Develop a trial simulation framework through causal-principled machine
learning approaches for Alzheimer?s (AD) disease and AD-related dementia 


   Note: An Official Proposal has not been Submitted to NACC.


   Lead Investigator:    Jiang Bian
   Institution      :    University of Florida
   E-Mail           :    bianjiang@ufl.edu
   Proposal ID      :    1743



   Proposal Description:
   In a recent work, we tested the ?target trial? approach of using large collections of observational real-world data (e.g., electronic health records and administrative claims) to emulate randomized controlled trials (i.e., NCT00478205), aiming to derive reliable treatment effects of AD drugs.  Built upon this work, we will use NACC data as the external validation datasets to validate the findings.  Through this process, we also aim to develop a trial simulation framework through causal-principles machine learning approaches based on observational data.       Such a trial simulation framework can be effective in obtaining causal estimates with observational data, that would have broader impact of using NACC data as both a knowledge discovery and a knowledge validation resource.  For example, in a NIA-funded project, we aim to develop methods for discovering drug repurposing signals for AD/ADRD from real world data (RWD) obtained via two large clinical research networks.  We proposed to evaluate and validate our top-ranked drug repurposing hypotheses on the high-quality NACC data, through the proposed trial simulation framework for causal estimates.  The success of this project will lead to novel computational tools for the identification of high-quality drug repurposing signals and promising AD/ADRD drug repurposing signals validated with diverse approaches.