Physical activity measurement in people with spinal cord injury: comparison of accelerometry and self-report (the Physical Activity Recall Assessment for People with Spinal Cord Injury).

Purpose: To (1) evaluate the level of agreement between individually calibrated accelerometers and the self-reported Physical Activity Recall Assessment for People with Spinal Cord Injury when assessing moderate-vigorous physical activity; and (2) qualitatively examine the different components of physical activity each measure assesses. Materials/methods: Nineteen manual wheelchair users with chronic spinal cord injury (19.0 +/- 12.9 years post injury, C5-L2 injury level) wore a wrist and spoke accelerometer for one week then returned to the lab and completed the Physical Activity Recall Assessment for People with Spinal Cord Injury for their last 3 days of accelerometer wear. Results: Bland-Altman plots revealed low levels of agreement between the two measures when measuring total (bias = -5.6 +/- 70.41 min/d, 95% agreement limits = -143.6-132.4 min/d), wheeled (bias = -9.7 +/- 30.2 min/d, 95% agreement limits = -69.0-49.5 min/d), and non-wheeled (bias =12.3 +/- 53.8 min/d, 95% agreement limits = -93.1-117.6 min/d) moderate-vigorous physical activity. The accelerometer was beneficial for detecting intermittent or brief activities while the self-report measure was useful for measuring lifting activities (e.g., resistance training) and wheeling on inclined surfaces. Conclusion: Total and wheeled moderate-vigorous physical activity measured by an accelerometer and a self-report measure showed low agreement at the individual level. Additional research is needed to examine whether physical activity may be best measured using accelerometers and a self-report measure concurrently.