Using a logarithmic model to predict functional independence after spinal cord injury: a retrospective study

Area of Science:

• Neuroscience
• Rehabilitation Medicine
• Biostatistics

Background:

• Predicting functional independence after spinal cord injury (SCI) is crucial for patient management and rehabilitation planning.
• Existing models may lack simplicity or accuracy across diverse SCI populations.
• A need exists for a straightforward tool to estimate long-term functional outcomes.

Purpose of the Study:

• To develop and validate a simple logarithmic model for predicting functional independence in adults with traumatic spinal cord injury (SCI).
• To assess the accuracy and agreement of the model's predictions against actual patient scores.
• To evaluate the model's utility in a clinical setting.

Main Methods:

• A retrospective cohort study involving 31 adults with traumatic spinal cord injury (SCI).
• Spinal Cord Independence Measure (SCIM) III scores at 1 and 3 months post-admission were used.
• A logarithmic model was constructed and validated using linear mixed models and Bland-Altman analysis.

Main Results:

• The logarithmic model demonstrated a non-linear correlation with actual SCIM III scores, which improved upon excluding Sphincter Management scores.
• Linear mixed model analysis showed a fixed effect slope of 1.12 [0.89, 1.36] and an intercept of -1.64 [-4.95, 1.68] after adjustment.
• Bland-Altman analysis indicated no significant bias in the total SCIM III score predictions, with most scores within limits of +15.3 and -19.3.

Conclusions:

• The developed logarithmic model accurately predicts functional independence scores in individuals with diverse SCI.
• The model's simplicity makes it a potentially valuable tool for daily clinical practice.
• Further validation in larger, diverse cohorts is warranted.