stroke populations. EVIDENCE ACQUISITION: We searched for guideline recommendations on stroke published between January 1st, 2010 and January 1st, 2020. Only the most recent guidelines for writing group were selected. Electronic databases (N.=4), guideline repertories and professional rehabilitation networks (N.=12) were searched. We systematically reviewed and assessed guidelines containing recommendation statements about upper limb robotic rehabilitation for adults with stroke (PROSPERO registration number: CRD42020173386). EVIDENCE SYNTHESIS: Four independent reviewers used the Appraisal of Guidelines for Research and Evaluation (AGREE) II instrument, and textual syntheses were used to appraise and compare recommendations. From 1324 papers that were screened, eight eligible guidelines were identified from six different regions/countries. Half of the included guidelines focused on stroke management, the other half on stroke rehabilitation. Rehabilitation assisted by robotic devices is generally recommended to improve upper limb motor function and strength. The exact characteristics of patients who could benefit from this treatment as well as the correct timing to use it are not known. CONCLUSIONS: This systematic review has identified many opportunities to modernize and otherwise improve stroke patients' upper limb robotic therapy. Rehabilitation assisted by robot or electromechanical devices for stroke needs to be improved in clinical practice guidelines in particular in terms of applicability.