The accuracy of the 24-h activity recall method for assessing sedentary behaviour: the physical activity measurement survey (PAMS) project
The accuracy of the 24-h activity recall method for assessing sedentary behaviour: the physical activity measurement survey (PAMS) project
复制标题
DOI：
10.1080/02640414.2016.1161218
复制
发表时间：
2017-02-01
影响因子：
3.4
通讯作者：
Welk, Gregory J.
中科院分区：
文献类型：
--
作者：
Kim, Youngwon;Welk, Gregory J.
久坐行为(SB)已成为一种可改变的危险因素，但对SB的测量误差知之甚少。本研究的目的是确定24小时体力活动回忆(24PAR)相对于SenseWear臂章(SWA)在评估SB方面的有效性。每个参与者(n=1485)在随机选择的两天进行了一系列数据收集程序：佩戴全套SWA 24小时，然后在第二天完成电话管理的24小时活动，以回忆过去24小时的活动。计算总久坐时间(TST)的计算不包括报告或记录的睡眠时间。使用等价性检验来比较TST的估计值。等价性检验分析表明，与SWA(等价区：580.7和709.8minday(-1))相比，TST的24PAR(90%CI：443.0和457.6minday(-1))没有显著的等价性。Bland-Altman图表明，在24PAR和SWA之间，极度或最少久坐的个体提供了相对可比的久坐时间。超重/肥胖和/或年长的人比正常体重和/或更年轻的人更有可能低估坐着的时间。24PAR的测量误差因久坐时间和人口统计指标的不同而不同。这一证据为今后开发测量误差模型以校正自我报告的误差提供了依据。
Sedentary behaviour (SB) has emerged as a modifiable risk factor, but little is known about measurement errors of SB. The purpose of this study was to determine the validity of 24-h Physical Activity Recall (24PAR) relative to SenseWear Armband (SWA) for assessing SB. Each participant (n=1485) undertook a series of data collection procedures on two randomly selected days: wearing a SWA for full 24-h, and then completing the telephone-administered 24PAR the following day to recall the past 24-h activities. Estimates of total sedentary time (TST) were computed without the inclusion of reported or recorded sleep time. Equivalence testing was used to compare estimates of TST. Analyses from equivalence testing showed no significant equivalence of 24PAR for TST (90% CI: 443.0 and 457.6minday(-1)) relative to SWA (equivalence zone: 580.7 and 709.8minday(-1)). Bland-Altman plots indicated individuals that were extremely or minimally sedentary provided relatively comparable sedentary time between 24PAR and SWA. Overweight/obese and/or older individuals were more likely to under-estimate sedentary time than normal weight and/or younger individuals. Measurement errors of 24PAR varied by the level of sedentary time and demographic indicators. This evidence informs future work to develop measurement error models to correct for errors of self-reports.