Defining physical activity: A comparison of actigraphy monitoring and self‐report measures

Abstract

AbstractBackgroundPhysical activity (PA) helps preserve age‐related brain health. However, PA is a broad construct and its operationalization can vary. Actigraphy monitors are gold‐standard measures of PA, but not always feasible. The Community Healthy Activities Model Program for Seniors (CHAMPS) and Physical Activity Scale for the Elderly (PASE) are widely‐used self‐report questionnaires of PA for older adults, but are limited by scope and subjectivity. We aimed to examine the ecological validity of objective and subjective measures of PA and determine how self‐reported PA relates to distinct actigraphy metrics.Method104 functionally normal older adults completed 30‐day Fitbit monitoring (average daily steps and calories burned), and reported PA levels via PASE (total activity) and CHAMPS‐METs (metabolic expenditure calories burned). Participants also completed measures of vascular (hypertension diagnosis, resting heart rate), neurobehavioral (cognition, Perceived Stress Scale, Everyday Cognition Scale), and brain MRI (medial temporal lobe volume) outcomes. Spearman’s rank correlations and regression models evaluated the associations among PA measures and with outcomes of interest.ResultGreater Fitbit step count associated with higher PASE (r=0.30, p=0.005), but not CHAMPS‐MET scores (r=0.15, p=0.20). On the other hand, Fitbit calories burned related to higher CHAMPS‐METs (r=0.35, p=0.003), but not PASE scores (r=0.16, p=0.15). Only Fitbit measures demonstrated expected positive associations with cognitive outcomes, with slightly stronger associations shown with step count (r range=0.23‐0.26). All 3 PA measures demonstrated expected negative associations with perceived stress and cognitive complaints (r range=‐0.07‐0.26). Greater PA, as measured by Fitbit metrics or CHAMPS‐MET, was associated with less hypertension (Fitbit steps t=0.15; CHAMPS‐METs t=‐2.04) and lower resting heart rate (Fitbit calories burned r=‐0.26). PASE did not significantly associate with vascular health (HTN t=0.93; HR r=‐0.02, ps>0.05). Lastly, greater Fitbit step count (β=0.32, p<0.01) and calories burned (β=0.29, p<0.01), but not PASE or CHAMPS‐MET (β range=0.05‐0.15, ps>0.05), were associated with larger medial temporal lobe volumes.ConclusionObjective indicators of PA demonstrate greater overall ecological validity compared to self‐report. Our findings suggest PA is comprised of several constructs. Metrics capturing PA intensity via calories burned may not be interchangeable with metrics of overall movement. Careful understanding what PA metrics represent will facilitate better‐informed recommendations for healthy aging.