End of Life Trajectories: A Prospective Model.

Principal Investigator: Douglas Wolf

Active Dates: 2012 – 2013

Description:

One of the consequences of the success of modern medicine is longer periods of survival in a fragile state, prompting growing recognition of, and attempts to define and describe, the increasingly common late-life condition of frailty. Improved understanding of the age profile of health and health care needs, along with their implications for disability levels and Medicare costs, requires extensive longitudinal data and complex analytical models. As a means of organizing and structuring the problem of modeling late-life patterns of disability, death, and care costs, this study will employ a recently developed typology of end-of-life functional trajectories, of which frailty is one type. The typology distinguishes among a set of distinctive pathways from full functioning to death, including variations in the timing of onset, the steepness and regularity of decline, and the period of time over which decline plays out. Although the concept of functional trajectories is plainly forward-looking, rooted in clinical observation of key health events and treatment junctures, nearly all of the empirical work to date that employs this conceptual scheme has used retrospective (or “look back”) methods. The retrospective approach, while useful and informative, is limited: it presupposes access to retrospective longitudinal data on decedents, and is therefore of little use for forecasting care needs or planning services. To further develop and apply this conceptual scheme, and to investigate its usefulness as a forecasting and planning tool in the area of population health and care costs, we will develop a prospective model of end-of-life functional trajectories. We will use the typology as a way of imposing a structure on what is otherwise ignored, or undifferentiated, heterogeneity in end-of-life functional and service-use patterns. In this project we will estimate estimate integrated models of mortality and disability using longitudinal data from the National Long Term Care Survey and the Health and Retirement Study, each of which can be linked to continuous Medicare claims records.