Purpose: The major aim of this continuation project is to follow an established cohort of outpatients and well adults for 2.5 years to determine the long-term stability and predictive validity of preferences for life-sustaining treatment and health states near to or worse than death. Methods: Subjects include 5 groups of outpatients and 2 groups of well- adults (N=350; n=50/group AIDS, elderly patients, stroke, cancer, nursing home residents, young well-adults, older well-adults). Measurements of preferences for life-sustaining treatment and health states near to or worse than death, previously made at entry to the study and 6 months, will be repeated at 1.5 yrs and 2.5 yrs. Measurements methods include a 5-point Likert scale (treatment preferences), and a seven-point comparative rating scale and standard gamble scaling technique (health state preferences). Data analyses: Data will be analyzed to determine 1) the stability (est- retest reliability) of preferences at 1.5 and 2.5 yrs in participants with stable health status, and 2) predictive validity (ability to predict preferences for previously unexperienced treatments and health states) of preferences in patients with unstable health status. In addition, data analyses will be performed to develop models to explain changes in preferences over time, and to compare the behavior of direct and indirect preference scaling methods. Implications: These results will have substantial public health and policy ramifications because they will address the substantial proportion of individuals who complete advance care directives or otherwise express preferences for life-sustaining treatment well in advance of life- threatening illness. These results will afford a better understanding of the long-term stability and validity of these preferences and their appropriateness to clinical decision making in various setting.
