Reliability of physical activity measured using triaxial accelerometer and the relationship between daytime activity and therapy-time activity in stroke inpatients

Abstract

Background: Increased physical activity (PA) during not only therapy-time, but also daytime, is important to improve functioning and activities of daily living in stroke inpatients. However, the methods of measuring PA levels in inpatients are not well established, and PA for stroke inpatients has not been described. Purpose: The purpose of this study was to assess the reliability of PA level measured using a triaxial accelerometer and to investigate the relationship between daytime activity (DA) and therapy-time activity (TA) in stroke inpatients. Methods: We recruited 17 inpatients who had experienced stroke a minimum of 1 month before. Each patient wore the triaxial accelerometer (OMRON Active Style Pro HJA-350IT), attached to their waist, for 14 consecutive days; PA was recorded every 60 s between 7.00 am and 7.00 pm. Metabolic equivalents (METs) were used to assess the recorded PA. We investigated 2 types of PA; DA (PA from 7 am to 7 pm), and the TA defined as (the PA during rehabilitation therapy-time, including physical therapy, occupational therapy, and speech therapy). The METs for DA and TA were classified into 3 different intensity categories: sedentary behavior [SB], 1.0-1.5 METs; light-intensity physical activity [LIPA], 1.6-2.9 METs; moderate-to-vigorous-intensity physical activity [MVPA], ≥3 METs. We calculated the total duration of PA in each of the 3 different categories for DA (DA-SB, DA-LIPA, and DA-MVPA) and TA (TA-SB, TALIPA, and TA-MVPA). Results: The triaxial accelerometer showed high test-retest reliabilities for all 6 categories of PA over 7 consecutive days (intraclass correlation coefficient, ICC (1,7) = 0.862-0.958). Furthermore, the accelerometer showed high test-retest reliabilities for all 6 categories over 2 consecutive weeks (ICC (1,2) = 0.913-0.979). There were significant differences among DA-SB (350.8±73.4 min), DA-LIPA (112.2±50.4 min), and DA-MVPA (9.5±12.6 min) (p < 0.05). There were also significant differences among TA-SB (100.0±21.7 min), TA-LIPA (49.5±23.6 min), MVPA (6.1±10.6 min) (p < 0.05). In the simple correlation analysis, DA-LIPA correlated with TA-LIPA (r = 0.807), and DA-MVPA correlated with TA-MVPA (r = 0.994). Conclusion(s): The triaxial accelerometer measured DA and TA with high reliability; the average DA and TA over 7 consecutive days may show the central tendency of DA and TA for stroke inpatients. DA was of extremely low intensity in stroke inpatients; DA-SB (related to static sitting activity) accounted for approximately 50% of all daytime, while the DA-MVPA, defined as≥3METsintensity activity, accounted for only approximately 2%. A simple correlations analysis showed a significant relationship between the moderate-to-high-intensity DA and TA, suggesting that TA may be an important factor in promoting light-to-high-intensity DA in stroke inpatients. Implications: This study indicates that measurement of PA using a triaxial accelerometer is highly reliable in stroke inpatients. High-intencity TA related to high-intencity DA was poor in inpatients, suggesting that it may be important to increase TA to improved DA in stroke inpatients.