Wearable Device Data in Clinical Evidence: Leveraging Real-World Data for Regulatory Submissions

The proliferation of wearable health devices has created an unprecedented opportunity to generate clinical evidence from continuous, real-world physiological measurements that were previously obtainable only through episodic clinical encounters or resource-intensive inpatient monitoring. Smartwatches that detect atrial fibrillation, continuous glucose monitors that track glycemic control around the clock, wrist-worn accelerometers that quantify physical activity and sleep patterns, adhesive biosensors that measure respiratory rate and skin temperature, and implantable cardiac monitors that record heart rhythm continuously for years, all of these technologies generate data streams that, when properly validated and analyzed, can provide clinical insights of a depth and continuity that traditional clinical measurements cannot match. For regulatory purposes, the question is no longer whether wearable device data can contribute to clinical evidence but rather how to establish the data quality frameworks, validation methodologies, study designs, and analytical approaches that ensure wearable-derived evidence meets the rigorous standards required for regulatory decision-making.

The FDA has made significant progress in developing the regulatory frameworks needed to evaluate and accept wearable-generated evidence in regulatory submissions. Through a series of guidance documents addressing digital health technologies for remote data acquisition, real-world evidence for medical devices, and the use of digital health technologies in clinical investigations, the agency has articulated an evolving set of expectations for how sponsors should validate wearable technologies, design studies that incorporate wearable data, and present wearable-derived evidence in regulatory submissions. These frameworks reflect the FDA’s recognition that wearable technologies offer genuine advantages for evidence generation, including the ability to capture endpoints that are more clinically meaningful than traditional surrogate measures, the potential to detect treatment effects with greater sensitivity through continuous monitoring, and the opportunity to generate evidence in real-world settings that more closely reflect how treatments are actually used.

This article provides a comprehensive guide to leveraging wearable device data for clinical evidence generation and regulatory submissions, addressing the technical validation requirements, study design considerations, data quality frameworks, and regulatory strategy decisions that determine whether wearable-derived evidence strengthens or undermines a regulatory submission.

The Wearable Data Revolution in Clinical Evidence

The transformation of wearable devices from consumer wellness products to clinical evidence generation tools represents a fundamental shift in how clinical data is conceptualized, collected, and analyzed. Traditional clinical evidence generation relies on measurements obtained at scheduled clinical visits, producing snapshots of patient status at discrete time points. Between visits, the patient’s physiological state, treatment adherence, symptom experience, and functional capacity remain largely unobserved. Wearable technologies fill this observational gap by providing continuous or near-continuous measurement streams that capture the full temporal pattern of physiological parameters, activity levels, and symptom manifestations.

From Episodic to Continuous Measurement

The shift from episodic to continuous measurement has profound implications for clinical evidence quality. A six-minute walk test performed at a clinic visit provides a single data point reflecting the patient’s functional capacity under specific conditions on a specific day. A wrist-worn accelerometer worn continuously for weeks provides thousands of data points reflecting the patient’s actual daily activity patterns, including the duration, intensity, and timing of physical activity; the frequency and duration of sedentary periods; sleep quality and duration; and day-to-day variability in functional capacity that may reflect disease fluctuations, treatment responses, or environmental factors. This rich temporal dataset enables analyses that are impossible with episodic measurements, including detection of subtle treatment effects that manifest as changes in daily activity patterns, identification of circadian rhythm disruptions that correlate with disease progression, and measurement of functional capacity in the patient’s natural environment rather than the artificial conditions of a clinical testing facility.

The Digital Biomarker Paradigm

The concept of digital biomarkers, physiological or behavioral measures derived from sensor data collected through digital devices, has emerged as a framework for translating raw wearable sensor data into clinically meaningful measurements. Digital biomarkers may directly parallel traditional clinical measurements, such as using a wearable photoplethysmography sensor to measure heart rate, or they may represent novel measurements that have no traditional analog, such as using accelerometer data to derive a continuous mobility score that integrates step count, gait speed, stride regularity, and movement complexity. The development and validation of digital biomarkers involves establishing the relationship between the sensor-derived measurement and the clinical construct it is intended to represent, demonstrating the analytical validity of the measurement, evaluating its clinical validity as a measure of the intended clinical concept, and assessing its utility as an endpoint in clinical studies or regulatory submissions.

FDA Regulatory Framework for Wearable-Generated Evidence

The FDA’s regulatory framework for evaluating wearable-generated clinical evidence draws on multiple guidance documents, policy statements, and regulatory precedents that together define the agency’s expectations for data quality, validation, and submission of wearable-derived evidence. Understanding this framework is essential for sponsors planning to incorporate wearable data into regulatory submissions.

