CASE STUDY: Impact of Data Quantity & Quality on Development of Robust Digital Endpoints in Decentralized Studies
COVID-19 pandemic has further accelerated the adoption of digital health technology in clinical trials enabling over 220 digitally augmented trials in 2021 alone. Researchers and trialists can reach a larger diverse target population and frequently assess their health by collecting individualized real-world data actively and passively. Such a scalable remote observational model helps investigators develop a better and holistic understanding of people’s day-to-day experiences of living with a health condition that ultimately will help improve our understanding of individualized health-related behavior and related outcomes.
However, several real-world challenges related to the quantity and quality of real-world data collected from the target population have surfaced. These include but are not limited to equitable recruitment and long-term retention of the target population. In this talk, I will summarize the ongoing learnings related to real-world data collection from three large-scale multinational studies that have enrolled over 20,000 participants (from young and healthy kids to patients with severe depression). Specifically, I will focus on two key topics:
1. Potential biases may impact the collection of health-related data using active and passive data streams, such as participants' socio-demographics, willingness/concerns, varying participation incentives, disease severity, and technical limitations across iOS and Android ecosystems.
2. Varying quality of sensor-based data collected from smartphones and wearables and its impact on developing robust digital endpoints. I will demonstrate digital signal processing-based data quality metrics to help assess the completeness, correctness, and consistency within and across participants and devices. For example, real-world factors such as device issues (e.g., hardware, software, network) and user-related (e.g., unintended or non-compliant device usage, non-wear).
Abhishek Pratap, Group Head AI & Digital Health, Centre for Addiction and Mental Health (CAMH), Toronto