A new prediction algorithm for mortality and nursing home admission Lead Investigator: Yian Gu Institution : Columbia University E-Mail : yg2121@columbia.edu Proposal ID : 306 Proposal Description: Using longitudinal grade of Membership model (L-GoM) we have set up a prediction algorithm that accurately predicts the time of death and entering nursing home for subject at the early stage of Alzheimer disease. The developed algorithm is trained with Predictors-1 cohorts of study data (10 years of semiannually follow up) and evaluated with similar Predictors-2 cohorts of study data. Using the NACC data we propose to evaluate this prediction algorithm one more time on a completely separate cohort of population.