[Recording the daily physical activity of COPD patients with an accelerometer: An analysis of agreement and repeatability].

OBJECTIVE The purpose of this study was to assess the agreement between different measurements of mean daily physical activity taken over a week in chronic obstructive pulmonary disease (COPD) patients with an accelerometer and to analyze the medium-term repeatability of these measurements. PATIENTS AND METHODS The study enrolled 12 healthy control subjects and 23 patients with stable COPD (mean [SD] forced expiratory volume in 1 second [FEV1] of 45% [13%] of predicted and a ratio of FEV1 to forced vital capacity of 53% [13%]). Accelerometer output, measured in vector magnitude units, was recorded in a physical activity log for a 1-week period. The results were then analyzed to compare output for a conventional recording period (Friday to Sunday) to that for 2 other periods (Monday to Wednesday and Tuesday to Thursday). The measurements were repeated 3 to 5 weeks later. RESULTS Activity counts were lower in the COPD patients than in the control subjects (184 [99] vs 314 [75]; P < .001). In the COPD patients, the results for the Friday to Sunday period correlated well with the results for both the Monday to Wednesday period (95% confidence interval, -29.21 to 28.81) and the Tuesday to Thursday period (95% confidence interval, -32.13 to 28.43). There were no significant differences in terms of medium-term repeatability of accelerometer readings between the COPD group and the control group (repeatability coefficient of 11.2% [4.6%] and 8.5% [4.7%], respectively). CONCLUSIONS Both agreement between the different measurements of physical activity taken during a 1-week period and medium-term repeatability for COPD patients and control subjects were very good.