The Real-World Evidence Framework for Devices

The FDA’s guidance on the use of real-world evidence to support regulatory decision-making for medical devices establishes the foundational framework for incorporating data from wearable devices and other real-world sources into regulatory submissions. The guidance describes the circumstances under which the FDA may consider real-world evidence, the characteristics of real-world data that support its use for regulatory purposes, and the methodological considerations for generating real-world evidence that meets regulatory standards. For wearable-generated data, this framework establishes that the data must be relevant to the regulatory question, generated through methods that ensure adequate quality and reliability, and analyzed using appropriate statistical methods that account for the specific characteristics and limitations of real-world data.

Digital Health Technology Guidance

The FDA’s guidance on digital health technologies for remote data acquisition in clinical investigations addresses the use of wearable devices, mobile applications, and other digital tools to collect clinical trial data outside traditional clinical settings. This guidance describes the FDA’s expectations for the validation of digital health technologies used in clinical investigations, the data management and quality assurance processes that should accompany digital data collection, and the documentation that should be included in regulatory submissions to support the reliability of digitally collected data.

Digital Health Technologies for Remote Data Acquisition

The FDA’s evolving framework for digital health technologies used in clinical investigations distinguishes between three categories of validation that must be addressed: verification that the technology accurately measures the physiological parameter it claims to measure, validation that the derived measurement is clinically meaningful for the intended purpose, and usability assessment confirming that the technology can be used reliably by the intended user population in the intended setting.

Verification and Validation (V&V) Framework

Verification for wearable devices involves demonstrating that the sensor technology accurately and reliably measures the intended physiological parameter under the conditions of intended use. For a wearable heart rate monitor, verification would demonstrate that the device accurately measures heart rate compared to a reference standard such as electrocardiography across the range of heart rates, activity levels, and skin types expected in the intended use population. For a wearable accelerometer used to measure physical activity, verification would demonstrate that the device accurately detects and quantifies movement patterns across the range of activities and intensities relevant to the clinical application.

Validation addresses whether the wearable-derived measurement is clinically meaningful for its intended use in the clinical investigation. A wearable device that accurately measures step count may be verified as a step counter, but its validity as a measure of functional capacity in patients with heart failure requires additional evidence demonstrating the clinical relationship between step count metrics and the functional capacity construct. This validation may involve correlation with established clinical measures, analysis of sensitivity to clinically meaningful changes, and assessment of the measurement’s discrimination between patient populations with known differences in the clinical construct.

Data Quality and Verification Frameworks

Data quality assurance for wearable-generated evidence requires a comprehensive framework that addresses the unique characteristics and failure modes of wearable sensor data. Unlike data collected in controlled clinical settings by trained personnel using calibrated instruments, wearable data is collected in uncontrolled environments by patients whose compliance with device wearing and usage instructions may vary, using consumer-grade devices that may be subject to interference, displacement, and environmental effects.

Data Completeness and Wear Compliance

One of the most significant data quality challenges for wearable studies is ensuring adequate data completeness, which depends primarily on participant compliance with device wearing protocols. Participants may remove wearable devices for comfort, hygiene, charging, or personal preference, creating gaps in the data record that can affect the validity of derived endpoints. Data completeness requirements should be defined prospectively in the study protocol, specifying the minimum wearing time required for a valid measurement period, the methods for detecting and quantifying non-wear periods, the statistical approaches for handling missing data, and the criteria for excluding participant-periods or participants with inadequate data completeness.

Artifact Detection and Signal Quality

Wearable sensor data is susceptible to artifacts arising from device displacement, motion interference, environmental factors, and sensor malfunction. Effective data quality assurance requires automated artifact detection algorithms that identify and flag data segments that may be corrupted by artifacts, signal quality metrics that provide quantitative assessment of data reliability, data cleaning processes that remove or correct artifact-affected data segments while preserving valid data, and quality control dashboards that provide study team visibility into data quality metrics across the study population.

Clinical Endpoint Selection Using Wearable Data

The selection of clinical endpoints derived from wearable data requires careful consideration of clinical meaningfulness, measurement validity, regulatory acceptability, and practical feasibility. Endpoints may be based directly on sensor measurements, such as mean daily step count or time in target glucose range, or may be derived through algorithmic processing of sensor data, such as a composite mobility score or a sleep quality index.

