Retrospective Study of Intensive Therapy Pathways in Children With Cerebral Palsy (P'TI-PARCOURS)

Background and Rationale Cerebral palsy (CP) is the most common cause of neurological disability in children, with a prevalence of approximately 2-2.5 per 1,000 live births in Europe. Hemiplegia, which accounts for approximately 30% of CP cases, predominantly affects the upper limb, impairing unimanual capacities, bimanual performance, and independence in activities of daily living.

The 2019 systematic review of evidence-based interventions for CP and subsequent international and French (HAS 2021) clinical practice guidelines strongly support intensive therapy approaches, including modified constraint-induced movement therapy (mCIMT), Hand and Arm Bimanual Intensive Training (HABIT), and HABIT-ILE. French HAS guidelines specifically recommend intensive rehabilitation programs of 60-90 hours incorporating motor learning principles targeting individualized functional goals.

Despite the strength of these recommendations, access to intensive therapy in France has historically been limited, leading many families to seek programs abroad. The Centre des Massues has proactively developed in-house intensive therapy programs integrated within its standard care pathway:

• COLHEMI (launched 2010): for children aged 6-12 years with unilateral CP; 70 hours over 2 weeks (10 therapy days), combining constraint-induced movement therapy, bimanual training, and goal-oriented therapy.
• Mini-COLHEMI (launched 2020): for children from age 2; 30 hours over 2 weeks (10 half-day sessions).

While these programs have been running for over a decade and generate systematic pre/post evaluations as part of routine care, their long-term impact on bimanual developmental trajectories and the factors associated with good response have not yet been formally studied.

Study Design Retrospective single-center study (Research Not Involving Human Subjects - RNIPH), using data collected during routine clinical care. Each patient serves as their own control (before/after design). All available longitudinal AHA data from each patient's medical record are collected to model bimanual developmental trajectories and quantify the effect of intensive therapy episodes relative to spontaneous maturation.

Data Collection and Management Data are collected retrospectively by Audrey COMBEY (occupational therapist) from electronic and paper medical records and occupational therapy databases. The dataset is pseudonymized using an inclusion number. The linkage table is maintained by the principal investigator (Audrey COMBEY) and by delegation by Rachel BARD-PONDARRE, stored on the Centre des Massues internal server with password protection. No data transfer to subcontractors is planned. Data will be retained for 3 years after study completion and archived for 15 years following last publication.

Statistical Analysis Approximately 75 patients will be included. Descriptive statistics will characterize the population (age, sex, side of hemiplegia, type of hemiplegia, follow-up setting, geographic distance to the Centre des Massues, history of botulinum toxin injections, intensive therapy sessions, and surgery). AHA score changes before and after each intensive therapy episode will be evaluated using paired Student t-tests (or Wilcoxon signed-rank test if normality is not met). GAS scores will be described by descriptive statistics and the proportion of patients reaching or exceeding their goal (score ≥ 0) will be calculated. A linear mixed model (LMM) will be applied to model bimanual developmental trajectories over time, with age as a fixed effect (natural maturation), COLHEMI/Mini-COLHEMI participation as an intervention variable, and random effects for individual baseline and progression rate. Logistic regression will be used to identify characteristics associated with good response. Parent satisfaction will be described by descriptive statistics per item. A 5% significance threshold is applied throughout. Analyses will be performed in R by Julie ROZAIRE.

Regulatory and Ethical Framework This study meets the definition of an internal study under French CNIL regulations and is classified as Research Not Involving Human Subjects (RNIPH) under the Jardé Law. It is exempt from CNIL authorization. The study is registered in the internal data processing activities register of the Centre des Massues (RGPD Article 30). The protocol has been submitted to the Internal Ethics Committee for Research (CIER) of the GHICL (IRB 00013355) and is pending approval.

Dissemination Results will be published in a peer-reviewed journal and presented at scientific society meetings (posters and/or oral communications). The principal investigator will be listed as first or last author. The study statistician will be co-author.