When researchers at NORD Rare Disease Center of Excellence Vanderbilt University Medical Center set out to understand how pulmonary arterial hypertension (PAH) affects patients’ daily lives, they turned to the NIH All of Us Research Program.
A study published in Pulmonary Circulation demonstrates how population-scale datasets, like All of Us, can strengthen rare disease research by providing well-matched comparison groups for analyses that would otherwise be limited by small sample sizes. Using wearable device data and matched controls from the All of Us study, Vanderbilt collaborated with institutions across the United States to examine physical activity, sleep, and quality of life in individuals with PAH. The study reflects a broader effort to integrate digital health data into the study of rare diseases, where traditional clinical measures often capture only a limited view of disease progression.
Extending rare disease research beyond the clinic
Rare diseases such as PAH are typically studied using disease-specific cohorts and clinical metrics such as a 6-minute walk distance test. While these approaches are essential, they often rely on episodic measurements collected during clinic visits. Measures such as the 6-minute walk distance provide important information about functional capacity, but they have known limitations and do not always capture how patients function in their daily environments. In this study, investigators used wearable devices to collect real-world longitudinal data on physical activity and sleep over a 12-week period in 110 individuals with PAH, with 44 participants followed longitudinally over one year. To interpret these findings, they constructed a matched control cohort using All of Us participants matched on age, sex, body mass index, and Fitbit monitoring dates.
Dr. Evan Brittain, the study’s corresponding author and a cardiologist at Vanderbilt University Medical Center, emphasized the value of this approach: “The All of Us Research Program offers a rare combination of scale, diversity, and multimodal data that is difficult to replicate in traditional cohorts. For conditions like PAH, where single-center datasets are inherently limited, it provides an opportunity to study risk, early signals, and heterogeneity in a broader population context. More broadly, we were motivated by the ability to link clinical data with wearable and behavioral data to better understand disease trajectories outside of the clinic.”
By incorporating All of Us data, the study was able to compare activity and sleep patterns in PAH with those observed in a broader population, providing important context for interpreting wearable-derived measures.
Detecting differences in activity and sleep
The analysis showed that individuals with PAH had substantially lower levels of daily activity compared to matched controls. PAH patients averaged approximately 5,200 steps per day at baseline, compared to 7,369 steps in matched controls — a gap that persisted at one-year follow-up. Minutes of moderate-to-vigorous physical activity were also markedly lower in PAH patients (8.9 minutes per day versus 31.7 minutes in controls). Daily step count correlated strongly with 6-minute walk distance and with patient-reported quality of life, indicating that wearable-derived activity captures meaningful aspects of disease burden.
Sleep patterns also differed significantly between groups. Individuals with PAH spent considerably less time in rapid eye movement (REM) sleep than matched controls at baseline (17.3% versus 21.6%), and this gap persisted at one year. Longitudinal follow-up showed that sleep quality worsened over time in PAH patients, with the percentage of light sleep increasing and REM sleep declining — changes that were statistically significant compared to controls. Over the same period, daily step counts fell from 5,200 to 4,651.
These findings highlight the value of continuous, real-world data in capturing aspects of disease progression that may be difficult to capture through traditional clinical assessments alone.
From continuous monitoring to earlier detection
The ability to measure activity and sleep over time raises the possibility of identifying early changes in health status before they are clinically apparent. This concept is particularly relevant for rare diseases, where progression may be gradual and difficult to quantify using standard endpoints.
Dr. Brittain noted that this work has reshaped his team’s research agenda: “One of the most important insights has been the potential to detect early, preclinical changes — what we think of as a ‘digital prodrome’ — using wearable and behavioral data. This has shifted part of our focus toward identifying early risk states and trajectories rather than studying disease only after diagnosis. It has also opened new directions in integrating consumer-generated health data with traditional clinical data across both common and rare conditions.”
More broadly, this work suggests that wearable data may serve as a complementary source of longitudinal phenotyping, helping researchers better understand how patients function in daily life and how those patterns change over time.
Using All of Us as a comparator resource
For rare disease research, one of the most consistent challenges is the lack of appropriate comparison populations. This study illustrates how All of Us can help address that gap by providing access to large-scale datasets with linked wearable, clinical, and demographic data.
On working within the platform, Dr. Brittain observed that “the platform is well designed and user-friendly, particularly for investigators working in cloud-based environments. The ability to work within a secure, centralized workspace with curated datasets is a major strength. They’ve also built strong support mechanisms and shared code resources that make it easier to get started.”
Rather than replacing disease-specific cohorts, population-based datasets such as All of Us can complement them, enabling analyses that place rare disease findings in a broader population context.
Implications for rare disease research
This study demonstrates that integrating wearable data with population-scale cohorts can provide new insights into disease burden and progression in rare conditions. By combining continuous behavioral monitoring with rigorously matched comparison groups, researchers can begin to characterize changes in function and health status that are not captured through traditional clinical methods.
For rare disease investigators, this approach offers a framework for incorporating digital measures and external comparator datasets into research design, to better characterize and contextualize disease progression.
Accessing All of Us data for rare disease research
All of Us is building one of the largest and most robust, multimodal health datasets available for research, with more than 880,000 participants enrolled and hundreds of thousands contributing genomic, clinical, and behavioral data. The forthcoming 2026 release of Curated Data Repository Version 9 (CDRv9) expands the cohort to more than 751,000 participants and substantially broadens available data types. This year, researchers will be able to access large-scale genomic data on more than 535,000 participants, as well as newly available tens of thousands clinical notes with natural language processing-derived concepts, along with a significant increase in digital health data. Wearable data from Fitbits will increase from 59,000 participants to approximately 68,000 participants — the largest publicly available Fitbit dataset — with Apple HealthKit data coming later this year.
“Rare disease research has always faced a fundamental numbers problem: too few patients, too little data, too few comparison populations. All of Us was designed to change that equation. The fact that nearly 23,000 researchers across the world are already using our data and have generated more than 1,350 peer-reviewed publications tells us the scientific community sees what’s possible here,” said All of Us CEO Josh Denny, M.D., M.S.
Scientists are using the program’s genomic and health history data to uncover rare gene‐disease links more reliably than before. Researchers are studying neurofibromatosis, cystic fibrosis, autoimmune diseases, Ehlers-Danlos syndrome, sickle cell disease, and many other conditions.
Researchers interested in conducting similar analyses can apply for access to the All of Us Researcher Workbench at researchallofus.org. The Workbench provides cloud-based tools for analyzing genomic data, electronic health records, physical measurements, and survey responses from a cohort that continues to grow.
This blog post is funded by the Division of Engagement and Outreach, All of Us Research Program, National Institutes of Health. Pyxis Partners Award Number: 1OT2OD038104-01
