evidence-based medicine and benchmarking to assess upper limb functions in neurological diseases

In neurorehabilitation and the assistance of disabled people, there has been rapid and continuous improvement in clinical tools to quantify body function and dysfunction following neurological conditions. The assessments of the motor functions and the influences of deficits on daily life activities are important to reveal relevant movement limitations, drive interventions for improving functional restoration, or identify the proper assistive technology according to the resulting benefits. Currently, this assessment relies on a visual and physical inspection by a clinician who uses criteria-based ordinal scales. This approach is part of the so-called Evidence-Based Medicine. The reliability and sensitivity to gross motor changes of standard clinical scales have been largely documented. Within this goal, technology-aided assessments could provide clinicians with objective, accurate, and repeatable measurements of a patient’s capacity. In the last years, hundreds of studies have exploited quantitative metrics for these goals, resulting in a plethora of potentially helpful but extremely variegated evaluation methods and protocols. Consequently, we are observing a growing awareness of the importance of benchmarking (i.e., standardized and reproducible evaluation). Considering the lower limb, standardized protocols have been developed for gait analysis, and they are now shared and commonly used by clinicians and researchers. When moving to the upper limb, instead, the current state of the art lacks a standardized and shared framework for the clinical assessment. Therefore, the goal of this thesis is to answer the growing need for evidence in the assessment of upper limb function in the field of neurorehabilitation and the assistance of people with neurological disorders. The evaluation is considered from three perspectives: i) the functional evaluation of the arm, ii) the quantification of the therapeutic effect (i.e., effectiveness and the long-term improvements of an intervention), and iii) the assessment of the orthotic effect (i.e., the immediate changes due to the use of an external device). In the first part of the work, we rigorously implemented the standard Evidence-Based Medicine in two clinical studies to determine the effectiveness of upper limb assistive devices in people with muscular dystrophy. We involved the PUL module (externally-assessed scale) and the Abilhand questionnaire (patient-reported questionnaire). These clinical scales represent the gold standard for assessment procedures in people with muscular dystrophy. We were able to depict a clear and efficient analysis of the impact of assistive devices. However, these clinical scales showed some drawbacks (e.g., inability to detect fatigue, ceiling/floor effects). As a consequence, the need to integrate quantitative measures into standard clinical scales in the evaluation of both arm functionality and robot effectiveness has emerged. In the second part of the work, we developed and validated a benchmarking framework for upper limb capacity. The work resulted from a multidisciplinary and iterative discussion among several partners (i.e., two research institutions, three clinical centers, and the Eurobench Consortium) with previous experience in benchmarking methodology, robotics, and clinical neurorehabilitation. The scheme was designed to enable an instrumented evaluation of arm capacity and to assess the effectiveness of rehabilitative interventions with high reproducibility and resolution. It includes four key elements: i) a taxonomy that identifies and classifies the relevant upper limb motor skills and motor abilities, ii) a selection of performance indicators able to quantify each motor ability, iii) the required sensor networks to extract the performance indicators, and iv) a set of standardized protocols that should be followed to obtain comparable results. We validated the benchmarking scheme in terms of repeatability, reproducibility, and clinical meaningfulness, performing a prospective multicenter study on neurologically intact young participants, elderly subjects, and post-stroke people. 60 subjects (30 young healthy, 15 elderly healthy, and 15 post-stroke) completed the benchmarking protocol. The framework was repeatable among different assessors and instrumentation. Age did not significantly impact the performance indicators of the scheme for healthy subjects. In post-stroke subjects, the movements presented decreased smoothness and speed, the movement amplitude was reduced, and the muscular activation showed lower power and lower intra-limb coordination. Spectral arc length and electromyography mean frequency were correlated with the Action Research Arm Test. We revised the original framework reducing it to three motor skills, and we extracted 14 significant performance indicators with a good correlation with the clinical scale. Finally, in the last part of the thesis, we investigated the reproducibility of the benchmarking scheme on the AGREE system, a platform for the rehabilitation of the upper limb based on a four degrees of freedom exoskeleton we developed at the Politecnico di Milano. The benchmarking scheme was integrated into the AGREE system. Through a Graphical User Interface, the assessor can choose the motor skills, the outcome measures domains of interest (kinematics and electromyography), and select the assistive contribution of the exoskeleton. Then, the system drives the user during the execution of the protocol while recording and saving the selected outcome domain signals and produces a final evaluation report. We performed a feasibility study on 15 healthy volunteers and three neurological patients to investigate the reproducibility of the benchmarking scheme with this technology. The reproducibility was verified for the motor abilities efficacy, efficiency, and muscular effort. It was only partially achieved for intra-limb coordination and power. For the other motor abilities, instead, the reproducibility of the scheme was not demonstrated. The AGREE system could not accurately measure the kinematics of the movement, and the induced movement was jerkier and slower. We must underline that the non-reproducibility of the benchmarking scheme is due to the AGREE technology, not to the framework itself. With this work, we could identify a subset of performance indicators that could adequately be exploited to monitor the patient's progress and, accordingly, customize the therapy and those Indicators that are affected by the exoskeleton and should not be involved in this evaluation. This work highlighted that Evidence-Based Medicine and benchmarking are nowadays the desirable approaches for evaluating the upper limb capabilities of frail subjects and assessing and comparing the performance of different rehabilitative or assistive interventions. Overall, the assessment tools proposed in this thesis can offer clinicians flexible, quantitative, and standardized methodologies. The proposed technology-driven solution could provide a promising complement to conventional clinical assessments. Further work is needed in this field, and it should be carried out in strict collaboration among all the figures involved in the processes of cure and care of people with neurological disorders (e.g., physicians, physiotherapists, engineers, and patients). This will lead not only to a significant reduction in the time spent assessing the patient's status but also to the creation of personalized and integrated rehabilitation and assistive technologies.