Comparison of Activity Monitor Output in Free Living Conditions

Abstract

PURPOSE: To compare Actigraph GT1M (AG) to Kenz Lifecorder EX (KL) outputs of steps counts, activity energy expenditure (EE), total EE, time spent in moderate (MOD) physical activity (PA), vigorous (VIG) PA and the combined classification of MOD+VIG PA during free-living conditions. METHODS: Ten male and 10 female participants wore three activity monitors on their waistline during all waking hours for one day. The AG was placed on the mid- axillary line on the right hip. To assess potential side differences in KL output, the KL was placed at the anterior midline of the right thigh (KL-R), and a second KL was placed at the anterior midline of the left thigh (KL-L). Manufacturer default thresholds for the KL were used to identify MOD (4-6) and VIG (7-9) PA. Freedson, Combo and WET formulae were used to estimate activity EE using the AG. Time spent in various intensity classifications was compared between the KL-R and KL-L to five established AG based intensity derivations (Freedson (1998), Nichols (2000), Swartz (2000), Hendelman1 (2000), and Hendelman2 (2000)). Paired sample t-tests and repeated measures ANOVA were used to analyze the data. RESULTS: KL-R and KL-L produced similar outputs for steps, EE and time within various PA intensity derivations. The AG counted significantly fewer steps compared to KL-L (p=0.002). KL-R, KL-L and AG-Freedson yielded significantly lower activity EE estimates compared to AG-WET and AG-Combo (p<0.005). The KL-R and KL-L yielded significantly lower estimates of time spent in MOD PA compared to AG-Freedson, AG-Nichols, AG-Swartz, and AG-Hendelman1 intensity derivations (p<0.002). There were no significant differences between activity monitors' output for time spent in VIG PA. KL-R and KL-L produced significantly lower estimates of time spent in MOD+VIG PA compared to AG-Swartz and AG-Hendelman1 intensity derivations (p<0.002). CONCLUSION: The KL produces similar output when placed on the right and left anterior aspects of the waistline. The KL tended to overestimate step counts compared to the AG and underestimate daily activity EE compared to established AG formulae. Overall, the KL yielded lower estimates of time spent in MOD and MOD+VIG PA compared to some established AG intensity derivations, however the KL produced similar estimates of VIG PA compared to all AG intensity derivations.