A study of predictive validity, responsiveness, and minimal clinically important difference of arm accelerometer in real-world activity of patients with chronic stroke

Objective: To investigate the predictive validity, responsiveness, and minimal clinically important difference of arm accelerometer in real-world activity of patients with chronic stroke. Design: Validation and psychometric study. Setting: Three medical centers. Subjects: Patients with chronic stroke came from three separated randomized controlled trials. Interventions: Patients with stroke received upper extremity rehabilitation programs for four weeks. Main measures: Real-world arm movements were measured by an arm accelerometer and three clinical measurement toolsthe Motor Activity Log, Stroke Impact Scale, and Nottingham Extended Activities of Daily Livingadministered before and after treatment. Results: A total of 82 subjects were recruited in the study (mean age: 55.32years; mean score of Fugl-Meyer Assessment: 39.91). Correlations between the arm accelerometer and three clinical measurement tools were fair to moderate (Pearson's r=0.47, 0.42, and 0.34, respectively). The correlation between the arm accelerometer and the quality of use of Motor Activity Log subscale was moderate to good (Pearson's r=0.57). The responsiveness of the arm accelerometer from pretreatment to posttreatment was medium (standardized response mean=0.72). The minimal clinically important difference range for the arm accelerometer was 547-751 mean counts. Conclusion: The arm accelerometer demonstrated acceptable predictive validity and responsiveness in patients with chronic stroke. The affected arm activity measured by the arm accelerometer was sensitive to change. The change score of a patient with chronic stroke on the arm accelerometer should reach 574-751 mean counts to be regarded as a minimal clinically important difference.