Cross-validation of a risk score algorithm to predict future cognitive

 
   Lead Investigator:    Daniel Brown
   Institution      :    University of Virginia Health System
   E-Mail           :    brown.daniel@gmail.com
   Proposal ID      :    778



   Proposal Description:
   1) As information regarding prevention and treatment interventions continues to   grow, the need for early and accurate prediction of future cognitive decline      becomes increasingly important. Doing so can allow those at higher risk for       subsequent neurocognitive impairment to initiate counteracting steps, and take    such action sooner. This in turn may lead to better care, including delayed and   slowed decline, ultimately resulting in higher quality and potentially longer     life.                                                                             A risk score algorithm was previously developed using whole brain and             hippocampal volume, the Montreal Cognitive Assessment, subjective cognitive       complaint, and sex to predict future cognitive status (impaired or unimpaired).   However, this algorithm has not been cross-validated, which substantially         limits current clinical utility. Cross validation using NACC data would           substantially increase the utility of such findings thereby providing a direct,   simple, feasible, and easily accessible clinical application. The goal of the     proposed project is to cross-validate the previously developed (aforementioned)   risk score algorithm. The expected outcome is that the algorithm will show        similarly strong psychometric properties when cross-validated against NACC        data, compared to the sample from which it was originally derived.                2) The objective is to cross-validate the aforementioned algorithm by testing     its psychometric properties against NACC data.