Established and Novel Digital Endpoints

Some wearable-derived endpoints have achieved broad acceptance through accumulating evidence of clinical validity and regulatory precedent. Step count and physical activity measures from accelerometry have been used in multiple regulatory submissions and are generally accepted as clinically meaningful endpoints for conditions affecting mobility and functional capacity. Continuous glucose monitoring metrics, particularly time in range, have achieved regulatory recognition and clinical guideline endorsement as clinically meaningful endpoints for diabetes management. Heart rhythm monitoring endpoints from wearable cardiac devices have established regulatory precedent through multiple device clearances and clinical trial applications.

Novel digital endpoints, measurements that are newly developed or that lack established regulatory precedent, require more extensive validation and regulatory engagement. Novel endpoints may offer significant advantages over established measures, potentially capturing aspects of disease burden or treatment benefit that conventional endpoints miss, but they carry greater regulatory risk because the evidence supporting their clinical meaningfulness is less mature. Sponsors should engage the FDA early through pre-submission meetings when planning to use novel digital endpoints, providing the available evidence for the endpoint’s clinical validity and seeking agency feedback on the validation evidence needed to support its use in a regulatory submission.

Study Design Considerations for Wearable-Integrated Trials

Incorporating wearable devices into clinical trial designs introduces protocol design considerations that are distinct from those in traditional trials. These considerations address device deployment and training, data collection procedures, monitoring and compliance management, and the analytical approaches for wearable-derived endpoints.

Hybrid and Decentralized Trial Designs

Wearable technologies are natural enablers of decentralized and hybrid clinical trial designs that reduce the dependence on in-person clinical visits and enable data collection in the participant’s natural environment. In a decentralized trial design, wearable devices may serve as the primary or sole means of collecting efficacy and safety data, with study visits conducted remotely and data transmitted electronically from the participant’s device to the study database. In a hybrid design, wearable data collection supplements traditional clinical visit assessments, providing continuous data between visits that enriches the clinical picture obtained through episodic measurements.

Trial designs that incorporate wearable technologies must address device provisioning and setup logistics, including device selection, configuration, distribution, and technical support for participants; participant training on device use, care, and troubleshooting; data transmission architecture, including the connectivity method, data transfer frequency, and backup procedures for transmission failures; remote monitoring processes for detecting and addressing device issues, data quality problems, and compliance concerns in real time; and device return, data reconciliation, and device retirement processes at the conclusion of the study.

Real-World Evidence from Wearables in Regulatory Submissions

The incorporation of wearable-generated real-world evidence into regulatory submissions requires careful attention to the evidentiary standards, documentation requirements, and presentation approaches that regulatory reviewers expect. The strength of wearable-derived RWE depends not only on the quality of the underlying data but also on the clarity with which the sponsor demonstrates data provenance, analytical validity, and clinical relevance.

Evidence Package Structure

A regulatory submission incorporating wearable-derived evidence should include comprehensive documentation addressing the wearable technology’s verification and validation evidence; the study design and protocol, including the rationale for using wearable-derived endpoints; the data quality assurance framework, including data completeness requirements, artifact detection methods, and missing data handling approaches; the statistical analysis plan, including the primary analysis methodology and sensitivity analyses addressing potential sources of bias; the results, presented with appropriate characterization of the data quality metrics and any limitations; and a discussion of the clinical significance of the findings, including how wearable-derived evidence integrates with other available evidence in the submission.

Sensor Technology Validation and Analytical Performance

The analytical performance of the wearable sensor technology is the foundation upon which all subsequent evidence generation rests. Inadequate sensor performance introduces systematic bias or random error into the data that propagates through all analyses and conclusions derived from that data. Sensor validation must be conducted under conditions that reflect the full range of use conditions expected in the clinical application, addressing performance across different patient demographics, activity states, environmental conditions, and wear positions.

Reference Standard Comparison

Sensor validation studies compare wearable measurements against established reference standards under controlled conditions. For heart rate measurement, the reference standard is typically electrocardiography. For blood oxygen saturation, the reference standard is arterial blood gas analysis or bench-top pulse oximetry. For physical activity measurement, reference standards include direct observation, video recording, or research-grade motion capture systems. The validation study should assess accuracy, precision, and reliability across the full measurement range and under the conditions of intended use, with particular attention to performance at the boundaries of the measurement range where sensor performance may degrade.

Patient Engagement and Compliance in Wearable Studies

The success of wearable-integrated clinical studies depends fundamentally on participant willingness to wear and use the devices as intended throughout the study period. Participant compliance with wearable protocols is influenced by device comfort, battery life and charging requirements, the perceived burden of the technology, the participant’s understanding of the study’s purpose, and the quality of technical support available when problems arise.

