Hospital and Home Rehabilitation Based on Social Care Robotics and Artificial Intelligence for Pediatric Patients

This study describes a randomized controlled clinical trial designed to evaluate the usability, user experience, and clinical effectiveness of Inrobics Rehab, a pediatric rehabilitation solution based on Socially Assistive Robotics (SAR) and Artificial Intelligence, in children and adolescents with Acquired Brain Injury (ABI) aged 7-16 years. The rationale is that SAR and virtual training can provide an interactive and motivating environment that may increase engagement and adherence, while enabling controlled, monitored exercises and objective follow-up of progress; however, evidence in pediatric ABI remains limited. The study hypothesis is that Inrobics Rehab can improve motor, cognitive, and socio-emotional outcomes in this population, while delivering a positive user experience with good usability and high adherence. The study is planned as an experimental, longitudinal, prospective, controlled, randomized clinical trial using a CE-marked medical device (Inrobics Rehab; CE Class I, registration no. RPS/777/2021). It is embedded in a 36-month collaborative research project involving INROBICS, the Hospital Nacional de Parapléjicos, and the Hospital Infantil Universitario Niño Jesús (HIUNJ), where HIUNJ plays a central role in patient selection, clinical assessments, delivery of therapy sessions, and clinical data collection in a real-world hospital setting. Participants will be pediatric patients with confirmed ABI receiving care at the HIUNJ Rehabilitation Service. Eligibility requires: age 7-16 years; sufficient comprehension to follow basic instructions and participate actively; confirmed acquired brain injury; clinical stability to engage in therapy; and ability to maintain independent sitting, with or without assistive supports (e.g., belts, wedges, trunk supports). Key exclusion criteria include unstable orthopedic injuries in the upper limbs (e.g., non-consolidated fractures or unstable osteosynthesis), severe pain/rigidity and/or severe upper-limb spasticity, severe pulmonary or cardiac disease requiring constant monitoring during exercise, a history of abnormal response to physical effort prior to ABI, severe visual impairment, severe cognitive impairment and/or disabling psychiatric disease, or any other condition judged by the medical team to prevent safe participation. The protocol emphasizes enhanced protection measures for this vulnerable population through reinforced consent/assent procedures, strict selection criteria, monitoring, and additional risk minimization. A total sample of 30 pediatric ABI participants is planned, split into two etiological cohorts to support balanced comparisons: 15 with oncological etiology and 15 with non-oncological etiology. Initial selection is by convenience sampling among eligible patients routinely followed at HIUNJ; subsequently, within each etiological subgroup, participants are randomly allocated to an experimental or control arm, with approximately half in each arm. Randomization is implemented using sequentially numbered sealed envelopes containing group assignment (control vs intervention), based on a computer-generated random list; hospital evaluators remain blinded to group allocation to support objectivity in outcome collection. The protocol acknowledges that the theoretically estimated sample size to detect moderate effects would be larger, but justifies feasibility constraints and the relevance of the proposed sample in the context of pediatric SAR studies. The experimental intervention consists of a structured rehabilitation program delivered through Inrobics Rehab Clinic (hospital) and Inrobics Rehab Virtual (home). The program includes 30 total sessions over 5 weeks of active intervention, distributed equally between 15 in-hospital sessions and 15 home-based virtual sessions, delivered in an alternating manner (weeks combining clinic and virtual work) to allow progressive adaptation while maintaining continuous clinical follow-up. Each hospital session lasts 30 minutes, with a planned frequency of 3 sessions per week, programmed and directed by a professional using the clinical platform and supervised throughout. In the home phase, participants complete 3 sessions per week via the virtual platform on a tablet, following exercises prescribed by the clinician and guided by an interactive virtual avatar; gamification elements are included to enhance motivation and participation, and the system records objective metrics to support monitoring of therapeutic progress and adaptation. A contingency window of ±2 additional weeks is allowed solely to recover missed sessions due to justified absences, without extending beyond this limit, to preserve a homogeneous "dose" of intervention and the internal validity of the design. The control arm receives no therapeutic intervention within the trial; control participants complete the same evaluation schedule as the experimental group to enable between-group comparisons and help isolate the effect of the robotics/virtual intervention from external factors. To ensure fairness, the protocol includes a post-trial access option for control participants: after the clinical trial ends, they may access Inrobics Rehab Virtual for 1-3 months through a temporary license, or opt to undergo the full intervention program (hospital + home phases) if desired. At study completion, all participants in both groups receive an individualized final report summarizing assessment results over the program. Outcome assessment is scheduled at baseline (pre-intervention), immediately after the intervention period (post-intervention at \~5 weeks for the intervention arm; at a comparable timepoint for controls), and at 3 months follow-up to determine whether observed changes are maintained. In addition, during the intervention, session-by-session ratings of satisfaction, fatigue, and/or pain are collected (hospital and home), and user-experience parameters are continuously recorded by the platform; heart rate is monitored only during hospital sessions. A dedicated Case Report Form (CRF) is included as an annex to standardize data capture and organization. The protocol defines a broad set of outcomes covering motor performance, cognition/emotional status, health perception and burden, and usability/user experience. Motor efficacy is assessed using standardized measures including trunk control (Trunk Control Measurement Scale, TCMS), manual dexterity (Box \& Block, complemented by Leap Motion measures), and manual function (ABILHAND-Kids, with optional Jebsen-Taylor), as well as dominance (Edinburgh inventory) and upper-limb strength (Daniels and/or dynamometry). Neuropsychological and emotional domains include visual attention (perception/difference or "faces" tests), working memory (WISC-V digit span), depressive symptoms (Children's Depression Inventory for those under 14), anxiety (STAIC), and executive functioning (EFE inventory). Health perception and burden include pediatric quality of life (EQoL-5D-Y5L), pain (VAS), fatigue (EPInfant), and caregiver burden (Zarit). Usability and user experience are examined using both questionnaires and platform analytics, such as motor accuracy, reaction time, engagement (e.g., whether the user remains oriented toward the platform/tablet), aggregated performance metrics, and execution speed; physiological response to effort is assessed with heart-rate monitoring using Polar Verity Sense during hospital sessions. Satisfaction is assessed via SAM (child) and QUEST 2.0 (caregiver), and involvement/adherence-related constructs include a Hopkins measure of engagement with the training program. Planned analyses include descriptive statistics and inferential comparisons across timepoints and between groups, selecting methods based on variable type and distribution, with standard significance thresholds; the protocol also identifies potential biases such as limited sample size, subjectivity in self-reported scales, and potential attrition particularly during the home phase. Ethical conduct and data protection are addressed in line with relevant regulations and the specific needs of pediatric participants. The study is conducted according to the Declaration of Helsinki, and participation requires formal informed consent (with the option to withdraw at any time without affecting usual care). Data are pseudonymized using a sequential alphanumeric code assigned by inclusion order; all questionnaires, clinical measures, physiological records, and platform data are labeled only with this identifier, while the code-to-identity link is stored separately with restricted access. Pseudonymized datasets for analysis are stored in the institutional OneDrive environment with restricted, credential-based access, and cross-site data sharing is limited to pseudonymized information through controlled shared folders with the sponsor (INROBICS). Documentation generated for the clinical investigation is retained for at least 10 years after study completion (and potentially longer depending on applicable medical device regulatory requirements and market placement), with measures to ensure integrity, accessibility, and protection against loss, alteration, or unauthorized access.