Assessment of the Accuracy of a Wrist-Worn Activity Tracker for Measuring Sedentary Behavior in Cardiac Outpatients: A Validation Study (Preprint)

BACKGROUND Wearable device–based monitoring of sedentary behavior is increasingly used in digital cardiovascular care; however, the validity of consumer-grade wrist-worn trackers in older cardiac patients remains unclear. OBJECTIVE This study evaluated the agreement between sedentary time measured by a consumer-grade wrist-worn activity tracker (Fitbit Inspire 3) and a research-grade waist-worn accelerometer (ActiveStyle Pro) in Japanese cardiac outpatients under free-living conditions. METHODS Fifty-one outpatients with heart failure or ischemic heart disease (mean age 78.4 years) simultaneously wore both devices for 7 consecutive days. Sedentary time was assessed within a standardized 12-hour daytime window (09:00–21:00). Agreement was examined using Bland–Altman analysis, Lin’s concordance correlation coefficient (CCC), and Cohen’s kappa for tertile classification. RESULTS The Fitbit systematically overestimated sedentary time by 62.1 minutes per day compared with the ActiveStyle Pro, with significant fixed bias but no proportional bias. Correlation between devices was strong (r = 0.905), while uncorrected concordance was moderate (CCC = 0.71). After applying a simple constant correction for the fixed bias, concordance improved substantially (CCC = 0.904). Tertile classification showed moderate agreement (κ = 0.588), with all misclassifications occurring between adjacent categories. CONCLUSIONS In older Japanese cardiac outpatients, a consumer-grade wrist-worn tracker systematically overestimated sedentary time by approximately 62 minutes/day compared with a research-grade waist-worn accelerometer, with fixed bias but no proportional bias. A simple constant correction markedly improved concordance (CCC ≈0.90), supporting calibrated use of consumer wearables for scalable remote monitoring and behavioral interventions in ambulatory cardiac care while underscoring the need for population-specific validation.