Study designs should incorporate strategies for maximizing and monitoring compliance, including selection of wearable devices that balance data quality with wearability, with attention to device size, weight, comfort, and aesthetic acceptability; clear, accessible participant training on device use, including written instructions, video tutorials, and hands-on training sessions; ongoing technical support through helpdesk services, troubleshooting guides, and device replacement processes; compliance monitoring with near-real-time visibility into wearing patterns that enables proactive intervention when compliance declines; and incentive structures that encourage consistent device wearing without creating coercion or bias.

Data Infrastructure for Wearable Evidence Generation

The data infrastructure supporting wearable evidence generation must address the unique characteristics of wearable data, including high volume, continuous generation, variable data quality, and the need for both real-time monitoring and retrospective analysis. This infrastructure encompasses the full data pipeline from sensor data capture on the wearable device through data transmission, storage, quality assessment, processing, analysis, and regulatory submission.

Cloud-Based Data Platforms

Most wearable evidence generation programs employ cloud-based data platforms that receive data transmitted from participants’ devices, store the raw and processed data securely, execute quality assessment and processing pipelines, and provide analytical and visualization capabilities for study teams and data scientists. These platforms must comply with applicable data protection and security requirements, including 21 CFR Part 11 for electronic records used in FDA-regulated clinical investigations, HIPAA for protected health information, GDPR for data from European participants, and GCP requirements for clinical trial data management.

Privacy, Consent, and Data Governance

Wearable devices collect physiological and behavioral data of extraordinary granularity and continuity, raising privacy and consent considerations that exceed those of traditional clinical data collection. A wearable accelerometer continuously recording a participant’s movement patterns generates a detailed record of the participant’s daily activities, sleep schedule, and behavioral patterns that may reveal information beyond the clinical parameters of interest, including patterns of social activity, mobility limitations, and daily routines.

Informed Consent for Continuous Monitoring

Informed consent for wearable studies must clearly describe the types of data collected by the device, the duration and continuity of data collection, how the data will be stored, transmitted, and protected, who will have access to the data and for what purposes, the participant’s rights regarding their data including the ability to withdraw consent, and any risks associated with the collection of continuous physiological and behavioral data, including privacy risks that extend beyond the clinical parameters of interest. The continuous nature of wearable data collection may create consent challenges that do not arise in traditional clinical settings where data collection is episodic and bounded by clinical visit schedules.

Emerging Applications and Future Regulatory Directions

The application of wearable device data in clinical evidence generation is expanding rapidly across therapeutic areas and regulatory use cases. Emerging applications include the use of wearable-derived digital biomarkers as primary endpoints in pivotal clinical trials, the integration of wearable data into pharmacovigilance and postmarket surveillance programs, the development of wearable-based companion diagnostics that identify patients likely to benefit from specific treatments, and the use of wearable-generated real-world evidence to support label expansion submissions and health technology assessment processes.

Advancing Digital Endpoint Acceptance

The acceptance of wearable-derived digital endpoints by regulatory authorities is advancing through a combination of scientific evidence generation, regulatory engagement, and multi-stakeholder collaboration. Initiatives such as the Digital Medicine Society’s library of digital clinical measures, the Clinical Trials Transformation Initiative’s work on mobile clinical trials, and the FDA’s Patient-Focused Drug Development guidance program are contributing to a growing body of evidence and consensus around the validation, regulatory acceptance, and clinical meaningfulness of digital endpoints. As this evidence base matures, the range of regulatory contexts in which wearable-derived endpoints are accepted is likely to expand significantly.

Convergence with AI and Advanced Analytics

The convergence of wearable sensor technologies with artificial intelligence and advanced analytics is creating new categories of digital biomarkers and clinical evidence that were not possible with either technology alone. Machine learning algorithms applied to wearable sensor data can identify patterns and features in physiological data streams that are invisible to conventional analysis, potentially enabling earlier disease detection, more sensitive treatment response measurement, and more precise patient stratification. This convergence introduces additional regulatory considerations related to AI/ML algorithm validation, transparency, and the interaction between sensor data quality and algorithm performance.

The regulatory frameworks for wearable-derived evidence will continue to evolve as the technology matures, the evidence base grows, and regulatory authorities gain experience evaluating submissions that incorporate wearable data. Organizations that invest now in the validation methodologies, data infrastructure, regulatory expertise, and clinical evidence generation capabilities needed for wearable-integrated regulatory strategies will be positioned to leverage these technologies effectively as their regulatory acceptance expands. Those that defer this investment risk finding themselves without the capabilities needed to compete in an evidence generation landscape that increasingly values the continuous, real-world, patient-centric data that wearable technologies uniquely provide.