Bi-level heterogeneity modeling of functional performance degradation for the aging population

• DOI: 10.1080/24725579.2017.1339147
• Publication Date: 2017-06-12T14:37:47Z
• Authors: Nazmus Sakib; Xuxue Sun; Nan Kong; Hongdao Meng; Mingyang Li
• Keywords: Degradation

As the majority of service recipients U.S. long-term care system, elderly with disabilities experience heterogeneous functional limitation and performance degradation. Successful modeling degradation becomes great importance for decision making. Existing heterogeneity approaches are mainly restricted to single-level older adults. This article proposes a novel bi-level quantification framework systematically investigate at both sub-population individual levels. Specifically, Bayesian non-parametric inference is first proposed characterize sub-population-level heterogeneity. It relaxes conventional assumption pre-specifying fixed number sub-populations allows joint model estimation selection. Functional data analysis techniques further considered quantify nested within sub-populations. Based on extracted temporal variation signatures profiles, post-analysis performed identify relevant risk/protective factors in explaining To illustrate work demonstrate its validity, cognitive sample community-dwelling adults from Health Retirement Study provided.