Artificial Intelligence for developing a personalised functional profile and supporting upper-limb rehabilitation decision-making in children with unilateral Cerebral Palsy (AInCP): protocol for an observational study

Introduction Cerebral palsy (CP) is a lifelong non-progressive disorder, resulting from disturbances in the developing or infant brain, with approximately 17 million people worldwide being diagnosed. More than one-third of the population with CP has unilateral CP (UCP), whereby the primary sensorimotor impairments affect predominantly one side of the body, mainly the upper limb (UpL), compromising independence and quality of life. Associated comorbidities (eg, cognitive, visual, behavioural impairments and epilepsy) can impede the establishment of a clear functional diagnosis, consequently compromising the formation of effective intervention strategies. Driven by the clinical presentation complexity of UCP, the main objective of the Artificial Intelligence in Cerebral Palsy (AInCP) study is to establish a valid functional diagnosis supporting accurate rehabilitative priorities, identifying multidimensional functional profiles that may inform personalised rehabilitation planning, utilising AI. Through the integration of multimodal data, the ultimate goal is to develop and test evidence-based clinical decision support tools for the personalised functional diagnosis and UpL assessment in children with UCP by developing, testing and validating AI-cost-effective strategies. This will enable more accurate, personalised functional diagnoses and treatment planning, enhancing quality of life and efficiency of care for children with UCP. Methods and analysis This prospective, multicentre, observational clinical study will include 250 children with UCP and 250 typically developing children (TDC). Participants will be enrolled and assessed within four clinical centres distributed across Europe. Data from children with UCP will be collected by three broad domains: clinical assessments and questionnaires, neuroimaging data and data from UpL daily-life monitoring using wearable sensors. TDC will be evaluated using clinical tests and wearable sensors. All the items identified and the whole methodology have been co-created and co-designed among a multidisciplinary team, including caregivers and persons with lived experience. Data will be analysed with AI and machine-learning algorithms to understand the interdependencies across functional domains and create functional profiles that can guide personalised treatment planning. Non-clinical data such as sleep quality, technology familiarity and technology acceptance will also be collected to provide an empirical foundation for conducting an early health technology assessment of the solutions developed within the AInCP project. Ethics and dissemination Ethical approval has been obtained by all involved clinical sites. Written informed consent will be obtained from the legal representatives of all participating children, and age-appropriate assent will be obtained from children according to local ethical requirements. All procedures will be closely monitored by ethicists during the implementation of AI to ensure patient privacy, prevent bias and avoid unintended harm while developing robust, fair and trustworthy AI solutions. The results of this study will be disseminated through open-access peer-reviewed journals relevant for paediatrics to ensure wide accessibility. Results will be communicated by the professionals directly involved in the clinical study through national and international academic conferences, relevant associations and organisations, including policy makers. Trial registration number NCT06073522.