Digital/Mobile Health in Clinical Trials

Organizers: Kevin Anstrom (Duke)
Chair: Kevin Anstrom (Duke)
Vice Chair: Xiaofei Wang (Duke)

Speakers:

Manisha Desai (Stanford)
Jessilyn Dunn (Duke)
Vanja Vlajnic (Bayer)

Abstracts:

Title: Design Lessons Learned from the Apple Heart Study
Speaker: Manisha Desai (Stanford)

Clinical trials with pragmatic features require special attention to their design and analysis. Often such features are incorporated in order to increase participant enrollment, engagement, and/or the amount of data collected on participants. Today’s pragmatic studies leverage real-world digital platforms such as electronic health records, continuous glucose monitors, and accelerometers that may alleviate the burden of data collection from the participant but that are not necessarily intended for research purposes. The Apple Heart Study was a prospective, single-arm, site-less pragmatic study designed to evaluate the ability of an app to detect atrial fibrillation. The implementation of the study contained numerous pragmatic features, such as the use of an app for enrollment and a wearable device for data collection, that enabled a high volume of patient enrollment and accompanying data. These pragmatic elements led to challenges surrounding the study implementation. Challenges we encountered included identifying the number of unique participants, linking complex and disparate data elements to each participant, ordering data longitudinally within participant, and engaging participants through study completion. Novel solutions to these challenges were derived, often in real-time. These solutions inform future designs, and data analysis and data monitoring plans for studies with similar pragmatic features. Importantly, the issues presented demonstrate the critical role of data science in modern pragmatic trials.

Title: Precision Healthcare Through Multi-scale Biomedical Data Integration
Speaker: Jessilyn Dunn (Duke)

The healthcare landscape faces rising costs and declining health outcomes as a result of a reactive, treatment-centered medical system. In particular, cardiometabolic disease is the leading cause of death and healthcare cost worldwide. Although this condition is often preventable and treatable when detected early, its incidence continues to rise. The shift toward a proactive, prevention-based healthcare system requires precision methods to assess risk and deliver effective interventions. The long-term goal of my research is to develop precision health tools that use continuous health monitoring data to reveal changes in health status and assist medical decision-makers in delivering precision therapies and just-in-time interventions.
Recent technological advancements make it possible to closely and continuously monitor individuals on multiple scales in real time while also incorporating genetic, environmental, and lifestyle information. We are collecting and using this multi-scale biomedical data to gain a more precise understanding of health and disease at molecular and physiological levels and developing actionable, predictive health models for improving cardiometabolic outcomes. We are simultaneously developing tools for the digital health community, including the Digital Biomarker Discovery Pipeline (DBDP), to facilitate the use of mobile device data in healthcare.

Title: Clinical Trials in a Digital Age: Analysis of High-Dimensional Data from Wearables
Speaker: Vanja Vlajnic (Bayer)

As the population continues to grow and live longer, the effects of chronic disease and illness become more salient to both the individual and society at large. As researchers continue to examine novel therapeutics and devices for potential treatments, new technologies are emerging as part of the landscape to assist in the collection of patient data. In many therapeutic areas, such as cardiovascular disease, the analysis of data from clinical trials play a fundamental role in identifying potential signals indicating an improvement in disease status or patient quality of life. As of late, wearable devices such as accelerometers have gained popularity in the clinical trial setting due to their non-invasiveness coupled with the ability to collect data continuously and passively. These data are collected at much higher resolutions than traditionally found in clinical trials and thus allow for the opportunity to examine such things as physical activity, sedentary behavior, and sleep of patients throughout the course of the trial and their relationships to clinical outcomes. This is in contrast to one of the more traditional endpoints in cardiovascular trials, the 6-minute walking test, which often only collects a single measurement on a patient at several time periods across the trial. This talk will provide a case study of the utilization of a wearable device in the cardiovascular space as well as discuss some of the regulatory implications therein. Furthermore we will examine traditional analytical methods as well as a deep learning approach to analyzing the high-resolution data.