The recent partnership between Oura and Eli Lilly has raised considerable attention across the healthcare industry. At first glance, it looks like another collaboration between a technology company and a pharmaceutical manufacturer. However, looking deeper, it signals a broader transformation that could introduce new aspects on how patients are monitored throughout their treatment journey.
Consumer wearables have progressed rapidly over the past decade. Devices that once focused primarily on step counting now continuously collect information on physical activity, heart rate, heart rate variability, sleep, body temperature, respiratory rate, and other physiological parameters. Millions of people generate large amount of data daily, creating an unprecedented source of longitudinal health information.
For pharmaceutical companies, this might represent an opportunity to complement traditional real-world data (RWD) with continuous patient-generated measurements. For researchers, it opens new scientific questions about data quality, interpretation, analytical methods, and clinical relevance. However, the greatest value of consumer-generated health data will depend less on the amount of information collected and more on our ability to transform these data into meaningful evidence.
A new generation of real-world data
Traditional sources of RWD include electronic health records (EHRs), administrative claims, disease registries, pharmacy databases, and laboratory results. These datasets describe patients during their interactions with the healthcare system. Every medical visit, prescription, hospital admission, or laboratory test contributes another piece to the patient's clinical history.
Wearable devices add a very different dimension. They provide continuous observations of patients during their daily lives, outside hospitals and clinics. Instead of isolated clinical measurements collected every few months, researchers can potentially analyse thousands of observations from the same individual over weeks, months, or even years.
This continuous perspective creates opportunities to potentially better understand disease progression, treatment response, recovery, behavioural changes, and patient adherence under real-world conditions.
The combination of healthcare-generated and consumer-generated data might then represent the next evolution of real-world evidence (RWE).
Why pharmaceutical companies are interested
The growing interest from pharmaceutical companies extends far beyond wearable technology itself. Continuous digital measurements can support multiple stages of the product lifecycle.
Clinical trials may benefit from remote patient monitoring, digital endpoints, and improved participant engagement. Wearable data can provide additional information between scheduled study visits, offering a more comprehensive picture of patient health over time.
After product approval, consumer-generated health data can contribute to post-marketing research by evaluating treatment effectiveness, monitoring recovery trajectories, studying medication adherence, and identifying patterns associated with adverse events.
The obesity field offers an excellent example. New GLP-1 receptor agonists influence body weight through complex behavioural and physiological adaptations. Physical activity, sleep quality, energy expenditure, recovery, and lifestyle changes all contribute to treatment outcomes. Wearables can capture many of these variables continuously, providing valuable context that traditional healthcare databases cannot offer.
As pharmaceutical companies increasingly invest in digital health partnerships, collaborations with wearable manufacturers may become more common across therapeutic areas including cardiology, metabolic diseases, neurology, respiratory medicine, and mental health.
More data require better interpretation
The expansion of wearable data brings exciting possibilities together with significant methodological challenges.
Modern wearable devices can generate millions of data points from a single individual. Heart rate may be measured every few seconds. Sleep metrics are updated every night. Activity patterns evolve throughout the day. Body temperature and heart rate variability fluctuate continuously.
The availability of these data creates opportunities for richer analyses, while simultaneously increasing the complexity of evidence generation.
Every physiological parameter reflects the combined influence of multiple biological and environmental factors. An increase in resting heart rate, for example, may coincide with infection, psychological stress, dehydration, poor sleep, intense physical training, environmental heat, medication changes, or normal biological variability. Similar patterns can emerge from very different underlying conditions.
Reliable interpretation requires context. Clinical history, behavioural information, environmental exposures, medication use, and individual baseline characteristics all contribute to understanding what a physiological signal actually represents.
This complexity highlights an important opportunity for epidemiologists and RWE scientists. The next generation of evidence will increasingly depend on robust analytical frameworks capable of transforming continuous digital signals into clinically meaningful information.
Data quality remains fundamental
High-quality evidence begins with high-quality data.
Consumer wearables have achieved remarkable technological progress; however, several methodological considerations deserve careful attention before incorporating these data into regulatory or clinical research.
Measurement accuracy varies across devices, sensors, firmware versions, and algorithms. Performance also differs depending on the physiological parameter being measured. Heart rate during resting conditions may achieve high accuracy, while energy expenditure, sleep staging, or stress metrics often involve greater uncertainty because they rely on proprietary algorithms.
Missing data represent another common challenge. Users remove devices for charging, discontinue their use temporarily, or replace them with newer models. Longitudinal datasets frequently contain interruptions that require appropriate analytical strategies.
Population characteristics also influence generalisability. Individuals who purchase wearable devices often differ from the broader population in socioeconomic status, education, health awareness, and lifestyle behaviours. These characteristics introduce selection bias that researchers should consider when designing studies and interpreting results.
Finally, many wearable outputs combine multiple physiological measurements into proprietary scores. Readiness Score, Recovery Score, Sleep Score, and similar metrics provide convenient summaries for consumers, but their underlying algorithms often remain undisclosed. These scores also influence user behaviour. People may modify their sleep habits, exercise intensity, recovery strategies, or daily routines in an effort to achieve a higher score rather than respond to their actual physiological needs. This behavioural feedback loop can alter the data being collected over time and should be considered when interpreting longitudinal wearable data. Researchers therefore need to understand exactly what each metric represents, how it is generated, and how user behaviour may evolve in response to these scores before incorporating them into scientific analyses.
Building the next generation of evidence
Consumer-generated health data create possibilities for pharmaceutical research, but their greatest contribution will come from rigorous methodological development.
Future research will require validation studies, standardised analytical approaches, clinically meaningful digital endpoints, and transparent interpretation frameworks. Integrating wearable data with EHRs, claims databases, patient-reported outcomes, genomics, and environmental information may provide a more comprehensive understanding of patient health than any single data source alone.
Artificial intelligence (AI) will likely play an increasingly important role by identifying longitudinal patterns across millions of observations. Human expertise, however, remains essential for defining clinically relevant questions, evaluating causal relationships, assessing bias, and translating statistical findings into healthcare decisions.
The Oura–Eli Lilly partnership may therefore represent more than a single business agreement. It reflects a broader movement towards continuous health monitoring and digital evidence generation. As collaborations between pharmaceutical companies and wearable manufacturers expand, consumer-generated health data may become an integral component of the evidence ecosystem.
