A historic dataset of wearable fitness tracker data from more than 59,000 Americans is giving researchers an unprecedented window into how daily activity and sleep patterns connect to chronic disease risk. The National Institutes of Health's (NIH) All of Us Research Program released Fitbit data spanning 14 years, containing 39 million step observations and 31 million sleep observations. This represents one of the largest and most demographically diverse wearables datasets ever made available for biomedical research.

The breakthrough comes from a deliberate effort to fix a major problem in digital health research: most wearables data has historically come from white, wealthy, highly educated people. This gap meant researchers didn't fully understand how activity patterns and health outcomes vary across different populations. The All of Us Research Program addressed this by distributing free Fitbit devices to invited participants nationwide through its Wearables Enhancing All of Us Research (WEAR) study, launched in February 2021.

Why Does Demographic Diversity in Health Data Matter?

Commercial wearable devices are now owned by 20 to 45 percent of Americans, yet most digital health research has disproportionately represented white individuals with higher educational and income levels. This creates a blind spot: researchers don't know whether the health insights they discover from wearables apply equally to everyone.

The WEAR study deliberately recruited participants with varying backgrounds and health needs. When researchers compared the demographics of participants who donated data through the free device program (WEAR study) versus those who brought their own devices (BYOD program), the differences were striking. For example, 55.1 percent of BYOD participants identified as white, compared to just 46 percent in the WEAR study. The WEAR cohort also included more people with lower household incomes, less formal education, and those reporting inadequate access to healthcare.

What Data Are Researchers Now Able to Study?

The expanded dataset enables researchers to examine patterns of physical activity and sleep across numerous population groups and link wearables data with other rich health information. Nearly half of the participants with Fitbit data also contributed electronic health records, physical measurements, genomic information, and survey responses. This multi-layered approach allows scientists to study relationships between digital health metrics and clinical outcomes in ways that weren't possible before.

The geographic reach is also significant. Participants come from all 50 states, with the highest concentrations in states like California, Wisconsin, Pennsylvania, and Illinois where the All of Us Research Program has partnerships with large healthcare organizations.

How to Use Wearable Data for Health Research

• Link Activity Data to Health Records: Researchers can now connect step counts and sleep patterns from wearables directly to electronic health records, allowing them to identify which activity levels correlate with lower risk of chronic diseases.
• Study Diverse Populations Separately: By including participants from different racial, ethnic, income, and educational backgrounds, researchers can determine whether health recommendations based on activity levels apply equally to everyone or need to be tailored.
• Examine Long-Term Trends: With 14 years of data spanning millions of observations, scientists can track how changes in daily steps and sleep duration over time predict future health outcomes.

The All of Us Research Program's mission is to address historical gaps in research experience and advance precision health for all, particularly those from populations with unique life experiences and health needs, such as older adults, rural populations, and individuals with less access to healthcare.

This dataset represents a shift in how biomedical research is conducted. Rather than studying health in a narrow slice of the population and assuming findings apply universally, researchers now have the tools to understand how digital biomarkers manifest across different demographic groups. The framework also includes methodological considerations for responsible use, helping researchers maximize both the scientific and societal impact of the data.

The implications are substantial. As wearable devices become more common, understanding how activity and sleep data relate to disease risk across diverse populations could lead to more personalized health interventions and better prevention strategies tailored to different communities. The All of Us Research Program has made this dataset available to registered researchers, opening the door to investigations that could reshape how we understand the connection between daily habits and long-term health.