Digital Biomarker Discovery Project
The Digital Biomarker Discovery Project (DBDP) is an open-source platform for the development of digital biomarkers.
Click here to learn more and get involved.
Discover our blogs and tutorials here.
Digital Biomarker Discovery Education (dbdpED)
dbdpED is an educational platform for digital biomarker discovery.
With tutorials, case studies, and educational videos, dbdpED provides a resource for learning the steps to using mHealth and wearables data to discover digital biomarkers. dbdpED is for all ages and abilities. We believe anyone can discover digital biomarkers!
Learn more about our efforts here and here.
Infection Watch
“Infection watch” is a new and unique research initiative to promote early detection of COVID-19 infections from wearable device data. Infection watch will primarily be a feasibility study to explore the potential of wearables to detect COVID-19 infection.
Jupyter Health
The BIG IDEAs lab is in collaboration with experts from Berkeley, UCSF, and Project Jupyter to build Jupyter Health. This innovative open-source platform aims to enhance the secure collection of data from wearable sensors to electronic health record and to streamline the development, sharing, and validation of digital biomarker algorithms. We develop modular features for Jupyter Health, ranging from comprehensive summary statistics to dynamic user visualizations that can provide immediate insights into health data. While Jupyter Health is initially piloted for diabetes management, we envision Jupyter Health to expand its capabilities to cover a wide array of health conditions.
Dynamic Time Warping: Biomedical Signal Processing
DTW group aims to develop metrics to evaluate the performances of different DTW algorithms to matching time-series signals. Yihang also proposed a novel algorithm called eventDTW based on the peak information of the signals, and we demonstrated its better performance in aligning signals with different sampling.
BASS CONNECTIONS PROJECT
Refining & Expanding Duke's Wearable Infection Detection Platform:
Our research has demonstrated that there are changes in physiologic and behavioral parameters measured by wearables during the course of infection. In particular, anomalies in resting heart rate (HR), HR variability, blood oxygen saturation, sleep patterns, and physical activity have been shown to provide a time-series signature of infection in symptomatic as well as asymptomatic phases. As a result, such wearable "digital biomarkers" can be used to help predict, detect, and monitor disease as was demonstrated by our CovIdentify Study. CovIdentify's overarching objective was to discover and implement digital biomarkers from off-the-shelf wearable devices to develop, validate, and translate a continuous screening tool for COVID-19 infection. With CovIdentify, we successfully highlighted the important role wearable devices can play in detecting COVID-19 related illness as soon as physiologically possible.
We have now started to build an online infection detection platform that populates and translates wearable data from a variety of sources in an easy to use manner. The goal of this platform is to inform the onset and trajectory of illness for not only COVID-19 but other disease as well (e.g., influenza). The platform will also support future studies investigating the role of comorbidity and demographics in disease outcomes. We hope to further develop and utilize this platform with the help of a team of undergraduate/graduate students and funding through Bass Connections.
Outsourcing the Digital Biomarker Discovery Pipeline (DBDP):
Digital biomarkers transform data from wearables into indicators of health outcomes, allowing rapid detection, prevention, and management of many diseases. As commercial wearables become more prevalent, digital health has the potential to improve healthcare accessibility and equity.
To facilitate the advancement of digital health, the Digital Biomarker Discovery Pipeline (DBDP) has been created as an open-source initiative from the BIG IDEAs Lab at Duke. The objective of the DBDP is to disseminate and promote toolkits, reference methods, and data as the consensus community standard for digital biomarker development. Until now, the members of the DBDP have focused on developing modular algorithms researchers can use for data cleaning and preprocessing tasks, analysis, and predictive model development. Progress was hindered, however, in expanding the DBDP as an open source platform for other stakeholders in digital health.
Currently, clinicians or the general members of the community without proficient programming skills/computational background find difficult to fully participate in discovery of digital biomarkers. The DBDP's current interface is not user-friendly, and the strategies for outsourcing the resources available at DBDP were limited. Efforts still lack in understanding the challenges the stakeholders in digital health face, promoting community engagement, outsourcing the material, and quantifying and assessing growth of DBDP as an open source community.
Individual Research Projects
Digital Biomarkers of Circadian Disruption and their Utility in Cardiovascular Risk Assessment
Will Wang
Will's thesis project aims to detect and quantify cardiovascular responses to circadian disruption and compare digital biomarkers against more conventional circadian metrics for evaluating the cardiovascular effects of circadian disruptions. Such digital biomarkers are meant to provide longitudinal, accessible and individualized metrics for evaluating the cardiovascular risk resulting from circadian disruptions (e.g. shift workers).