Performance-based tests in subjects with stroke: outcome scores, reliability and measurement errors.

Area of Science:

• Rehabilitation Medicine
• Neurology
• Biomedical Engineering

Background:

• Stroke survivors often experience impaired motor function, necessitating reliable assessments.
• Performance-based tests are crucial for evaluating functional recovery in stroke patients.
• Understanding the reliability of different outcome scoring methods is vital for clinical practice.

Purpose of the Study:

• To evaluate the intra- and inter-rater reliability of seven common performance-based tests in stroke subjects.
• To determine the impact of different outcome score calculations (e.g., single trial, mean of trials) on reliability and measurement error.
• To identify the optimal number of trials for consistent and accurate assessment of functional performance in stroke.

Main Methods:

• A reliability study involving two independent examiners evaluating sixteen stroke subjects.
• Seven performance-based tests were administered over two sessions, seven days apart.
• Intra-class correlation coefficients (ICCs) and standard errors of measurement (SEM%) were used to analyze reliability and error.

Main Results:

• All seven tests demonstrated significant and adequate intra- and inter-rater reliability across all outcome score types.
• No significant differences in score values, reliability, or measurement errors were found between single trial, mean of trials, or best/worst trial scores.
• Measurement errors (SEM%) were consistently low, ranging from 5.01% to 14.78%.

Conclusions:

• A single trial is sufficient for obtaining reliable and consistent results for the evaluated performance-based tests in stroke subjects.
• The choice of outcome score calculation (single vs. multiple trials) does not significantly affect the reliability or measurement error of these tests.
• These findings simplify the assessment process for functional performance in stroke rehabilitation, enhancing clinical efficiency.