# Improvements in blood and fitness tracker biomarkers in a longitudinal real-world cohort of digital health platform users

## TL;DR

A digital health platform helped users improve blood biomarkers like LDL cholesterol over time, with results linked to lifestyle changes and genetic predispositions.

## Contribution

Large-scale, real-world evidence that integrating blood, wearables, and genetics can guide personalized actions to maintain healthier biomarker levels.

## Key findings

- Users with suboptimal LDL cholesterol and HgbA1c showed sustained improvements over years of platform use.
- Increased daily steps and higher REM sleep correlated with healthier cholesterol levels.
- Higher polygenic risk for traits like elevated LDL cholesterol was associated with smaller biomarker improvements.

## Abstract

Digital health technologies offer new opportunities for personalized health management and disease prevention. In this retrospective, long-term longitudinal study of over 20,000 users of a digital health platform (DHP), we aimed to determine whether improvements in health-related biomarkers could be observed in users and, if so, how they were maintained over time. We further explored whether genetic predisposition and physiological patterns, such as sleep and activity, were associated with variability in these biomarker responses. The DHP evaluates a user’s individual biological profile, consisting of blood biomarkers, polygenic risk scores (PGS), and fitness tracker data, and provides personalized lifestyle interventions based on knowledge about nutrition, supplements, exercise, and recovery collected from over 7,000 clinical studies. Here, we show improvement in suboptimal levels of the primary outcome, key blood biomarkers that are sustained or increased in the long-term with DHP use. We additionally show the correlation of biomarker improvement with secondary outcomes, including specific sleep and activity patterns in users. Lastly, we find significant correlations between polygenic risk and both baseline levels and longitudinal change in biomarkers, including low-density lipoprotein cholesterol (LDL-c), suggesting that genetic predisposition for a negative trait (e.g., elevated LDL-c) could make it more difficult to improve that trait. This longitudinal, integrated biomarker dataset highlights the potential of digital health tools in fostering improvements in health-related biomarkers through personalized data analytics and targeted behavioral interventions.

Digital health applications promise to turn personal data into improved health and longevity, but large-scale, long-term evidence has been limited. We retrospectively analyzed multi-year data from more than 20,000 adults using a digital health platform that combines blood tests, wearable data (activity and sleep), and genetic information to give personalized lifestyle guidance about food, supplements, exercise, and recovery. We found that users who started with suboptimal blood biomarkers such as LDL cholesterol and HgbA1c showed meaningful improvements that were sustained across years and multiple follow-up tests. Data from wearables provided information on who improved: users who increased their daily steps by an average of ~1,000 steps from baseline to follow-up activity levels, and those with higher REM sleep percentages, were more likely to shift cholesterol in a healthier direction, despite substantial month-to-month variability in step counts. Genetics also characterized biomarker improvement: individuals with a higher inherited risk for certain traits (for example, high LDL) tended to show smaller improvements, suggesting a genetic influence on how much people benefit from lifestyle changes. Our results provide large-scale, real-world evidence that integrating blood, wearables, and genetics can guide practical, personalized actions and may help people maintain healthier biomarker levels over time.