Data-driven frailty subtypes in normal cognitive aging, mild cognitive
impairment, and Alzheimer?s disease: A latent transition analysis 
   Lead Investigator:    Linzy Bohn
   Institution      :    University of Alberta
   E-Mail           :    linzy@ualberta.ca
   Proposal ID      :    1462



   Proposal Description:
   Frailty is a geriatric condition that reflects multisystem decline and an increased vulnerability to adverse aging outcomes. The frailty phenotype and the frailty index are the two leading approaches for measuring and conceptualizing frailty. A small but growing body of research has explored complementary, data-driven approaches to frailty measurement and conceptualization. Specifically, our team recently examined a database of baseline multimorbidity indicators (n = 30) on a large sample (n = 649) of cognitively normal older adults. We used latent profile analysis to detect, classify, and validate early frailty subtypes (Bohn, Zheng, McFall, Dixon, under review).     Objective: In this proposed study, we seek to access the NACC UDS in order to conduct similar data-driven analyses and test the generalizability of our approach and findings across a wider range of clinical cohorts, measurement occasions, and aging morbidity indicators.    Specific Aims: We plan to assemble two-wave longitudinal data for a large number of indicators of aging multimorbidity (n = 50+) using the following clinical cohorts: normal aging, mild cognitive impairment (MCI), and Alzheimer?s disease (AD). We intend to examine the following specific research aims using data-driven analytic techniques:     1. Ascertain which modalities of frailty subtypes can be empirically determined at baseline and then validated across one follow-up wave and clinical cohorts  2. Examine transitions between frailty subtypes over time within cohorts   3. Identify significant predictors of the probability of transitioning between frailty subtypes over time (e.g., APOE)  4. Explore how frailty subtypes relate to performance and decline in specific domains of cognition (e.g., memory variable).     Novelty and Impact: This research marks a new contribution to the emerging literature on data-driven frailty subtype assessment and frailty transitions in normal aging, MCI, and AD. Findings from this study may advance