Litcius/Paper detail

Fitbit Physical Activity and Sleep Data in the All of Us Research Program: Data Exploration and Processing Considerations for Research

Caitlin P. Bailey, Kevin W. Dodd, James J. McClain, Isabell Seo, William Wheeler, Dana L. Wolff‐Hughes

2025Medicine & Science in Sports & Exercise14 citationsDOIOpen Access PDF

Abstract

PURPOSE: The All of Us Research Program is a national effort to enroll ≥1 million participants for health research. The program enables participants to donate Fitbit data, providing a unique dataset for physical activity (PA) and sleep research. This study characterizes Fitbit device data in the All of Us cohort version 8 (v8) data release to facilitate use by researchers. METHODS: Data were inspected and quality control guardrails were implemented. Days 15 to 21 (inclusive) postconsent date were selected for analysis of demographic characteristics, wear days, and wear time proxy variables. RESULTS: On days 15 to 21 postconsent, the v8 All of Us Fitbit cohort consisted of n = 30,445 participants with PA and/or sleep summary data from 160,487 person-days. Almost all participants (99%) provided both PA and sleep data. The sample consisted of 57% bring-your-own-device participants and 43% Wearables Enhancing All of Us Research (WEAR) participants provided with a device. The majority of participants were non-Hispanic White (66%) and female (68%). Seventy percent of the sample had ≥5 d of summary data available. Using heart rate data as a proxy for wear, participants averaged 1310 ± 170 min·d -1 of wear time on wear days, or roughly 21 h and 50 min. The Fitbit Charge series were the most common devices worn in the cohort. CONCLUSIONS: Considerations for working with All of Us Fitbit data (v8) are discussed and include rolling dates of primary consent, date shift implementation, proprietary Fitbit algorithm, valid day methodology, missing data assumptions, photoplethysmography technology, device version data, and population representation.

Topics & Concepts

Proxy (statistics)CohortWearable computerData collectionWearable technologyMedicineCohort studySample (material)Sleep (system call)Physical therapyInformed consentGerontologyComputer scienceAlternative medicineStatisticsMathematicsEmbedded systemChromatographyMachine learningChemistryOperating systemInternal medicinePathologyPhysical Activity and HealthSleep and related disordersMobile Health and mHealth Applications