Abstract
Objective: The aim of this study is to evaluate the relationship of leading causes of death with gradients of cognitive impairment and multimorbidity. Method: This is a population-based study using data from the linked 1992-2010 Health and Retirement Study and National Death Index (n = 9,691). Multimorbidity is defined as a combination of chronic conditions, functional limitations, and geriatric syndromes. Regression trees and Random Forest identified which combinations of multimorbidity associated with causes of death. Results: Multimorbidity is common in the study population. Heart disease is the leading cause in all groups, but with a larger percentage of deaths in the mild and moderate/severe cognitively impaired groups than among the noncognitively impaired. The different "paths" down the regression trees show that the distribution of causes of death changes with different combinations of multimorbidity. Discussion: Understanding the considerable heterogeneity in chronic conditions, functional limitations, geriatric syndromes, and causes of death among people with cognitive impairment can target care management and resource allocation.