Exploring the Feasibility of Using Real-World Data from a Large Clinical Data Research Network to Simulate Clinical Trials of Alzheimer's Disease

Clinical trials are essential but often have high financial costs and long execution time. Trial simulation using real world data (RWD) could potentially provide insights on a treatment9s efficacy and safety before running a large-scale trial. In this work, we explored the feasibility of using RWD from a large clinical data research network to simulate a randomized controlled trial of Alzheimer9s disease considering two different scenarios: an one-arm simulation of the standard-of-care control arm; and a two-arm simulation comparing treatment safety between the intervention and control arms with proper patient matching algorithms. We followed original trial9s design and addressed some key questions, including how to translate trial criteria to database queries and establish measures of safety (i.e., serious adverse events) from RWD. Our simulation generated results comparable to the original trial, but also exposed gaps in both trial simulation methodology and the generalizability issue of clinical trials.