Identifying Postural Instability in Children with Cerebral Palsy Using a Predictive Model: A Longitudinal Multicenter...

Area of Science:

• Neurology
• Pediatrics
• Rehabilitation Medicine

Background:

• Children with cerebral palsy (CP) often experience significant challenges with postural control and trunk instability.
• Truncal tone (TT) impairments, including spastic or hypotonic types, are common and impact functional abilities.

Purpose of the Study:

• To develop and validate a predictive model, TT-PredictMed, for identifying factors associated with spastic and hypotonic truncal tone in children with CP.
• To improve the understanding of the complex interplay of factors contributing to postural impairments in this population.

Main Methods:

• A longitudinal, double-blinded, multicenter descriptive study involving 102 teenagers with CP.
• Development of a multiple logistic regression model (TT-PredictMed) following "Transparent Reporting of a multivariable prediction model for Individual Prognosis or Diagnosis" guidelines.
• Collection of clinical and functional data from 2006 to 2021.

Main Results:

• Predictors of hypotonic TT included hip dysplasia, postnatal etiology, male gender, and poorer manual and gross motor function.
• Predictors of spastic TT included neuromuscular scoliosis, prenatal etiology, specific spasticity patterns (quadri/triplegia), dystonia, and refractory epilepsy.
• The TT-PredictMed model demonstrated an average accuracy, sensitivity, and specificity of 82%.

Conclusions:

• The TT-PredictMed model effectively identifies key factors associated with hypotonic and spastic truncal tone in children with CP.
• These findings contribute to a better clinical understanding and management of postural instability in pediatric CP.
• The model's accuracy supports the application of machine learning in clinical prognosis for CP.