Leveraging “Big Data” for the Design and Execution of Clinical Trials
Randomized clinical trials form the cornerstone of evidence-based medicine and are required to accurately determine cause-effect relationships and treatment effects of medical interventions. Nonetheless, contemporary clinical trials are becoming increasingly difficult to execute and are hampered by slow patient enrollment, burdensome and extensive data collection, and high costs. Over the past decades, there has been an infusion of digital technology and computing power within healthcare. “Big data,” defined as data so large and complex that traditional mechanisms and software used to store and analyze data are insufficient, offers the potential of innovation and improvement for contemporary clinical trials. The primary focus of health technology to date has been direct patient care, but these platforms offer further potential to change the paradigm for conducting clinical trials and generating medical evidence. The digitalization of medical information allows data across multiple health systems to be integrated and centralized within readily analyzable common data models with standardized data definitions. Moreover, these technologies favor embedding clinical research within everyday clinical care, offering the benefits of generalizable study results, “re-use” of data already collected during routine patient care, and minimal burden of trial participation on patients and local study sites. “Big data” approaches and machine learning also may aid in phenotyping complex medical conditions and identifying optimal patient subsets for study in clinical trials. In this chapter, we review the current challenges facing traditional clinical trials and discuss the conceptual framework and rationale for merging clinical trials with the evolving field of health data science. We follow by outlining specific avenues through which “big data” have potential to reshape the way clinical trials are performed and by discussing respective advantages for purposes of generating high-quality, highly actionable, and patient-centered medical evidence.
