The convergence of technology and healthcare is transforming the landscape at an unprecedented pace in life sciences. Digital biomarkers have emerged as powerful tools that hold immense potential for revolutionizing the biotech and pharmaceutical industries. The annual growth rate of clinical trial adoption of digital products across all clinical phases from 2000-2020 is 34% (Marra, C. et al., 2020). These digital indicators derived from various digital devices and sensors are reshaping the way we collect, analyze, and interpret patient data. In this blog post, we will explore the remarkable rise of digital biomarkers and their profound impact on the life sciences sector.
What are Digital Biomarkers?
“The FDA defines a digital biomarker to be a characteristic or set of characteristics, collected from digital health technologies, that is measured as an indicator of normal biological processes, pathogenic processes, or responses to an exposure or intervention, including therapeutic interventions.”
Vasudevan, Srikanth, et al. NPJ digital medicine (2022).
The process of digital biomarker discovery employs sensors and wearable devices to collect data on physiological and behavioral characteristics which must first be processed and validated, and then can be engineered into digital biomarkers -- often via bespoke algorithms and machine learning approaches.
Our team has completed several projects related to digital biomarkers:
1. A Data Science Platform to Support Biomarker Discovery: diagnostic biomarkers, for example, using video and voice recordings to diagnose autism spectrum disorder.
2. Treatment Response Monitoring: pharmaco-dynamic response biomarkers such as those acquired when using a wearable device to detect movement changes in response to a prescribed drug to treat Parkinson’s disease.
3. Guide Medical Diagnosis with Biometric Eye Tracking: monitoring characteristics over time to assess the status of a medical condition; For example, using performance and eye-tracking on tasks completed on an iPad to assess cognitive function in traumatic brain injury patients over the course of a year.
There are several checkpoints to be a viable digital biomarker candidate. For example, the collection and engineering processes used to generate the biomarker must be reliable, reproducible, and have reasonable compliance from patients. They must also be clinically relevant i.e. distinguish between controls and a disease state or correlate with clinical endpoints (Song, Youngjae, et al., 2022).
Once validated, digital biomarkers offer many advantages over traditional biomarkers: less resource intensive and invasive data collection methods, enhanced data collection, early detection/predictive insights, continuous real-world evidence, remote patient monitoring, and many others. However, collecting digital biomarkers presents its own unique challenges. For one, extensive compute and storage resources may be required for data collection and analysis. Additionally, since it may be somewhat easier to collect more potential biomarkers, it is also easier to introduce noise and irrelevant data which then require sophisticated data science and modeling workflows to extract meaningful signals. Both challenges usually necessitate expertise in data engineering and machine learning approaches and the resources to apply that expertise. Furthermore these data collection, processing, and biomarker discovery approaches need to be standardized to facilitate generalizability and reproducibility. Overcoming these challenges in digital biomarker discovery requires expertise in data engineering and data science that many companies lack the resources and knowledge to take advantage of.
1. Enhanced Data Collection: Digital biomarkers overcome the limitations of traditional, intermittent data collection methods by providing continuous, real-time monitoring. This abundance of data allows for a more comprehensive understanding of an individual's health and disease progression, leading to improved decision-making processes and the possibility of personalized interventions.
2. Objective and Quantifiable Measurements: Subjective assessments and self-reporting can introduce biases and inaccuracies in data collection. Digital biomarkers, on the other hand, offer objective and precise measurements, reducing the potential for errors and improving the reliability of data analysis.
3. Early Detection and Predictive Insights: By remotely monitoring patients over time, digital biomarkers can provide early warning signs of health deterioration or the onset of diseases. This early detection enables timely interventions, leading to better patient outcomes and potentially reducing healthcare costs.
4. Increased Efficiency in Clinical Trials: Digital biomarker data collection is usually less resource intensive (does not require interaction with clinician) and invasive than traditional biomarker collection such as blood-based biomarkers. By providing remote access, clinical trials can be accessible to more diverse patient populations, as they are not limited to individuals who live nearby and have the ability to travel to a trial site. This can help with clinical trial recruitment by widening the recruitment pool and help retention by reducing the burden of having to travel to a trial site.
5. Better Endpoints for Clinical Trial Success: Relying on patient reported outcomes (PROs) that are subjective presents many challenges. Relying on episodic clinical visits is not ideal. Capturing continuous biomarkers that are more consistent, more quantified, less biased, and less reliant on patient compliance, can serve as better proxies for treatment response than current standards such as the 6-minute walk test. Further, digital approaches support decentralized clinical trials (DCTs).
