Assessing Alzheimer's Disease progression across multiple domains with ordered

 
   Lead Investigator:    Linus Jonsson
   Institution      :    Karolinska Institute
   E-Mail           :    linus.jonsson@ki.se
   Proposal ID      :    733



   Proposal Description:
   This project aims to develop a disease model that simulates progression on  discrete states across three domains (cognition, function and behavior).  Parametric methods will be applied to allow estimation of transition  probabilities across a wide state-space. The resulting model will be useful for  predicting long-term consequences of novel therapies affecting any or all of  the three domains. Second aim:  to identify determinants of disease progression in patients with prodromal AD. This project would be conducted in collaborations with IBM Watson Health and H. Lundbeck A/S.    Sub-aim (Handels et al.)  Modelling the natural disease progression of Alzheimer's disease    Our goal is to develop a model that enables the simulation of the natural   disease progression of AD for health-economic decision analytic frameworks. We aim to explore various parametric (survival) models to estimate the natural progression in terms of (changes in) continuous or categorized scores of cognition, behavior, function, and institutionalization and mortality based on a set of (demographic) predictors. We expect these models can be used to address heterogeneity, structural uncertainty and external validity in decision analytic frameworks.