Agreement of Activity Counts and Minutes of Moderate to Vigorous Physical Activity Assessed with Actical and Actigraph Accelerometers

Abstract

PURPOSE: The purpose of this study was to evaluate the agreement between Actical and Actigraph accelerometers for assessment of activity counts and minutes of moderate to vigorous physical activity (MVPA). A second purpose was to develop a cut point for the Actical that is equivalent to the Freedson et al. (1998) cut point for the Actigraph. METHODS: Physical activity of 33 participants (mean age = 19.7 ± 1.3 yrs) was assessed for one complete day that included a structured exercise session with Actical and Actigraph accelerometers. Mean counts/min were calculated for both accelerometers and compared. Minutes of MVPA were calculated for the Actigraph for each participant using the Freedson et al. cut point of 1,952 counts/min. MVPA for the Actical was calculated using a cut point developed from Heil (2006). Regression analysis was then used to estimate the Actical cut point equivalent to an Actigraph cut point of 1,952 counts/min. Paired t-tests and Bland-Altman plots were used to examine agreement among the different MVPA cut points. RESULTS: Raw activity counts from the Actical and Actigraph (mean = 309 ± 135 vs. 405 ± 151 counts/min, respectively) were highly correlated (r = 0.88), although mean activity counts were higher (p < .05) for the Actigraph. Cut points were evaluated by comparing minutes of MVPA recorded by the Actical and Actigraph accelerometers. MVPA from the Heil equation cut point was 40 min greater than MVPA from the Actigraph cut point (p < .05) and the correlation was .79. Bland-Altman plots demonstrated a significant mean bias and a large scatter of residuals (95%CI = −83 to 4 counts/min). The Actigraph cut point of 1,952 was found to be equivalent to an Actical cut point of 1,527 counts/min. Using this cut point for the Actical, mean minutes of MVPA for the Actical and Actigraph did not differ (p > .05; 52 ± 28 vs. 54 ± 31) and the correlation was r = 0.95. Bland-Altman plots demonstrated no mean bias and a much smaller 95% CI (−16 to 21 counts/min). CONCLUSIONS: Activity counts from the Actical and Actigraph were highly correlated over one day. The Actical cut point from the Heil equation did not provide similar estimates of MVPA to the Freedson et al. Actigraph cut point. The cut point of 1,527 counts/min developed in this study for the Actical provided similar estimates of MVPA to the Freedson et al. Actigraph cut point.