1. Clinical Trials: Digital biomarkers are reshaping the landscape of clinical trials by enabling remote monitoring, real-time data collection, and improved patient compliance. These capabilities streamline the trial process, accelerate recruitment, and enhance data quality, ultimately expediting the development of novel therapies. People who traditionally could not come on site can now receive a device in the mail and be monitored and participate remotely.
2. Personalized Medicine: With the advent of digital biomarkers, more forms of personalized medicine are becoming a reality. By leveraging the data obtained from wearable devices and other sensors, healthcare providers can tailor treatments based on an individual's unique characteristics, optimizing efficacy, and minimizing adverse effects.
3. Disease Management and Remote Patient Monitoring: Digital biomarkers offer tremendous potential in managing chronic diseases. By continuously monitoring key physiological parameters and behaviors, healthcare professionals can proactively intervene, optimize treatment regimens, and enhance patient engagement and adherence.
4. Real-World Evidence Generation: The availability of large-scale, real-time data through digital biomarkers facilitates the generation of real-world evidence (RWE). RWE plays a pivotal role in assessing the safety and effectiveness of treatments in diverse patient populations, supporting regulatory decision-making, and informing healthcare policies.
Protocols, preprocessing, and feature handling algorithms need to be standardized and applied across wearable/sensor brands to ensure digital biomarker work is reproducible across clinical trials. This includes using adequate amounts of training and testing data for biomarker identification. To ensure this is possible, it is critical that published clinical trial outcomes report sufficient details on their methods so they can be replicated in future trials. Data management, preprocessing pipelines, and feature engineering/extraction pipelines require expertise in the fields of data engineering and machine learning. Understanding feature importance and interpretability is often a key part of the usability of clinical biomarkers; Therefore, researchers need to be able to understand how to apply feature reduction and machine learning approaches to improve signal to noise ratio within data and interpretability of algorithms that are often described as a “black box”.
Looking ahead, the integration of digital biomarkers with advanced technologies like artificial intelligence, machine learning, and big data analytics will further amplify their impact. As the field continues to evolve, collaboration between biotech, pharmaceutical companies, and technology providers will be crucial in harnessing the full potential of digital biomarkers and translating them into tangible improvements in patient care and outcomes. Currently, physical activity and cardiac rhythm studies are the most common types of digital biomarker published research (Motahari-Nezhad, et al. JMIR mHealth and uHealth; 2022). However, there is significant potential for digital biomarkers in studying neurological and psychological disorders (psychologytoday): assessing real time, continuous patterns of sleep, mood, eating habits, activity levels, autonomic activity, cognition, etc.
Our platform, VIVO, is our answer to the clinical data challenges in adopting digital biomarkers. VIVO can operate as a stand alone application or be configured to seamlessly integrate with existing data infrastructure to extract, analyze, and identify both digital and molecular biomarkers.
The rise of digital biomarkers marks a significant paradigm shift in the life sciences sector, offering a transformative approach to patient monitoring, disease management, and drug development. The integration of digital biomarkers in the biotech and pharmaceutical industries has the potential to enhance clinical trial efficiency, enable personalized medicine, and improve patient care.
We build data science applications to support drug development, digital & molecular biomarker discovery, and digital health. Our team works at the intersection of biology and technology to accelerate innovation. If you have an AI/ML-related question or would like to discuss your AI strategy, we’d love to hear from you! Reach out today at email@example.com, on Twitter @mercurydatasci, or on LinkedIn.
Vasudevan, S., Saha, A., Tarver, M. E., & Patel, B. (2022). Digital biomarkers: convergence of digital health technologies and biomarkers. NPJ digital medicine, 5(1), 36.
Song, Youngjae, et al. "Pathological Digital Biomarkers: Validation and Application." Applied Sciences 12.19 (2022): 9823.
Marra, C. et al. Quantifying the use of connected digital products in clinical research. NPJ Digit. Med. 3, 50 (2020).
Motahari-Nezhad, H., Fgaier, M., Mahdi Abid, M., Péntek, M., Gulácsi, L., & Zrubka, Z. (2022). Digital Biomarker–Based Studies: Scoping Review of Systematic Reviews. JMIR mHealth and uHealth, 10(10), e35722.
A rich text element can be used with static or dynamic content. For static content, just drop it into any page and begin editing. For dynamic content, add a rich text field to any collection and then connect a rich text element to that field in the settings panel. Voila!
Below is an image