Predicting community walking after stroke

Area of Science:

• Neurology
• Rehabilitation Medicine
• Biomedical Engineering

Background:

• Personalized stroke rehabilitation requires accurate prediction of functional recovery, specifically community ambulation.
• Early identification of individuals likely to achieve community walking facilitates tailored intervention strategies.

Purpose of the Study:

• To assess the predictive value of patient characteristics, clinical assessments, and Inertial Measurement Unit (IMU) data for community walking 6 months post-stroke.
• To determine if measures at admission or discharge during rehabilitation offer stronger prediction of long-term walking outcomes.

Main Methods:

• Collected data from 35 stroke survivors, including IMU-based gait and balance tests, clinical assessments (e.g., Berg Balance Scale, Motricity Index), and patient demographics.
• Utilized univariate and multivariate ordinary least squares regression to analyze the predictive power of admission and discharge measures for community walking at 6 months.

Main Results:

• Age, gait features, and daily step count at admission and discharge significantly predicted 6-month step count.
• 2-minute walk test (2MWT) speed, Trunk Control Test (TCT), Motricity Index (MI), and baseline daily gait speed predicted 6-month average and maximum gait speeds.
• Admission measures demonstrated greater predictive value than discharge measures.

Conclusions:

• Age, trunk stability (TCT), affected leg strength (MI), and clinical/daily-life gait assessments are valuable predictors of community ambulation 6 months post-stroke.
• Findings suggest that early assessment data is more predictive of long-term outcomes.
• Larger sample studies are recommended to further validate these predictive models.