The overall objective of this proposal is to develop proper methods for characterizing aging-related changes in activities of daily living (ADLs) while accounting for differential survival patterns among older persons. The Yale Precipitating Events Project (PEP) is an ongoing longitudinal study of ADL disability, which has followed a large cohort of community-living older persons for over ten years. Monthly interviews and comprehensive assessments at eighteen-month intervals have been conducted to assess ADL disability, possible risk factors, potential intervening events, and deaths. PEP has documented relevant information that helps to elucidate the mechanisms underlying the occurrence and recovery of ADL disability and provided strong evidence that disability is a reversible and recurrent event with considerable heterogeneity and multiple potential pathways. However participants with greater disability tend to die earlier, hence the estimation of the extent of and overall burden of disability using available follow-up observations may be biased because the amount of ADL data varies according to participants'survival times. In current analytic methods for longitudinal study, death is typically dealt with in a similar fashion as dropout in non-aging research or treated as independent censoring, and often retrospective analysis is conducted conditioning on the proximity of death time. Most dropout models assume the existence of potential response after dropout, which does not apply to the aging-related functional change in older persons when death is the main reason for truncated measurements. Thus, a different analytic strategy is needed to deal with death in longitudinal aging research or any research in which death is an inevitable event.
Our first aim i s to depict the ADL ordinal response, which we view as manifested trajectory of underlying disabling process that is related to the risk process of death.
Our second aim i s to discover the distinct patterns representing the complex relationship between the disabling process and death.
Our third aim i s to study the dynamic transitions among the ADL states (i.e., (re)occurrence, worsening, improvement, and recovery of ADL disability) and death. In all the aims, the role of risk factors such as age, gender, cognitive impairment, physical activity level, depression, existence of acute or chronic comorbid conditions (e.g., stroke, hearing and sensory impairment) and physical frailty will be investigated. The findings from our proposal will provide aging researchers with an informative picture of aging-related ADL trajectories with varying survival patterns. The use of PEP data with monthly assessments addresses the previous concern that lack of frequent measurements that cover more than the period immediately before death has limited our understanding of the risk of death and biased the findings on ADL trajectories. The methods developed in this proposal are generally applicable to other longitudinal studies where death is a competing event, such as the longitudinal study of cognitive function in Veterans with Alzheimer's disease, study of the immune function in AIDS patients and study of quality of life in post-treatment cancer patients.

Public Health Relevance

Disability in activities of daily living (ADL) in older persons incurs high cost in personal and long-term cares often leads to adverse health outcomes. Investigating patterns and factors associated with occurrence, deterioration, improvement and recovery of ADL and with death in older persons thus constitutes an important clinical and scientific task.

Agency
National Institute of Health (NIH)
Institute
National Institute on Aging (NIA)
Type
Research Project (R01)
Project #
5R01AG031850-03
Application #
8063930
Study Section
Neurological, Aging and Musculoskeletal Epidemiology (NAME)
Program Officer
Joseph, Lyndon
Project Start
2009-05-15
Project End
2013-04-30
Budget Start
2011-06-15
Budget End
2013-04-30
Support Year
3
Fiscal Year
2011
Total Cost
$313,533
Indirect Cost
Name
Yale University
Department
Public Health & Prev Medicine
Type
Schools of Medicine
DUNS #
043207562
City
New Haven
State
CT
Country
United States
Zip Code
06520
Murphy, Terrence E; Allore, Heather G; Han, Ling et al. (2015) A longitudinal, observational study with many repeated measures demonstrated improved precision of individual survival curves using Bayesian joint modeling of disability and survival. Exp Aging Res 41:221-39
Makris, Una E; Fraenkel, Liana; Han, Ling et al. (2014) Risk factors for restricting back pain in older persons. J Am Med Dir Assoc 15:62-7
Lin, Haiqun; Han, Ling; Peduzzi, Peter N et al. (2014) A dynamic trajectory class model for intensive longitudinal categorical outcome. Stat Med 33:2645-64
Lin, Haiqun; Allore, Heather G; McAvay, Gail et al. (2013) A method for partitioning the attributable fraction of multiple time-dependent coexisting risk factors for an adverse health outcome. Am J Public Health 103:177-82
Gill, Thomas M; Gahbauer, Evelyne A; Lin, Haiqun et al. (2013) Comparisons between older men and women in the trajectory and burden of disability over the course of nearly 14 years. J Am Med Dir Assoc 14:280-6
Han, Ling; Allore, Heather; Murphy, Terrence et al. (2013) Dynamics of functional aging based on latent-class trajectories of activities of daily living. Ann Epidemiol 23:87-92
Murphy, T E; Han, L; Allore, H G et al. (2011) Treatment of death in the analysis of longitudinal studies of gerontological outcomes. J Gerontol A Biol Sci Med Sci 66:109-14
Cao, Yanrong; Lin, Haiqun; Wu, Tracy Z et al. (2010) Penalized spline estimation for functional coefficient regression models. Comput Stat Data Anal 54:891-905
Lin, Haiqun; Guo, Zhenchao; Peduzzi, Peter N et al. (2008) A semiparametric transition model with latent traits for longitudinal multistate data. Biometrics 64:1032-42