A Novel Statistical Approach to Evaluating Rehabilitation Outcomes in Children with Hemiparetic Cerebral Palsy

Background. Gait impairments are a primary contributor to disability in children with cerebral palsy (CP). Traditional clinical scales and group-level statistical tests often fail to detect small but clinically meaningful improvements following rehabilitation, underscoring the need for more sensitive, objective evaluation methods.

Objective. To compare an individualized magnitude-based decision (MBD) statistical approach applied to inertial-sensor-based instrumental gait analysis (IGA-IS) data with conventional group statistical methods (p-value) for assessing the efficacy of a rehabilitation course in children with the hemiparetic form of CP.

Methods. This observational, single-center, prospective study analyzed IGA-IS records obtained immediately before and after a rehabilitation program in 23 children aged 8–17 years with hemiparetic CP, treated in the pediatric medical rehabilitation department of the Russian Children’s Clinical Hospital (RDKB).

Results. In this cohort, standard paired t-tests failed to reveal any statistically significant group-level changes in spatiotemporal gait parameters (p > 0.05 for all measures). In contrast, individualized MBD analysis demonstrated clinically significant improvements in the majority of patients, including:

• Increased step velocity and stride length on the paretic side in 47 % and 33 % of patients, respectively.
• Expanded knee (67 %) and ankle (40 %) joint range of motion.
• Greater foot-lift height during swing (47 %).
• Reduced pathological compensations in the lumbosacral spine (53 %).

Thus, the MBD method proved more sensitive to individual rehabilitation effects, identifying improvements that remained undetected by the classical group-based approach.

Conclusion. Individualized MBD analysis provides an objective means of detecting clinically meaningful gait improvements overlooked by conventional group tests, supporting its integration into rehabilitation outcome assessment.