A digital biomarker is a measurable indicator of a biological state or treatment effect captured using digital technologies—typically via sensors in wearables, smartphones, or connected medical devices. It can be collected continuously or event-based and frequently provides real-world data, such as activity, sleep, movement patterns, or physiological signals.
Distinction and Typical Applications
Digital biomarkers are not equivalent to conventional laboratory parameters such as HbA1c or CRP, but rather complement them through digital, often high-frequency measurement series. They can take various forms: (i) passive measurements such as step count, heart rate variability, or tremor signals, (ii) active tests such as cognitive tasks in an app or standardized voice samples, and (iii) contextual data, such as usage and interaction patterns, provided these can be meaningfully interpreted from a medical perspective. It is important to clearly define whether a digital biomarker is used as an exploratory variable, as a secondary endpoint, or (less commonly) as a primary endpoint.
In practice, digital biomarkers are often discussed in the context of real world evidence, patient-reported outcomes, and electronic patient diaries. The key distinction: a patient-reported outcome is based on self-reporting, whereas digital biomarkers are objectively measured using sensors. Nevertheless, both are often combined in studies to integrate clinical relevance and the patient perspective.
Methodological Requirements: Validity, Reliability, Context
For a digital biomarker to be reliable in clinical trials, fundamental methodological questions must be addressed: Does the sensor measure what it claims to measure (validity)? Does it deliver comparable results upon repetition (reliability)? And can a clinically relevant statement actually be derived from the signal (clinical validation)?
- Analytical validation: Sensor and algorithm performance, comparison with reference methods, influence of artifacts (e.g., movement, skin contact, device position).
- Clinical validation: Correlation of the digital signal with clinical endpoints, disease activity, or treatment response, including subgroup analyses.
- Context and bias: Usage patterns, adherence, technical failures, different device generations, and software updates as potential confounders.
A common error is insufficient specification of analysis algorithms. Particularly with machine-learning-based methods, training data, model versioning, and change management must be transparently described; otherwise, reproducibility issues and audit findings are likely. Furthermore, missing-data mechanisms and protocol deviations should be addressed early, as digital measurements often exhibit more dropouts and data gaps than conventional visit-based measurements.
Regulatory classification in the EU/Germany
In the EU, regulatory classification depends heavily on whether the digital solution is classified as a medical device and whether it has a medical intended purpose. If a wearable or app is used as a medical device, requirements under EU MDR (EU) 2017/745 typically apply, including clinical evaluation, risk management, and, where applicable, post-market clinical follow-up. For medicinal product trials under EU CTR 536/2014, a digital biomarker may be part of the clinical trial protocol; in such cases, data integrity, data protection, and verification of data sources must be clearly described.
For Germany, data protection requirements (GDPR) and interaction with ethics committees are also relevant. Ethics committees frequently focus on informed consent regarding data processing, the burden on participants from sensor use, and the handling of “incidental findings” (e.g., abnormal cardiac rhythm data). In audits and inspections, it is also verified whether electronic data capture meets ALCOA principles and computerized system validation requirements.
Relevance for clinical trials
Digital biomarkers can make trials more efficient by enabling denser data collection and earlier detection of changes. This is particularly relevant in chronic diseases, neurological indications, or in oncology for functional parameters. At the same time, requirements for data management and statistical analysis increase: time series require clear aggregation rules, predefined analysis windows, and a precise definition of the analysis population.
From the sponsor and CRO perspective, early decisions are critical: Which devices will be permitted (bring-your-own-device vs. study-provided device)? How will software updates be controlled? What processes exist for device accountability, training, helpdesk, and data transmission? Full-service CROs such as mediconomics typically support specification in the clinical trial protocol, setup of interfaces (e.g., to electronic data capture), and demonstration of data quality to investigators and authorities.
Frequently Asked Questions (FAQ)
Can a digital biomarker be a primary endpoint?
Yes, in principle, but it requires particularly robust clinical validation and transparent documentation of the measurement and analysis procedures. In many projects, secondary or exploratory use is initially chosen to build evidence.
How are data gaps from wearables handled in the analysis?
This should be planned statistically in advance, including rules for minimum wear time, imputation, or sensitivity analyses. Without predefined rules, the risk of bias and discussions during results interpretation increases.
What role does the MDR play for digital biomarkers?
As soon as the digital solution is used as a medical device or derives medical statements, MDR requirements may apply. In such cases, clinical evaluation, risk management, and post-market surveillance must be considered, among other aspects.
Regulatory References
- EU Regulation (EU) No 536/2014 (Clinical Trials Regulation, CTR): Requirements for clinical trials and the clinical trial protocol for medicinal products.
- EU Regulation (EU) 2017/745 (MDR): Regulatory requirements for medical devices including software and wearables with medical intended purpose.
- ICH E6(R3) Good Clinical Practice: Principles on data quality, documentation, systems, and oversight in clinical trials.
In practice, it is also important to note that the technical infrastructure (device management, data transmission, backup, role and rights concept) should be tested before first-patient-first-visit. A data management plan with clear responsibilities, monitoring strategy (e.g., remote monitoring), and defined quality metrics helps to detect deviations early and initiate CAPA measures. Training materials and documented instructions are also essential to achieve inspection readiness.
In practice, it is also important to note that the technical infrastructure (device management, data transmission, backup, role and rights concept) should be tested before first-patient-first-visit. A data management plan with clear responsibilities, monitoring strategy (e.g., remote monitoring), and defined quality metrics helps to detect deviations early and initiate CAPA measures. Training materials and documented instructions are also essential to achieve inspection readiness.