Skip to main content

Advertisement

Table 3 Characteristics and results of studies that examined reliability of ActiGraph models for measuring sedentary behaviour in older adults (mean age ≥ 60 years), ordered from largest to smallest sample size

From: Validity of objective methods for measuring sedentary behaviour in older adults: a systematic review

Study Participants and data source Monitor and epochs analysed Methods Results for Sedentary Behaviour
Kocherginsky, et al., 2016 [36] n = 2208 57% female (weighted) Mean age = 74.2 y (95% CI 73.7–74.7) USA: Data collected for National Health and Nutrition Examination Survey (NHANES) ActiGraph 7164 Worn on right hip 60-s epochs Free-living Activities: VA < 100 cpm Observation period: 7 consecutive days during waking hours Valid hours and days: ≥10 h; ≥1 day Non-wear algorithm: 60 min of consecutive zeros, no allowance for interruptions Analysis: used linear regression to examine variations across days of the week; computed Lin’s concordance coefficients to compare 2-day and then 3-day averages to 7-day average among participants with 7 valid days of data Average daily percent of time spent in SB Monday to Friday: 65.3–65.9% Saturday: 66.3% Sunday: 69.6% Difference between Sunday and Monday to Saturday was significant (p < 0.001). Difference between Saturday and Monday to Friday was significant (p = 0.045). Comparison of % time spent in SB between 2 & 3 day averages with 7-day average Lin’s concordance r: For 2-day vs 7-day: 0.91 For 3-day vs 7-day: 0.94
Keadle et al. (2017) [27] n: 209 Only females Mean age: 70.6 ± 5.7 y USA: Data collected for an observational ancillary study of participants from the Women’s Health Study, a randomized trial of aspirin and vitamin E to reduce risk of cardiovascular disease and cancer. Data collected after completion of the trial. ActiGraph GT3X+ Worn on hip 60-s epochs Free-living Activities: VM < 200 cpm Observation period: two to three 7-day periods over 2–3 y during waking hours Valid hours and days: ≥10 h; ≥4 day Non-wear algorithm: Choi algorithm [61] and ≥ 600 min/day Analysis: computed reproducibility of sitting time across time periods ICCs; used linear mixed models; assessed utility of one 7-day assessment for classifying 2–3 year behaviour by cross-classifying participants using the baseline quartile distribution for SB and the quartile distribution of the average of two follow up assessments ICCs (95% CI) over 2–3 years All participants: 0.75 (0.69, 0.80) Younger: 0.74 (0.66, 0.81) Older: 0.74 (0.65, 0.81) Normal weight: 0.73 (0.65, 0.80) Overweight: 0.76 (0.68, 0.83) Less active: 0.75 (0.67, 0.82) More active: 0.64 (0.54, 0.73) Percent agreement in classification of SB into same quartile at baseline and average of follow-up assessments: 50 and 7% misclassified by ≥2 quartiles
Wanner et al., 2013 [32] n = 65 32 males, 33 females (50.8%) Mean age = 60.8 ± 9.9 y Switzerland: Data collected for ancillary study of the Swiss Cohort Study on Air Pollution and Long and Heart Disease in Adults, after completion of the main study. ActiGraph GT3X Two worn on right hip 60-s epochs Free-living Activities: VA < 150 cpm, <  100 cpm and < 200 cpm Observation period: 8 consecutive days Valid hours and days: not reported Non-wear algorithm: 60 min of consecutive zeros, no allowance for interruptions Analysis: compared normal filter to low-frequency extension (LTE) filter using Spearman correlations, Wilcoxon rank sum tests, scatter plots, and Bland–Altman plots; used linear regression to compute correction factors in half the sample and re-analyse results using correction factor NORMAL VS LTE FILTER FOR VA < 150 CPM Non-wear time Spearman r: 0.97 Mean difference: 8.9 ± 13.3; 1.5% ± 2.2%, p < 0.001 Sedentary time (min/day) Spearman r: 0.96 Mean difference: 25.7 ± 17.6; 4.5% ± 3.1%, p < 0.001 Other findings Results for mean differences did not change if cut-point changed to < 100 or < 200 cpm for SB. Plots showed non-wear time and SB time were systematically lower for low-frequency extension vs normal filter. CORRECTION FACTOR FOR LFE FILTER FOR VA < 150 CPM Nonwear min/day: 2.996 + (1.01 x nonwear time from LFE) Sedentary min/day: 62.74 + (0.93 x sedentary time from LFE) COMPARE NORMAL VS LTE USING CORRECTION FACTORS FOR VA < 150 CPM Non-wear time: Mean difference: − 0.8 ± 9.1; − 0.2% ± 1.5, p = 0.30 Sedentary time (min/day): Mean difference: 0.1 ± 15.6; − 0.1% ± 2.7%, p = 0.72
Hart et al., 2011 [35] n = 52 13 males; 39 females Mean age: 69.3 ± 7.4 y USA: Data collected from participants of larger ongoing study of physical activity patterns. ActiGraph 7164 Worn on right waist 60-s epochs Free-living Activities: VA ≤50 cpm Observation period: 21 consecutive days during all waking hours Valid hours and days: not reported Non-wear algorithm: 60 min of consecutive zeros, no allowance for interruptions Analysis: computed reproducibility of sitting time using Spearman-Brown Prophecy Formulas based on ICC; computed RMANOVA to examine differences between days of the week Number of days of measurement required for: ICC = 0.80: 5 days ICC = 0.85: 7 days ICC = 0.90: 11 days ICC = 0.95: 21 days No significant differences between days of week in time spent in SB (p = 0.48)
  1. Abbreviations: cpm Counts per minute, IQR Inter-quartile range, ICC Intraclass correlation coefficient; valid hours and days: for free-living studies lasting at least 7 days, number of hours per day and days during observation period that were required for data to be included in analysis; VA Vertical axis, VM Vector magnitude, m Minutes, s Seconds, h Hours, y Years